V O L U M E I FINAL REPORT FOR THE IFPRI/ISRA ...
V O L U M E I
FINAL REPORT FOR THE IFPRI/ISRA STUDY
OF
CONSUMPTION AND SUPPLY IMPACTS
OF
AGRICULTURAL PRICE POLICIES IN THE PEANUT BASIN
A
N
D
SENEGAL ORIENTAL
BY
Valerie Kelly
Thomas Reardon
A. Abdoulaye Fall
Bocar Diagana
Lisa McNeilly
(USAID Contract No. PIO/T 685-0281-3-70042)
September 1993
International Food Policy
Insti tut Senegalais de
Research Institute (IFPRI)
Recherches Agricoles (ISRA)
.
i
VOLUME I
FINAL REPORT FOR THE IFPRI/ISRA STUDY
OF
CONSUMPTION AND SUPPLY IMPACTS
OF
AGRICULTURAL PRICE POLICIES IN THE PEANUT BASIN
AND
SENEGAL ORIENTAL
BY
Valerie Kelly
Thomas Reardon
A. Abdoulaye Fall
Bocar Diagana
Lisa McNeilly
(USAID Contract No. PIO/T 685-0281-3-70042)
September 1993
International Food Policy
Institut Senegalais de
Research Institute (IFPRI)
Recherches Agricoles (ISRA)
TABLE OF CONTENTS
I .
INTRODUCTION
Background of Study and Policy Context . . . . . . . . . . . .
1
Study Objectives and Research Questions
. . . . . . . . . . .
3
Research Methods . . . . . . . . . . . . . . . . . . . . . . .
5
Analysis Procedures and Conventions
8
.
.
.
.
.
.
.
.
.
.
.
.
.
Organization of the Report . . . . . . . . . . . . . . . . . .
9
2.
CHARACTERISTICS OF AGROCLIMATIC REGIONS AND STUDY ZONES
. . . 1 1
Location and Characteristics of Agroclimatic Regions . . . . . 1 1
Representivity and Characteristics of Study Zones
. . . . . . 16
Summary of Main Points . . . . . . . . . . . . . . . . . . . .
31
3.
SAMPLE CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . 3 4
Introduction
. . . . . . . . . . . . . . . . . . . . . . . . . 3 4
Sample Demographics
. . . . . . . . . . . . . . . . . . . . .
35
Household Demographics . . . . . . . . . . . . . . . . . . . .
35
Socioeconomic Characteristics of the Household . . . . . . . . 40
Relevance of Household Characteristics . . . . . . . . . . . .
52
for Agricultural Price Policies
Summary of Main Points . . . . . . . . . . . . . . . . . . . . 54
4.
DESCRIPTION AND ANALYSIS OF RURAL AND URBAN INCOME
. . . . . . 56
Introduction
. . . . . . . . . . . . . . . . . . . . . . . . . 56
Research Methods . . . . . . . . . . . . . . . . . . . . . . .
57
Levels, Distribution, and Sources of Income
. . . . . . . . .
61
Intersectoral Links of Income Generating Activities
. . . . . 74
in Rural Areas
Importance of the Informal Sector in Urban Zones . . . . . . . 75
Patterns of Income Diversificat i o n . . . . . . . . . . . . . . 7 6
Summary of Main PO ints . . . . . . . . . . . . . . . . . . . .82
5.
CROP PRODUCTION . . . . . . . . . . . . . . . . . . . . . . . 8 5
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 85
Methodology . . . . . . . . . . . . . . . . . . . . . . . . . 8 5
Gross and Net Cropping Incomes . . . . . . . . . . . . . . . . 88
Land Cultivation and Input Use Patterns
. . . . . . . . . . . go
Returns to Family Labor
. . . . . . . . . . . . . . . . . . .
gI
Distribution of Income by Gender . . . . . . . . . . . . . . .
93
Cereal Production Sufficiency Levels . . . . . . . . . . . . .
g4
Summary of Main Points . . . . . . . . . . . . . . . . . . . .
96
i
.
:,
.;
APPENDIXES
VOLUME 2 PART I
4.1
-
Tables of Average Household Income by Sector, Village, Zone,
and Year
5.1
-
Tables of Analysis Results Describing Characteristics of
Crop Production and Income
6.1
-
Tables of Analysis Results for Sales, Purchases, and Gifts
of Agricultural Products
6.2
-
Comments on zones
7.1
-
Frequency Distribution of Net Income from Livestock
Transactions
7.2
-
Tables of Analysis Results Describing Characteristics of
Livestock Sales and Changes in Stocks
7.3
-
Gini-Coefficients of Cattle and Small Ruminant Distribution
from Ndiaye (1993)
7.4
-
Statement of Net Annual Income (in FCFA) Associated with
Small Ruminants for the Typical Household from Ndiaye (1993)
7.5
-
Summary, Conclusions, and Policy Recommendations from Ndiaye
(1993) Analysis of IFPRI/ISRA Data
8.1
-
Detailed Zone Specific Description of Food Consumption
Patterns
8.2
-
Note sur les Resultas de la Transformation du Sorgho et du
Mais
8.3
-
Tables and Figures of Analysis for Food Consumption Patterns
9.1
-
Consumer and Producer Prices
9.2
-
Comments on Quality of Price Data for Specific Products and
Zones
9.3
-
Consumer Price Trends for Cereals and Pulses
iii
.’
:
.
.;
:
.’ ~
..
LIST OF TABLES
Table 2.1 --
Demographic characteristics of agroclimatic..........17
regions and study zones
Table 2.2 --
Infrastructure and institutional coverage............19
of study zones suggested by reconnaissance survey
Table 2.3 --
Relative importance of various crops in 1987 and.....23
1988 as declared by village representatives
during reconnaissance survey
Table 2.4 --
Agricultural constraints, fallow, and animal.........25
traction patterns suggested by the village
reconnaissance survey
Table 2.5 --
Relative importance of animal husbandry and other....26
off-farm activities as suggested by the
reconnaissance survey
Table 2.6 --
Food security and the relative importance of.........28
different coping strategies (to compensate for food
production shortfalls) suggested by the
reconnaissance survey
Table 3.1 --
Characteristics of the rural sample by zone..........36
Table 3.2 --
Characteristics of the urban sample by zone..........37
Table 3.3 --
Age and gender distribution of rural sample..........38
compared with theoretical population
Table 3.4 --
Household characteristics for the rural zones........39
Table 3.5 --
Household characteristics for the urban zones........41
Table 3.6 --
Characteristics of household heads for rural zones...42
Table 3.7 --
Characteristics of household heads in the
urban sample........................ . . . . . . . ..*....... 44
Table 3.8 --
Access to land for persons >= 15 years in rural......47
zones
Table 3.9 --
Livestock assets in rural zones......................48
Table 3.10 --
Standard of living indicators for households in......50
urban sample
iV
,
.
.
.
.’
.
.’
:
.
Table 3.11 --
History of households in the urban zone..............51
Table 4.1
--
Classification of income-generating activities.......60
by sector
Table 4.2
--
Level and distribution of net household income.......62
per AE (FCFA) by zone and year
Table 4.3
--
Frequency distribution of average net household......65
incomes (FCFA/AE) by zone and year for rural zones
Table 4.4
--
Frequency distribution of average net household......66
incomes (FCFA/AE) in urban zones
Table 4.5
--
Net household income per AE by zone, year, and.......67
sector for rural sample
Table 4.6
--
Net household income per AE (FCFA) by town and.......68
sector for urban sample
Table 4.7
--
Average net total income per AE (FCFA) and shares....72
by sector and tercile for each rural zone and year
Table 4.8
--
Average net total income per AE (FCFA) and shares....73
by tercile and sector for each urban zone
Table 5.1
--
Cereal/pulse production sufficiency rations by.......87
zone and harvest year
Table 5.2
--
Comparison of average gross and net cropping.........88
income by crop for 1989/90
Table 5.3
--
Levels of net cropping income and the cost of........89
minimum needs
Table 5.4
--
Median returns to family labor in cropping...........92
activities for 1989/90
Table 5.5
--
Distribution of crop income by gender................93
Table 5.6
--
Average cereal production sufficiency ratios.........94
for 1988/89 and 1989/90
Table 5.7
--
Average kilos of surplus cereals per AE:.............95
1988/89 and 1989/90
Table 6.1
--
Share of sales revenue from sales of home............102
production by zone, harvest year, and
crop purchased
Table 6.2
--
Volume of cereal sales by zone and year..............104
,
w
.
.
.
:
.’
:
Tab1 e 6.3
--
Estimated volume of coarse grain sales by............104
zone and year
Tab1 e 6.4
--
Share of purchases for home consumption by zone,.....105
harvest year, and product
Tab1 e 6.5
--
Volume of cereals purchased by zone, year, and.......106
product
Tab1 e 6.6
--
Share of purchases for home consumption by zone ,..... 107
harvest year, and product
Table 6.7
--
Estimated volume of coarse grain purchases by........109
zone and year
Table 6.8
--
Estimated net vol ume of coarse grain transactions....109
by zone and year (quantity marketed minus
quantity produced)
Table 6.9
--
Share and total value of seed purchases by zone,.....110
year, and product
Table 6.10 --
Share and tota
value of gifts out by zone, year,.. ..llO
and product
Table 6.11 --
Share and total value of gifts received by zone,.....112
year, and product
Table 6.12 --
Share of total cereal production marketed by zone....124
and year
Table 6.13 --
Share of total peanut production marketed by zone....124
and year
Table 6.14 --
Distribution of households by kilos of cereals.......126
marketed in 1988/89 and 1989/90
Table 7.1
--
Summary of net income from animal transactions.......131
Table 8.1
--
Summary of findings on caloric intake levels.........155
and shares
Table 8.2
--
Share of minimum calorie requirements covered........159
by product group, zone, and year
Table 8.3
--
Product composit on of cereals consumed by zone......160
and year
Table 8.4
--
Product composition of pulses consumed by zone.......161
and year
Table 8.5
--
Classification of consumption
patterns...............164
\\
vi
Table 8.6
-- Share of cereals and pulses purchased by zone........164
and year
Table 8.7 --
Share of rice in cereal purchases by zone and year...166
Table 9.1
-- Income patterns......................................206
Table 9.2
-- Policy effects via production output side............207
Table 9.3
-- Production input expenditure patterns and policy.....209
effect
Table 9.4 --
Expenditure patterns.................................210
Table 9.5
-- Policy effects via decreases in real income on.......212
consumption side
Table 9.6
-- Calculation of net effects on real household.........214
1988/89
vii
‘.
.’
:
LIST OF FIGURES
Figure 2.1 --
Agroclimatic and Study Zones . . . . . . . . . . 12
Covered by the IFPRI/ISRA Survey
Figure 6.1 --
Relative Shares of Sales by Gender
113
.
for (B) Millet, (B) Peanuts, (C) Cotton,' ' l
(D) Cowpeas, and (E) Other for Harvest
Years 1988/89 and 1989/90
Figure 6.2 --
Relative Shares of Purchases by Gender . . . . 118
for (A) Millet, (B) Peanuts, (C) Cotton,
(D) Cowpeas, and (E) Other Products for
Harvest Years 1988/89 and 1989/90
Figure 8.1 --
Consumption by Product Group, Zone, and . . . 154
Year
Figure 8.2 --
Cereal and Pulse Consumption by Product, . . . 158
Zone, and Year
Figure 8.3 --
Composition of "Other Food" by Product,
. . , 162
Zone, and Year
Figure 8.4 --
Sources of Cereals and Pulses by Zone . . . . . 165
and Year
Figure 8.5 --
Product Composition of Cereal and Pulse . . . . 168
Consumption Compared for (A) NonMarket and
(B) Market Villages
Figure 8.6 --
Sources of Cereals and Pulses . . . . . . . . . 170
Compared for (A) NonMarket and
(B) Market Villages
Figure 8.7 --
Composition of Cereals and Pulses . . . . . . . 171
Purchased Compared (A) NonMarket and
(B) Market Villages
viii
,
:
:
:
:’
I
‘.
.I
‘.
Figure 8.8 --
Income Effects on Level and Composition . . . 173
of Cereal and Pulse Consumption for two
Harvest Years
Figure 8.9 --
Income Effects on Sources of Cereals . . . . . 174
and Pulses Consumed for Two Harvest Years
Figure 8.10 --
Income Effects on Composition of Cereals . . . 177
and Pulses Purchased
Figure 8.11 --
Seasonality of Cereal and Pulse Consumption . . 179
Figure 8.12 --
Seasonality of Cereal and Pulse Purchases . . . 184
Figure 9.1 --
Consumer Price Trends for Millet/Sorghum . . . 193
Figure 9.2 --
Consumer Price Trends for Maize . . . . . . . . 195
Figure 9.3 --
Consumer Price Trends for Shelled Peanuts . . . 196
Figure 9.4 --
Consumer Price Trends for Cowpeas . . . . . . . 197
ix
.:
.’
.I.
:
.
.
.
:
:
1 .
INTRODUCTION
BACKGROUND OF THE STUDY AND POLICY CONTEXT
In the early 1980s the Government of Senegal (60s) began a
structural adjustment program.
Some of the most important structural
adjustment reforms being considered at that time were the liberalization
of cereal markets, the privatization of agricultural input markets, and
adjustments in exchange rates.
These are reforms that can be expected
to change levels of real income, and prices of inputs, outputs and
consumption goods.
The IFPRI/ISRA
study, Supp7y and Consumption Impacts
of Agricultura7 Price Policies in the Peanut Basin and Senegal Oriental,
was designed to provide a better understanding of how households are
affected by and respond to these types of policy reforms.
In the mid-1980s, the GOS relied heavily on price policies to
elicit desired production and consumption responses.
I n
1 9 8 6
the
official price for both coarse grains and peanuts (the major cash crop
in most of the rainfed zones) was raised from 70 to 90 FCFA/kilo. These
increases in producer price were designed to compensate for increases in
input costs brought about by the elimination of subsidies and sharp
restrictions on access to credit -- measures introduced between 1980 and
1985 in conjunction with Senegalese structural adjustment programs.
A series' of market liberalization policies was introduced during
the second half of the 1980s.
Among the steps taken to liberalize the
cereals market was the replacement of the official coarse grain price
with a designated floor and ceiling price in 1987. The Commissariat de
Securite Alimentaire (CSA) was given the mandate to intervene in cereal
markets to keep prices within the designated ranges.
By 1988, CSA's
role was diminished from one of intervention to one of simple price
monitoring.
Other important changes were the lifting of the ban on
private sector coarse grain trading and the development of a market
information system that provided the public with weekly cereal prices
for major markets.
These actions were designed to reduce the consumer
price of local cereals by lowering marketing costs and increasing
competition.
Such a policy, if successful, would permit the consumer
price to fall without significant reductions in producer prices.
The potential demand augmenting impact of steps taken to lower
consumer prices of coarse grains was,
however, diminished by the GOS's
decision to maintain control of the imported rice sector and to lower
rice prices from 160 to 130 FCFA/kilo following political disturbances
after the 1988 elections.
This action reduced the price differential
between coarse grains and imported rice thereby reducing the consumer's
price incentive to consume local cereals.
Steps were also taken to reduce government intervention in the
peanut sector and permit licensed private traders to participate in ttie
first stages of peanut collection activities.
These actions were to
)
..
.’
.
2
make the exportation of peanuts more profitable by realizing economies
in marketing rather than by reducing producer prices.
As stated in the
NPA, the GOS was seeking policies that would encourage farmers to
maintain levels of peanut production realized in the early 1980s while
increasing coarse grain production and overall farm income.
In brief, the decade of the 1980s witnessed a vast array of policy
initiatives that would significantly change rural incomes (through both
production and consumption effects) and urban incomes (through changes
in consumer prices).
Unfortunately,
at the time the changes were
implemented it was extremely difficult for policy analysts to gauge the
potential impact of their interventions.
One of the most important types of knowledge needed to predict the
likely consumption and supply impacts of these policy changes is an
understanding of how producers and consumers respond to price and income
changes.
To fully analyze the potential impacts of these policy
initiatives one would need to understand rural and urban consumption
patterns for locally produced millet, maize, sorghum, and rice vs.
imported rice and wheat.
One would want to know how these patterns vary
with changes in prices, as well as such nonprice factors as income,
location,
and infrastructures for marketing and processing. In rural
households, one also needs to know how decisions concerning production,
stocking, and marketing of cereal and cash crops respond to the same set
of price and nonprice factors.
Very little is known about Senegalese consumption patterns, except
for short periods in specific zones (Kramer 1984). It is not clear how
the large demand for imported rice, or for that matter any other cereal,
is distributed over the population.
While it is clear that rice and
wheat are widely consumed in Dakar, various limited-extent surveys have
turned up evidence of significant rice consumption in rural areas (e.g.,
Benoit-Cattin 1987).
Moreover the responsiveness of these patterns to
changes in prices and incomes is not known.
In 1984 Kramer pointed out
that "income elasticities are largely unknown for Senegal. The same is
true of price elasticities of demand... there are almost no estimates of
cross-price elasticities for major cereals." Researchers have found the
lack of such estimates a severe constraint when they attempt to model
aggregate response to price policy options in Senegal (see Braverman and
Hammer 1986; and Abt Associates 1985).
More recently, Kite (1992) reviewed the current evidence on food
consumption patterns,
particularly estimates of price and income
elasticities.
He concludes that ' . ..there is very little "hard"
(recent,
or reliable) information about consumption parameters for
Senegalese families, so a quantitative measure of likely impacts of
pricing policy is difficult."
The picture is not much better on the crop production and supply
side.
Although there is substantial information available on crop
output response to various technologies, there remains a lack of "hard"
:
.
.’
.
3
data on how producers respond to changes in prices of both inputs and
outputs.'
Work by Martin (1988) uses a linear programming model to
predict farmer response to changes in relative prices given a set of
input/output relationships for several different levels of technology.
However,
his input/output data were developed in consultation with
knowledgeable individuals (agronomists, extension workers, and so on)
and are not based on household data.
The model also does not take into
account nonagricultural options for generating income that are available
to rural households.
The IFPRI/ISRA study was designed during the above described period
of rapidly changing agricultural policies.
The study was designed as a
partial response to the need for better data and analysis of the
potential impacts that changes in relative prices might have on the
consumption and supply behavior of Senegalese households. The response
is partial because the study covers only the Peanut Basin and Senegal
Oriental.
These agroclimatic regions do, however, cover 85 percent of
the rural population, 60 percent of the total Senegalese population and
90 percent of both peanut and coarse grain production.
The study may also be thought of as a partial response to data and
analysis needs because it provides only two years of household data,
limiting the inferences that can be drawn from the study to the specific
combination of prices and policies that prevailed during the study
period.
In recognizing these limitations, it is also important to view
this project in the broader Senegalese context where most household data
sets cover only one year.
The fact that the IFPRI/ISRA study collected
data on the same households over two years permits both cross-sectional
and inter-annual comparisons of responses to changes in policies.
STUDY OBJECTIVES AND RESEARCH QUESTIONS
The ISRA/IFPRI project was designed to improve the data base for
and the analysis of income generating and consumption behavior in rural
and urban households throughout the Peanut Basin and central Senegal
Oriental.
The emphasis in rural zones is on the strategies employed by
households to secure both income and food over time through a wide
variety of economic activities.
The study puts the agricultural sector
in the context of the overall rural economy by covering all aspects of
household production and consumption behavior rather than limiting
itself to the cropping sector (as is often done in rural household
studies).
The urban areas are studied to gain insights on levels of
income and consumption patterns , important determinants of urban demand
for crops produced by rural households.
By looking in detail at how farmers allocate time and resources
across agricultural and non-agricultural activities, the study provides
information concerning the factors that influence decisions to produce
different kinds of crops
and to undertake various non-cropping
activities.
By looking at expenditures and consumption of home produced
4
products,, the study provides a detailed picture of how households use
their income and the factors that influence consumption behavior.
Relative prices and income levels are but two of the explanatory
factors examined on both the production and consumption side. Others
are size of household, access to markets, access to credit, availability
of transportation services, access to land, livestock opportunities and
livestock holdings, to name but a few.
Survey data are used to describe household behavior prevailing
during the survey period.
The data permit one to examine how selected
policy interventions would be likely to influence behavior in the future
-- particularly crop production patterns, food security strategies, and
real incomes.
The study permits one to address the following questions:
Composition of Grain Consumption
0
Are households in the study regions consuming substantial
quantities of rice and wheat? To what extent have these products
"penetrated" extra-Dakar consumption patterns?
0
Is there substantial effective demand for maize (a crop thought to
have greater productive potential than millet)?
How does this
differ between households which produce maize and those which do
not?
0
How are decisions about the product composition of consumption,
production, and sales linked to the overall resource and cash flow
strategies of rural households? How do millet and sorghum fit in?
maize? peanuts?
cotton?
Source of Products
0
For rural households in these regions, how important is purchased
food in total consumption?
How important are transfers such as
gifts from relatives in Dakar, donor food aid, or gifts from
neighbors?
What effect is this having on the composition of
consumption?
0
Are the purchases and sales mainly taking place in the village
itself, in weekly markets, in regional markets?
Levels of Consumption
0
What is the average level of calories consumed per adult equivalent
per day in each study zone?
0
Are there entire zones, or at least specific groups within zones,
which are at
risk
in terms of food consumption?
What
characteristics identify these households?
., . .
:
‘.
I
.
_’
.;
5
Responsiveness to Prices
l
How sensitive is the consumption of millet in the face of changes
in its price?
to changes in income? in the face of changes in the
prices of rice, maize, or sorghum?
If millet consumption drops in
the wake of a millet price increase, toward what product do
consumers move?
l
How sensitive are production and sales of millet and maize to
changes in their own prices?
to changes in peanut or cotton
prices? to changes in income from noncropping sources and overall
income?
RESEARCH METHODS
The research reported here was a cooperative effort between staff
economists at IFPRI and ISRA. Project design began in 1987 as a follow-
up to a jointly sponsored IFPRI/ISRA regional seminar on cereal
marketing in West Africa.
Project funding was approved by U.S.A.I.D. in
April 1988 and fieldwork began the following June.
Sample Selection
A village reconnaissance was used to obtain a broad understanding
of how variables that normally influence economic behavior differed
within and across agroclimatic regions in the Peanut Basin and Senegal
Oriental.
Knowledge of differences and similarities among the
agroclimatic regions was used to assure that the purposive selection of
study zones and villages would produce as representative a sample as
possible, given resource constraints.
The reconnaissance survey collected data on zone characteristics
such as,
ethnic
composition,
marketing
and
transportation
infrastructure,
and access to various services
(health, credit,
education, extension, water).
The reconnaissance also provided data on
villagers'
perceptions
of their economic well being and income
generating capacity (cereal
sufficiency
levels,
crop production
constraints, degree and nature of income diversification).
The reconnaissance survey used an informal interview procedure, not
a structured questionnaire.
Data were collected in group meetings
attended, at a minimum, by the village chief and a group of elders.
Usually, other residents of the village (both men and women) attended
and participated in the discussions.
Reconnaissance results suggested that five agroclimatic regions
needed to be covered to obtain a
sample that was generally
representative of the entire Peanut Basin and central Senegal Oriental.
The five agroclimatic regions were divided into eight representative
study zones -- six rural and two urban as follows:
0
1
,
.
.’
:
6
Agroclimatic regions
Study Zones
Sample
Villages
1.
Northern Peanut
Sagatta
3
Basin
2.
Central Peanut
Niakhar
3
Basin
Colobane
3
Kaolack (urban)
1 town
3.
Southwest Peanut
Passy
3
Basin
4.
Southeast Peanut
Dioly
3
Basin
5.
Central Senegal
Missirah
3
Oriental
Tambacounda (urban)
1 town
In selecting study zones within each agroclimatic region, both physical
and socio-economic factors were considered to assure that study zones
were representative of the larger agroclimatic region.
Three villages were selected to represent each rural study zone.
One village was always the site of a weekly market, while the remaining
villages had poorer access to market infrastructure.
This
stratification was done to permit analysis of the influence that access
to market infrastructure has on household behavior. The three villages
selected were not more than 50 kilometers from each other, thereby
permitting coverage by a single enumerator. The populations living in
the villages selected had to reflect the full diversity of the study
zone (as revealed by the reconnaissance survey) in terms of economic,
agronomic,
infrastructural,
and socio-cultural characteristics.
In each village 12 households were randomly selected from a list of
all households in the village.
This provided a total sample of 36
households per study zone,
Due to attrition, the number of households
enumerated and analyzed declined in some zones.
In most cases, however,
the number of households did not drop below thirty,'
The urban study zones of Kaolack and Tambacounda were selected
because they are the urban focal points for the majority of rural zones
covered.
Forty households were randomly selected in each town by
' Readers desiring more information on sampling procedures
'
should refer to Kelly and Reardon (1989).
:
7
referring to detailed ma;ps of the two areas prepared for the national
census conducted in 1988.
Survey Methods
The study period covered two harvest years (1 October 1988 through
30 September 1990) for all rural areas but the Northern Peanut Basin
(October 1988 through September 1989). The latter zone was dropped in
the second year so more researcher time could be devoted to the
preparation of the urban study.
The urban study covered one year (1
July 1990 - 30 June 1991).
Data were collected using intensive, repeated, household-level
interviews.
The data collected fall into the following categories:
Household demographics;
Food consumption;
Prices and quantities for all purchases and sales of principal
crops and livestock;
Prices and quantities
for expenditures on services and
products other than principal crops;
Cash transfers;
Household assets;
Net income from all economic activities;
Gross income and input costs for most economic activities;
Full input/output data (including labor) by plot for 1989/90
and 1990/91 crop seasons.
An extensive set of analyses has been completed to date and
reported in numerous project reports, working papers, and conference
papers.
These analyses can be grouped into the following general
categories:
* In addition to the data collected on the zones covered by this
report, we have some supplementary data from both the reconnaissance
and the household surveys for the Central Peanut Basin (Ndioum
Gainthe), the Southeastern Peanut Basin (Mereto), and the Upper
Casamance (Diaobe).
These data can be used in the future as a control
on results obtained for other study zones in the agro-climatic
regions.
:
*’
‘Y’
:
:
.
‘.
:
8
Food consumption: Calculation of average levels of caloric
intake per adult equivalent and product composition of food
consumption, by zone, season, income group, and access to
commercial infrastructure.
Income levels:
Calculation of average levels, share by
sector,
and distribution (Gini coefficient) of income per
capita and per adult equivalent by zone and income tercile;
Income diversification: modeling of the determinants of income
diversification and development of a topology of income
diversification.
Crop production:
Calculation of the value of crop production
per adult equivalent, returns to labor in crop production, and
degree of cereal
self-sufficiency provided by household
production; calculations are differentiated by zone, by level
of commercial infrastructure, and by income category.
0
Crop and livestock transactions:
Calculation of average values and quantities of products
bought and sold per adult equivalent plus average transaction-
derived prices by month for each principal product;
description of frequencies for place of transaction, persons
involved (household status and gender), sources of financing,
uses of revenues, and reasons for transactions.
0
Price trends:
Development of a full consumer price series for
agricultural product purchases during harvest years 1988/89
and 1989/90 based on transaction derived survey prices
supplemented with data from the Commissariat de la Securite
Alimentaire (CSA) and Michigan State University.
0
Price analysis: Simulation of the short run impact that
changes in the relative prices of cereal and cash crops will
have on real household income.
ANALYSIS PROCEDURES AND CONVENTIONS
Analyses to date have been done separately for each zone, no
attempt has been made to generalize results across all zones.
Within each zone, observations are weighted by household size (AE)
and a coefficient to adjust for the intentional over sampling of market
villages required for analyses of the influence of infrastructure on
household behavior.
The results presented for each individual chapter are based on the
data for all households having a complete set of data for that topic.
This means that the results presented in different chapters are not
s
.’
t
.
.’
.:
. -.
‘.
9
based on the exact same set of households.
This procedure was used
because:
(1) Data entry and analyses were performed sequentially, topic by
topic, so that interim reports on selected themes such as food
consumption and crop transactions could be issued.
When the
consumption data were analyzed, for example, it was not yet known
which households would also have a full set of consistent data for
income or for crop transactions.
(2) Given the relatively small sample size in each zone,
it was
considered best to use the maximum number of observations possible
for each type of analysis.
Once all the data were entered and the households with a full set
of data for all topics were identified, a few analyses were rerun with
the reduced set of households.
A comparison of these results with those
obtained in the earlier analyses suggests that the results are not very
sensitive to changes in sample size.
Given time constraints, it was not
feasible to rerun all analyses with the same set of households.
ORGANIZATION OF THE REPORT
The physical and socio-economic characteristics of the agroclimatic
regions and study zones are described in Chapter 2. The characteristics
of sample households are described in Chapter 3.
Chapters 4 through 9 contain the study's principal findings
concerning household income, crop production, crop disposal, livestock
transactions, food consumption, and prices.
Chapter 10 contains a resume of the principal findings and a
discussion of the policy implications.
.,,
:
.
.J
.’
:
10
REFERENCES
Abt Associates, Inc. 1985.
"Senegal agricultural policy analysis,
agricultural
policy
analysis project", April
(submitted to
USAID/Senegal).
Benoit-Cattin, M.
unpublished.
"Explaining rice consumption in the
Senegalese peanut basin", in Delgado, Reardon Matlon (eds.)
Commodity priorities and changing food patterns in West Africa.
mimeo, IFPRI, Washington, D.C.
Braverman, A. and J.S. Hammer. 1986. "Multimarket analysis of
agricultural pricing policies in Senegal", in I. Singh, L. Squire,
J.
Strauss, eds.
Aqricultural households models, Baltimore,
Maryland, USA: Johns Hopkins University Press for the World Bank.
Kelly, V. and T. Reardon. 1989. Samplinq methods used in the IFPRI/ISRA
study of consumption and supply impacts of aqricultural pricinq
policies. June. IFPRI, Washington, D.C.
Kramer, C. 1984.
"Food consumption, food demand in Senegal: Summary of
literature and suggestions for needed research", submitted to
USAID/Senegal.
Kite, Rod. 1992. "Evidence on food consumption patterns and behavior in
Senegal:
Implications
for the food policy dialogue", mimeo,
USAID/Dakar, October.
Martin, F. 1988. Food security and comparative advantaqe in Seneqal: A
micro-macro approach. Ph.D. diss., Michigan State University, E.
Lansing, U.S.A.
.
.
;
..i.
!“.
:
11
2.
CHARACTERISTICS OF AGROCLIMATIC REGIONS
AND STUDY ZONES
This chapter describes the agroclimatic regions and study zones
covered by the IFPRI/ISRA survey.
The information helps the reader to
interpret the report's descriptive and analytical findings in the
context of the physical and socioeconomic characteristics of the study
zones.
The first section describes the characteristics of the agroclimatic
regions while the second section describes the characteristics of the
rural and urban study zones selected to represent each agroclimatic
region.
Information concerning the agroclimatic regions is drawn
primarily from secondary sources.
The description of IFPRI/ISRA study
zones draws primarily on the village reconnaissance surveys and not at
all on the IFPRI/ISRA household surveys. The picture of the study zones
obtained from the reconnaissance data provides a general idea of how the
interviewers' perceive of their zone.
These data permit one to develop
hypothesis about the zones that are later tested with much more detailed
household survey data.
LOCATION AND CHARACTERISTICS OF AGROCLIMATIC REGIONS’
Figure 2.1 is a map that locates each of the agroclimatic zones and
study zones covered by the survey.
Northern Peanut Basin
The Northern Peanut Basin covers the administrative region of Louga
and all its arrondissements, except Darou Mousty and Dahra.
It covers
10,746 square kilometers (urban centers not included) and had a rural
population of 264,136 persons in 1988, and a rural population density of
' All the demographic statistics on the agroclimatic regions come from
preliminary 1988 census data provided by the Bureau National de
Recensement (September 1988).
The area data come from Secretariat
d'Etat a la Decentralisation (1986). The description of soils,
climate, and cropping systems of the regions draws heavily from
l/Atlas du S&-&al Editions Jeune Afrique, 1983, second edition. The
rainfall data come from l/Atlas du Senegal (Jeune Afrique 1983) and Le
Borgne (1988).
.
12
Figure 2.1 -- Agroclimatic and study zones covered by the IFPRI/ISRA
survey
.,’
.
.
.
i
..‘,,
.
:.
.
.’
:
:
.
.
.
.
:
13
25 persons per square kilometers.'
The main ethnic groups are the Wolof
in the west and the Fulani in the east.
The profile of the region is very uniform, with sandy plateaus and
dunes. The soils called dior by the Wolofs are ferric and not leached.
The Northern Peanut Basin is found in the most arid zone of the Sahel.
The area is prone to the harmattan, a hot, dry wind. The average annual
rainfall varies between 300 and 500 mm and rises as one goes from north
to south.
Agriculture is characterized by a traditional cropping system with
a biannual rotation of millet and peanuts.
Toward the middle of the
198Os, there was substantial growth in the production of cowpeas because
its short cycle is well adapted to the low rainfall and short rainy
season.
The stock of agricultural equipment dates from the Programme
Agricole (1960-80). Animal traction (horse and donkey) is common.
Climate and technical factors affect the type and extent of animal
husbandry practiced in the region.
It is mainly pastoral, based on
transhumance (seasonal migration). In the rainy season, the herds stay
in the region around temporary watering holes; in the dry season they
are led south and west where there are pastures, wells, and markets.
This animal husbandry focuses on cattle and is practiced mainly by the
Fulani.
Sedentary animal husbandry is practiced by the Wolof in the
western part of the region.
The infrastructure network is not well developed. A dearth of
correctly functioning wells creates a lack of fresh water.
The medical
and sanitary infrastructure is very insufficient. The dispersion of the
population aggravates poor access to medical care. Louga (the regional
capital), Kebemer,
and Linguere are the administrative and urban
centers.
The road and rail networks are limited to joining
administrative centers.
The embryonic
industry that exists is
concentrated in Louga.
The rural institutions in the region are Societe de Developpement
et de la Vulgarisation Agricole (SODEVA) (extension of cereal production
technologies and administration of the cowpea program) and some private
voluntary organization (PVO) activity such as World Vision's dry-season
vegetable gardening projects.
Central Peanut Basin
The Central Peanut Basin covers a part of the administrative region
of Louga (i.e., the arrondissements of Darou Mousty and Dahra), part of
the administrative region of Thies,
and part of the administrative
region of Diourbel, and partially includes those of Fatick, Kaolack, and
Tambacounda. It covers 33,958 square kilometers and had a population of
' The total population (including urban areas) was 336,570 persons in
1988, comprising 31,811 compounds and 35,357 households.
.
14
2,254,355 persons
in 1988 of which 1,740,233 are rura1,3 with a
population density of 66 persons/square kilometers (of which 51 are
rural).
The main ethnic groups are the Wolof in most of the region, and
the Serer, especially in the administrative regions of Thies and of
Fatick. In the eastern parts of the zones, there are also Fulani,
Toucouleur, and Mandingue.
Apart from the cliffs of Thies in the far west of the study region,
the land is even and flat.
Dior soils are predominant in all the zones
except for the ferric, leached soils in the northern part. The region
is in rainfall isohyets of 500 to 700 mm. The climate is hot and subject
to the harmattan.
Despite ethnic and rainfall variation over the region, the cropping
system everywhere is dominated by millet and peanuts. Millet continues
to be mainly a subsistence crop.
Peanuts are both a food crop and the
main cash crop.
These two are followed by cowpeas and sorghum in the
center, and by maize as one goes east. Cropping practices are relatively
homogeneous,
although there is a modest diversity linked to differences
in ethnic groups, to the degree of land constraint, and to the degree of
integration between cropping and animal husbandry.
The infrastructure is relatively well developed, having benefitted
from the important peanut economy.
The large towns (Kaolack, Thies,
Diourbel) grew with the peanut economy. The main industrial facilities
are based on phosphate mining and on peanut processing (Societe
Electrique et Industrielle du Basl--SEIB--and Societe Nationale de
Commercialisation des Oleagineux du Senegal--SONACOS). These provide
substantial local employment. There is also intense commercial activity,
especially in the Kaolack region, where there are a number of well-
developed permanent and seasonal markets. The road and rail network in
this region is much more dense than in the other study regions.
Many firms and other rural institutions are present in the region
for marketing, or for technical training and organization of producers
(SODEVA, Caisse de Securite Alimentaire, Societe de Developpement des
Fibres Textiles). The marketing of peanuts brings with it a strong
presence of SONACOS, of cooperatives, and of private storage units
(OPS), which manage peanut collection points. Tourism is developing,
mainly on the Petite Cdte, which attracts substantial foreign exchange
and stimulates local industry (particularly cottage manufacturing).
There are also many PVOs with development projects in the region.
Southwest Peanut Basin
This region covers a part of the administrative regions of Fatick
and Kaolack. It covers 6,765 square kilometers and had a population of
3 In 1988, 195,178 compounds, subdivided into 243,585 households were
registered in the census.
1
,’
.;
.
.
:
15
398,396 persons in 1988, and a density of 59 persons/square kilometers.
The main ethnic groups are the Serer and the Wolof.
As one goes from north to south in this region, the soils go from
being ferric/leached to red-ferric and semileached in the lower plateaus
of the northern Saloum.
In the western coastal area, the soils are
salinized and of the mangrove type because of seawater. The sea climate
affects the overall climate of the zone, and the annual average rainfall
is between 800 mm in the north and 1000 mm in the south.
The main crops are millet and peanuts, although there is some crop
diversification into maize and sorghum.
Use of animal traction is
widespread; horse-drawn plows predominate but cattle-driven traction is
more prevalent than in the center and north of the Peanut Basin.
Vegetable gardening is important in the dry season.
In contrast to the North and Central Peanut Basin, there are no
large towns; instead there are a number of semiurban areas where the
administrative infrastructure is located. The delta of the Saloum River
is an important tourist spot.
SODEVA operated a maize program in the
rural community of Keur Samba Gueye.
Southeast Peanut Basin
The region is bordered in the south by the frontier with The
Gambia;
the west-east boundaries go from Nioro (Kaolack) to the
arrondissement of Koumpentoum (Department of Tamba). This is a mainly
rural region with an area of 6,659 square kilometers and a population of
214,829 persons in 1988,
and a density of 32 persons per square
kilometers.
The Wolof are the main ethnic group, followed by the
Fulani.
There are some Manding in the eastern part of the region.
The Southeast Peanut Basin has plateaus with rocky soils. The
climate is typically sudanian, characterized by a hot, dry season and a
rainy season, with average rainfall varying between 800 and 1000 mm per
year.
It is an extensive-agriculture region. Rotation between millet and
peanuts is common, as is fallowing. Maize and sorghum are also
important.
Cotton is grown in the east of the region, which is the
beginning of the cotton zone.
Cattle-drawn traction is common in the west of the region, but in
general horse traction predominates.
Livestock husbandry is important,
especially among the Fulani.
The region has relatively less infrastructure than the Central
Peanut Basin.
However, there are fairly dynamic weekly markets along
the Gambian border. There are no large urban centers. Apart from the
4 The total population (rural and urban) of the region was
574,151 persons in 1988, hence a density of 85 persons/km' (BNR 1988).
16
national highway (Dakar-Tamba) that runs through this region (as does
the rail), the road network is quite poor.
The region was once the beneficiary of the New Lands program
(Terres Neuves), now defunct. The program was designed to stimulate and
organ-ize the colonization of the empty lands in the Koumpentoum zone by
colonists coming from the Central Peanut Basin where there are severe
land constraints.
Societe de Developpement des Fibres Textiles (SODEFITEX) supplies
technical services in this zone, as well as a market outlet for maize
and cotton. It also builds roads and helps form Farmer Associations
(ABP). SODEVA is present in the west of the region.
Central Senegal Oriental
The region covers the departments of Tamba (except for the
arrondissement of Koumpentoum) and of Bakel. Its area is 16,748 square
kilometers and it had low population density (five persons/square
kilometers in 1988) and a small population of 82,342 persons in 1988.
There is great ethnic diversity.
Fulani are in the majority, but there
are also the Mande groups (Manding, Diakhanke, Soninke) as well as
migrants from the Peanut Basin (Wolof and Serer).
The region is composed of plateaus cut by valleys with heavy bush
cover. The average rainfall was 770 mm during 1961-85 with a minimum of
460 mm and a maximum of 1,250.
Agriculture is substantially diversified in the region, with cotton
as an important cash crop, and substantial production of maize and
sorghum.
There is little use of agricultural equipment and animal
traction because the Programme Agricole did not operate in this zone.
There is some irrigated rice and banana production. Vegetable gardening
is important in the dry season. Animal husbandry and migration income
are important in the region.
SODEFITEX is the main development institution in the region. Its
activities are focused on cotton and maize, as well as on road-building
in some remote areas.
The region has relatively little infrastructure
(roads, schools, and so on).
Tambacounda is the principal semiurban
center (a secondary city).
Its distance from urban centers in the west
or center makes transport costs an important problem.
REPRESENTIVITY AND CHARACTERISTICS OF STUDY ZONES
Representivitv of Study Zones
As mentioned in the introduction, study zones were selected to
represent each of the agroclimatic zones covered by the IFPRI/ISRA
study.
Table 2.1 shows the relationship between the IFPRI/ISRA study
zones and the agroclimatic regions that they represent.
The table
1 7
Table 2.1 - Damographic charactaristicr of agroclimatic ragions and study zanas
Uorthrn
Central
Southwattarn
Southaastarn
Cantral
Agroclimatic tagfons’
Paenut Basin
Peanut Basin
Peanut Basin
Peanut Basin
Sanagal Oriantal
Rural CommunitieslCommunes
32
131
29
15
9
Rural Villages
1,606
4,492
1,313
713
454
Urban Communes)Quartiers
3124
131202
4147
II?
1115
Area (sq.km.)
10.748
33,058
6,765
6,659
16,748
Population
345,336
2,269,401
573,985
215,264
124.227
Households
35,935
246,599
59,626
20,841
14,401
Population Density
Total (peoplelsq.km)
32
67
85
32
7
Rural only
26
52
59
31
5
Ethnic composition
Wolof, Fulani
Wolof, Serer, Fulani
Wolof. Sww
Wolof, Fulani
Fulani, Diakhanke,SarakolC, Manding
Study zonas covatad
Sagatta
Niakhat
Colobana
Passy
Urban
Oioly
Miasirah
Urban
Number of Villages Covered by
reconnaissance
24
28
15
27
25
16
Study villagas(towna covarad by household
Sagatta (mkt)
Niakhar (mkt)
Colobane (mkt)
Passy (mktl
Kaolack
Dioly Mandakh mkt
NiaoulelTouba
lambs
survey
Khelcom Peulh
Mboltogne
TeckMar
K.Aly Gueye
T . ThieckenelMDioukoul
Medina Alylhfoukouty
Darou Cisse
Mboyene
Khayane
Ndiathiang
Kouthia
ffousahold rurvay raprarantivity’
Rural Community)Communes
29
26
66
26
1
14
8
1
Rural Villages/Ouartiers
1,606
649
3,047
1,313
35
379
454
15
Area (sq.km)
10,746
4,604
14,833
6,765
2,965
16.748
Rural Population
People
275,335
373,312
958,819
399,277
150,861
144,229
82,342
4,1885
Households
28,181
41,103
102,295
38,141
18,638
13,082
8,585
5,816
Percent People in market
_...:
villages’
14.3
20.5
17.1
13.7
15.1
14.9
Approximata rhara of total rural (or urban)
6
8
21
8
5
3
2
1
population taprasantad by study zone
Source:
The data used in the table come from the Secretarial d’Etat 6 la DBcentralisation 1986 (number of rural communilies, villages, and area), from the BNR (population in 1988) end from the village reconnaissance sm-~~eys of the )FPRt/lSRA project
(ethnic composition).
Notes:
’ Data in this section describe the entie agroclimatic region. Parts of these regions are not covered by the IFPRIIISRA study (see section on study zone representivity below)
’ Results from the IFPRIIISRA reconnaissance surveys suggest that each study zone reflects the general pattern of household behavior for the areas and populations indicated in this sub-section of the table.
’ Market village is defined as a village with commercial infrastructure. Villages where the rural community or arrondissement is located fall into this category, as do all villages that host weekly markets.
1 8
illustrates the relative importance of the population and land areas
represented by the sample households in each study zone.
Sampling was not proportional to population, therefore, it is
important to keep in mind that the behavior of households in the zones
with larger populations and land areas will have a greater impact on
aggregate
consumption
and production
behavior than behavior of
households in less populated areas.
For example, households in the
areas represented by the Colobane sample in the central Peanut Basin
account for 21% of the total Senegalese rural population.
While the
study zone in the southeastern basin and that in central Senegal
Oriental account for only 3% and 2% of the rural population,
respectively.
Characteristics of Rural Study Zones
The following description of the rural study zones is based on the
village reconnaissance done before the household survey began. Hence,
the behavioral points (such as importance of nonagricultural income) are
general points made b.y villagers and merely serve as hypotheses to be
tested usinq the more detailed observations from the household survey.
The description of rural zones is organized according to the
following categories: demographics, infrastructure and institutions,
cropping systems, animal husbandry, nonagricultural activities, and food
situation.
In each category, the different rural zones are compared.
Following the discussion of rural zones is a brief description of the
two urban zones that were added in July 1991.
Demographics
Table 2.1 presented above shows the area, population, and
population density of each study zone,
The most densely populated zones
(>50 persons per square kilometer) are found in the western parts of the
central and southern Peanut Basin represented by the Niakhar and Passy
study zones.
The least densely settled area is central Senegal Oriental
(5 persons per square kilometer).
The Wolof constitute the principal ethnic group in four of the five
agroclimatic regions.
They are the predominant ethnic group in all
rural study zones but the zone of the Niakhar (central basin), which is
predominantly Serer, and in central Senegal Oriental where the Fulani
and Mandingue are more highly represented.
The Fulani are the second
largest group in
all zones dominated by the Wolof, except the
southwestern Peanut Basin where the Serer are in second place.
Institutional Presence
Table 2.2 shows the types of public or private organizations that
distribute inputs, purchase outputs, supply credit, or provide extension
services
to farmers
in the study zones.
Three categories of
institutions are examined: (a) parastatals; (b) village sections (VS)'
:
:
:
.,;
.
,‘,
19
Table 2.2 --
Infrastructure and institutional coverage of study zones suggested by reconnaissance survey results.
Northern
Central
Central
Southwest
southeast
Central
Peanut
Basin
Ba.Sin
Peanut Basin
Peanut
Senegal
Basin
Niakhar
Colobane
Basin
OliEittd
Characteristics
n=16
II=15
n=24
Percent of Villages surveyed in reconnaissance having indicated attribute
Institution5 uresent
. SDR/other
79
85
22
44
68
100
agencies
NGOs
37
31
1village groups (VS)
21 46
29 79
87 14
30 78
40
80
Access to credit
.Credit VS/RDS/NGOh
62
29
20
52
52
100
of which VS
1 2
1 8
7
52
8
1 2
FZD9
54
18
7
7
48
100
Marketiue and diet.
u
Store in village
1 7
44
73
30
36
31
.Markt 5 5km from
33
61
47
56
12
31
village
.C?olIectn point
62
82
67
89
76
50
-< 5km from village
Processsine Center
Mill in village
1 7
25
27
30
16
40
Access to education
and health
. School in village
1 2
43
38
48
1 2
50
>50%’ of children
21
68
13
70
4
57
in school
health care in
4
7
27
22
4
37
village
Water SUDDI~
Problem very
75
40
63
7
72
25
important
Source: IPPRI/ISRA village reconnaissance survey (1988)
(4
The data in this table represent the percentage of villages among all the villages surveyed in the zone that declared the presence of the given
characteristic.
(b)
This line gives the percentage of villages in each zone having declared receipt in 1988 of agricultural credit obtained by a Village Group (VS),
a Rural Development.
Society (RDS) or an NGO
This represents the share of villages having declared receipt of credit obtained by VS.
;;
This represents the share of villages having declared receipt of credit obtained by RDS.
Note that “a” is not always equal to “b+c”
because certain villages received credit by VS et RDS at the same time.
(4
This line shows the percentage of villages surveyed in each zone having declared that more than half of the village’s children go to French-
speaking school.
(0
The question asked was “where do the village residents go for primary health care?”
\\
In many villages, there is a health post that is not used by the villagers because of lack of medicines or personnel. We counted these cases
as ones where access was absent.
”
2 0
which are cooperatives that have legal status; and (c ) non-governmental
organizations (NGOs).
Parastatal activity is ubiquitous in central Senegal Oriental
(Missirah).
It consists of activities by SODEFITEX for production of
maize and cotton.
By contrast, in the zones of the Peanut Basin
parastatals are less present (22 to 85 percent of the villages depending
on the zone).
The cowpea program was the reason a large number of
villages were working with SODEVA in the northern and central basin; the
maize and cotton programs were the reason for many villages working with
parastals in the southeastern Peanut Basin.' The relative dearth (due
to gradual withdrawal over the past decade) of parastatals is noteworthy
in the central Peanut Basin (particularly Colobane) and in the southwest
Peanut Basin (Passy).
More than 70 percent of villages in the central
and southwestern basin and in Senegal Oriental participate in VSs.' One
finds a smaller share (40-45 percent) of villages in the northern and
southeastern Peanut Basin with functional VS.
NGO's are uniformly
present in 15-30 percent of villages across all rural study zones.
Access to credit is generally positively correlated with the
presence of parastatal, NGOs and VSs.
Senegal Oriental has the highest
level of institutional presence and highest access to credit while
Colobane in the central basin has low institutional presence and low
access to credit.
The Niakhar zone in the central basin is, however, an anomaly. It
has substantial institutional presence but a low level of declaration of
credit access.
Among villages saying they received credit, the most
important sources are parastatal institutions, followed by VSs.
Infrastructure
Table 2.2 also shows infrastructure in the study zones. We examined
the availability of the following resources: water, education, medical
care, markets,
supply and/or processing of agricultural products and
other consumption goods.7
In interpreting these institutional and
infrastructural findings, keep in mind that half of the population of
Senegal is located within 200 km of Dakar.'
5 The maize programs existed at the beginning of our survey but was
phased out during the survey period.
6 Only the VSs that declared to have undertaken activity for their
members (sale of peanuts, provision of seeds, etc.) during the
previous or then current growing season (1987/88) are counted. In
Senegal Oriental, APBs organized by SODEFITEX are classed as VSs.
7 A discussion of the content of the concept of infrastructure can be
found in Ahmed and Hossain (1988).
' See Enders (1989) for a more detailed discussion of population
'
density and distribution in Senegal.
I
:
:
I
,’
)
*
,.
1.
:
21
Access to water used for household consumption, for animal
husbandry, and for dry-season gardening is a serious problem in several
zones.
The problem is
especially common in
the northern and
southeastern basin where two-thirds of the villages say that access to
water is a very serious problem.g
The problem is somewhat less common
but still important (40-50 percent of villages) in both zones of the
central Peanut Basin.
Long distances and poor roads aggravate the
problem of getting to water sources and transporting water back to the
village.
In villages where access to water poses a problem, a great
deal of labor is expended to transporting water and is, therefore, not
available for alternative activities.
Some villages declared that
children were not sent to school because they had to transport water.
Access to formal education (government or private school taught in
French)" is considered easier when the village has or is close to a
school. If a school is in a village, the attendance rate is higher, as
expected. The survey showed that Senegal Oriental, the southwestern
basin and Niakhar in the central Peanut Basin have the largest share of
villages with schools."
More than half (57 to 70 percent) of the
villages in these three zones claim an attendance rate of more than 50
percent.
By contrast, the southeastern basin zone has the fewest
villages with schools (only one of eight); in almost all of these
villages (96 percent), respondents said that less than 50 percent of the
children go to French-speaking school. Similar results were obtained for
the northern basin and the Colobane zone of the central basin.
Access to primary medical care is a general problem in the study
zones.
Over all study zones, more than two-thirds of the village
respondents declared that the villages do not have primary health
facilities in the village, and villagers have to travel more than 5 km
to get care.
In this situation of general dearth, Senegal Oriental is
relatively well off; a third of its villages said that they had health
units in the village.
However,
our secondary data and the local
authorities gave us a dimmer view of health facilities in Senegal
Oriental. In this matter the study villages may be better equipped than
the average in the zone.
There is a great diversity of access to marketing and processing
infrastructure for agricultural products across zones.
Less than 50
percent of the surveyed villages have a shop, except in the Colobane
zone of the central basin where three-quarters have one or more shops.
' The question posed to the chief and/or notables was whether water
supply was sufficient; if not sufficient, was water a minor, serious,
or very serious problem.
" Another source of formal education that was not included in the
survey is the Koranic schools where Arabic and the Koran are taught.
" To determine the attendance rate for children, the question posed %
was whether school-attending children represent at least half the
children of school age in the village.
22
Fifty to sixty percent of villages in the southwestern Peanut Basin and
the Niakhar zone of the central basin were less than 5 kilometers from
a weekly market.
The high population density in these zones favors the
denser market coverage.
By contrast, the southeastern basin has many
weekly markets, but they are relatively spread out, and most villages
are more than 5 kilometers from a market.
Access to collection points for peanuts is similar to that for
weekly markets.
More than 80 percent of villages interviewed in the
southwest and in the Niakhar zone of the central basin have fewer than
5 kilometers to travel when selling peanuts.
In Senegal Oriental 50
percent of the villages surveyed had to travel more than 5 kilometers.
Mills for processing coarse grains are most common in Senegal
Oriental (40 percent of villages) and least common in the northern and
southeastern Peanut Basin (16-17 percent of villages surveyed).
Agriculture
Crop Mix
The village respondents' views on the structure of their
agricultural production are presented in Table 2.3.12.
In all the
Peanut Basin, millet and peanuts are the main crops. The reconnaissance
survey suggests that millet dominates in the central and southwestern
basin, while peanuts predominate in the northern and southeastern Peanut
Basin.
In Senegal Oriental maize dominates, followed by millet and cotton.
Cowpeas are important as the third ranked crop in the north and
center of the Peanut Basin, mainly because of the promotion campaign by
SODEVA.
Dry-season cropping is only important in Senegal Oriental where
about 60 percent of villages grow dry season crops.
Fewer than 25
percent of villages in
other zones cultivate dry-season plots.
Vegetable gardening is the only dry-season farming in the Peanut Basin.
In Senegal Oriental irrigated rice and bananas are grown in the dry
season, with the help of SODAGRI and OFADEC.
'* To classify crops, we asked each village chief to identify and
categorize the crops grown in the village in order of importance. The
latter was implicitly based on land dedicated to each and the number
of producers growing it.
.
.’
:
23
Table 2.3 --
Relative importance of various crops in 1987 and 1988 as declared by village
representatives during reconnaissance survey.
Northern
C@lWal
central
southwest
Southea.5t
CC3ltKd
Peanut
BaSill
Basin
Peanut
Peanut
Senegal
Ba.SiU
Niakhar
Colobane
BaSill
BaSiIl
OIifXttal
Characteristics
(N=16)
(N=26)
(N=lS)
(N=24)
(N=Zl)
(N=16)
PriucipaI rainfed crop
peanuts
71
47
22
68
6
millet
29
100
53
78
28
31
sorghum
-
-
cowpeas
maize
4
38
cotton
25
rice
TOTAL
100
100
100
100
100
Secondary rainfed crop
peanuts
21
100
63
78
24
19
millet
67
40
22
64
6
sorghum
7
8
31
cowpeas
12
maize
25
cotton
4
19
rice
-
none
TOTAL
100
100
100
100
100
100
Tertiary rainfed crop
peanuts
4
8
6
millet
-
7
4
sorghum
4
43
33
18
36
25
cowpeas
88
36
63
4
maize
7
52
44
19
cotton
-
4
8
60
rice
11
none
4
11
other
21
TOTAL
100
100
100
100
100
100
Principal dry-season crop
vegetables
7
7
26
8
19
rice
12
banana
12
sorrel
-
none
92
93
93
74
84
67
other
4
-
8
TOTAL
100
100
100
100
100
100
Source:
IFPRI/ISRA village reconnaissance survey (1988).
Remark:
The interpretation of the table is as follows: In the Northern Peanut Basin, 71% of the
villages surveyed declared peanuts as the most important crop and 29% placed millet in
first place. In the same zone 67% ranked millet in second place while 21% placed
peanuts second and 12% placed cowpeas second. Third place is dominated by cowpeas
(88% of the villages).
.*
,
:
.
.:.
.
.
*
‘.
.
.
.
.
24
Crop Production Constraints
Table 2.4 shows the constraints to crop production declared by the
village representatives.13
Land constraints are considered most important in the Niakhar zone
of the central basin (about half of the villages). The next most severe
case is Colobane, still in the central Peanut Basin. These zones are
among the most densely populated. In the other zones, the constraint was
seldom mentioned (only one in 10 declared a problem). These two land-
constrained zones plus the southwestern Peanut Basin had the smallest
share of villages (15-30 percent) claiming to fallow land regularly.
The low prevalence of fallowing in these zones, supports the villagers
claim that a land constraint exists.
Other zones had 50 to 75 percent
of villages claiming to use fallows regularly.
It is surprising how few villages in all zones declared a labor
constraint, despite the size of the migration phenomenon in the Peanut
Basin and in Senegal Oriental. In fact, fewer than 10 percent of the
villages said they were seriously affected by labor constraints. One
possible reason is that the labor that migrated is surplus labor, with
marginal productivity lower than the average productivity, and so these
laborers can leave the rural areas without reducing output. Another
possible explanation is that much of the migration takes place in the
dry season and does not conflict with cropping activities.
A third
hypothesis is that the principal bottleneck is at harvest time when
temporary laborers from Guinea are available.
An agricultural equipment constraint is declared by 10 to 20
percent of villages in all zones but the southeastern Peanut Basin where
the constraint is considered important in 40 percent of villages. The
"Programme Agricole",
which provided equipment credit throughout the
1960-80 period, was not as well developed in the southeastern basin.
Villages in this zone are also not extremely active in the SODEFITEX
cotton program, which is the only other major source of credit for
annual traction equipment.
Horse-drawn traction is declared to be the most widely used
form
of traction, except in Senegal Oriental where donkeys are more common.
Oxen traction predominates in about 15 percent of villages surveyed in
the southern Peanut Basin zones and in Senegal Oriental.
Animal Husbandry and Off-Farm Activities
Table 2.5 provides an extremely simplified picture of the relative
importance of income sources for men and women as revealed in the
l3 Here the question was whether a given constraint exists; if it
does, was it of medium severity or great severity? Despite the "moral
hazard" of this type of question,
it provided broadly indicative
I
information and grist for hypotheses to be tested with the household
survey data.
25
Table 2.4 --
Agricultural constraints, fallow, and animal traction
patterns suggested by the village reconnaissance survey.
Northern
Central
Central
Southwest
Southeast
Central
Peanut
Basin
Basin
Peanut
Peanut
Senegal
Basin
Niakhar
Colobane
Basin
Basin
Oriental
Characteristics
N=16)
(~=26)
(N=15)
(N=24)
(N=Zl)
(~=i6)
Animal traction type:
(Percent of villages declaring category as principal form of traction)
principal'
ox
horse
donkey
18
16
19
96
100
87
82
84
37
TOTAL
4
13
44
100
100
100
100
100
100
Animal traction type:
(Percent of villages declaring category as second most important form of
traction)
secondaryb
Zrse
4
13
11
76
25
donkey
13
1 9
16
31
none
aa
100
67
70
8
44
a
7
TOTAL
100
100
100
100
1 0 0
100
Fallow
76
28
17
ia
50
6 9
(Percent of villages claiming to use fallows to restore soil fertility)
Constraints:
(Percent of villages declaring the constraint to be important)
. Land
a
46
21
4
1 2
12
Equipment
17
20
40
12
Labor
8
4
12
Source:
IFPRI/ISRA village reconnaissance survey (1966)
a)
The question asked was "what is the principal type of animal traction used in the village?"
b)
The question asked was "what is the second most used type of animal traction?"
2 6
Table 2.5 --
Relative importance of animal husbandry and other
off-farm
activities as
suggested by
the
reconnaissance survey
E
Northern
Central
Central
Southwest
Southeastern
Central
Peanut
Basin
Basin
Peanut
Peanut
Senegal
Basin
Niakhar
Colobane
Basin
Basin
Oriental
E
Characteristics
(N=16)
(~=26)
(N=15)
(N=24)
(N=21)
(N=I~)
Sources of Incane'
iPercent of villaqes Declarinq Activity as Principal, Secondary, or
Tertiary)
E
For Hen
Principal activity
crop
crop
Crop prod
crop prod
crop prod
crop
% villages
prod
prod
100
100
100
prod
E
Secndry activity
92
100
hsbndry
hsbndry
hsbndry
9 4
% villages
hsbndry
hsbndry
67
89
52
mig/ext
Tert. activity
62
8 9
Migrat.
mig/Sen
comnerc
2 5
% villages
comnerc
Mig/Sen
33
41
12
hsbndry
E
42
65
2 5
For Women
E
principal activity
crop
crop
crop prod
crop prod
crop prod
crop
% villages
prod
prod
87
100
100
prod
secndry activity
a3
100
commerce
hsbndry
hsbndry
7 5
% villages
hsbndry
hsbndry
27
8 9
4 0
hsbndry
tert activity
4 2
54
hsbndry
conrnerc
other
19
E
% villages
commerce
mig/Sen
2 0
26
8
comerc
8
3 5
12
E
Husbandry
(Percent of villaqes where husbandry is declared
important)
E
. cattle herding
4
25
27
56
24
50
sheep herding
29
68
73
9 3
8 0
9 4
. head herding
29
71
6 0
9 3
7 2
9 4
Cattle fattening
4
11
7
11
0
12
E
sheep fattening
4 6
18
15
26
3 6
--
17
Source:
IFPRI/ISRA village reconnaissance survey (1988).
The reconnaissance results presented here provide a very general picture of zone tendencies.
Detailed
i
enumeration of income for every member of the household study suggests that the actual composition of income
is somewhat different than the picture obtained from the reconnaissance data.
See Chapter 4 for household
survey results.
a)
The declared income sources for men and for women are classed in order of importance.
The figures below each activity title show the proportion of villages surveyed in each zone mentioning the
activity in a given rank.
Example: In the northern Peanut Basin,
92% of the villages surveyed declared that crop prod agriculture is
the principal source of income for men in their village; hence for 8% of the villages the principal source
of income was declared not to be agriculture.
As the second scurce ofincome 62% cited animal husbandry and as the third source of income 42% declared
commerce.
"Mig/Sen" means migration inside Senegal while
"Mis/ext" means migration outside Senegal; "Migrat"
includes both Senegalese and foreign migration.
27
reconnaissance data.
Analysis of household survey data (Chapter 4)
shows that these general perceptions of income sources are not always
confirmed by detailed enumeration of income earned by each household
member. Nevertheless, the reconnaissance data did permit us to identify
zones
where off farm incomes from migration and commerce were
potentially important.
The table 2.5 shows that in all the zones, for men and for women,
rainfed cropping is declared to be the main source of income. (As shown
in Chapter 4, this contrasts with household survey findings in some
zones.) The table also shows the importance of animal husbandry and
fattening activities in each zone.
Cattle raising was declared important in more than 50 percent of
villages in the southwestern basin and in Senegal Oriental, but in fewer
than 25 percent of villages in other zones.
Small ruminants were
important in more than 70 percent of villages in all zones but the
northern basin.
Animal fattening (primarily of small ruminants) was
most common in the northern and southern Peanut Basin.
For men, sales of animals and animal products were declared to be
the main noncropping income sources throughout the Peanut Basin. In
second place is petty commerce (especially in the zones of Sagatta and
Dioly) and migration.
Migration is declared more important than
livestock income in Senegal Oriental.
Migration is also declared an
important income source in the central and southwest basin.
For women, animal husbandry was declared the main noncropping cash
generator in 5 of 6 zones.
Only in the central Peanut Basin zone of
Colobane was petty commerce declared to be more important than animal
husbandry.
Women's migration was a commonly declared source of income
in the central basin zone of Niakhar.
Food Situation
Table 2.6 shows the declarations of the village representatives
concerning the village's general food situation in 1987 and 1988. The
table also shows the relative importance of methods used to redress
deficits.
The data indicate how many months at least half of the
households in the village could feed themselves based only on their 1986
and 1987 harvests.14
This is a rough measure of self-sufficiency.
Three ranks were created based on the declared number of self-
sufficient months: (1) self-sufficient or not very deficient households
can eat from own production for at least nine months of the year; (2)
moderately deficient, from five to eight months; and (3) very deficient,
fewer than five months.
Analysis of the results led to the following
conclusions.
l4 Again, we stress that these results are only broadly indicative and
serve as hypotheses for the household survey data.
-*-
28'
Table 2.6 --
Food security and the relative importance
of different coping strategies (to
compensate food production shortfalls)
suggested by the reconnaissance survey.
Northern
Central
Central
Southwest
Southeast
Central
Peanut
Basin
Basin
Peanut
Peanut
Senegal
Basin
Basin
Basin
Oriental
Niakhar
Colobane
Characteristics'
(~=16)
(~=26)
(N=15)
(N=24)
(N=Zl)
(N=I~)
(Percent of villages surveyed falling into each category)
1987 Harvest
2 9 months
5-6 months
8
68
40
81
50
25
I4 months
21
28
60
15
46
75
71
4
4
4
1988 Havest
19 months
5-8 months
< 4 months
8
71
27
81
38
25
25
25
73
19
54
75
67
4
8
Means to cover deficit
(Percent of villages declaring activity as principal, secondary or tertiary means)
. print. means
hsbndry
hsbndry
hsbndry
hsbndry
hsbndry
hsbndry
67
73
73
67
89
81
. second. means
migrat
migrat
credit
migrat
misc
comnerc
37
11
40
30
26
31
. tert. means
conxnerc
comnerc
comnerc
off-farm
29
7
16
8
Source:
IFPRI/ISRA village reconnaissance survey (1988).
a)
The indicator for degree of self-sufficiency used is the "number of months covered by
the cereal harvest for more than half of the households in the village.
29
The southwest basin and Niakhar in the central basin have the most
villages declaring themselves self-sufficient (81 and 68 percent,
respectively).
In these zones, one notes a large share of villages
ranking millet as their first crop.
These are also zones with a large
share of Serer households.
The food (production) situation is poorest in the northern Peanut
Basin where,
70 percent of the villages declared themselves very
deficient in 1987.
Almost all the other zones were moderately
deficient, with the southeast having about the same number of villages
each in the first and second categories.
A comparison of the 1986 and
1987 declarations indicates that the food (production) situation was
more or less stable. Some downturn was declared in the northern and
southwestern Peanut Basin zones.
Most villages declared that the first means that they used to buy
cereal to redress their deficits was animal sales (especially sheep and
goats).
Sales of poultry and milk were also mentioned frequently. In
second place was the use of migration income, especially in the northern
basin,
in Niakhar,
and in Senegal Oriental.
Earnings from petty
commerce were the second source in Senegal Oriental and the third in the
northern and southern basins.
Consumption credit (from merchants or
family) was important in Colobane (second after animal sales).
Characteristics of Urban Study Zones
Demography
Kaolack, founded in 1853, is located in the heart of the Peanut
Basin, approximately 200 km southeast of Dakar. Forty-eight percent of
Kaolack's 151,000 inhabitants are men and 42% are women, divided into
14,600 concessions and 18,600 households (BNR 1988). Kaolack is also
the capital of the region and is comprised of 21 districts and 5
villages, officially incorporated into the town. The predominant ethnic
groups are the Serer.
Tamba is situated 500 km east of Dakar and is the most important
city in the eastern part of the country.
It is comprised of six
official districts, plus there are six affiliated villages which make up
the commune of Tamba, the administrative capital of the region. It has
a population of 41,900 (11% of the total population of the region) of
which 51% are men and 49% are women.
The population forms 4330
concessions and 5820 households (BNR 1988). The ethnic composition is
diverse,
with Mandh being the predominant group (Manding, Bambara,
etc...), followed by the Fulani group residing in upper Casamance and
Guinee, and finally the Wolof group from the Peanut Basin.
Demographic evolution of the two zones
1955
1965
1976
1988
Kaolack
46600
81600
135473
150961
Tamba
4600
12700
25780
41885
Source: Atlas JA (1983) et BNR (1988).
3 0
Kaolack was three times more populated than Tamba in 1988, while
three decades before, the proportion was ten times greater.
Infrastructure and Economy
The following table presents information which is available for the
two cities in terms of health, education, communication, and leisure
infrastructure.
Note that the two cities are relatively well endowed with schools,
especially at the elementary level.
Access to water for household
consumption appears to be adequate.
For those who do not subscribe to
SONEES (public utility company providing water) have access to public
faucets where one must pay.
At this level, the problem lies in the
number of public faucets and the establishment of a price scale for
water for the local populations.
Almost all of the districts are
covered by one or more paved roads.
Electric power is available to
households near main roads; but in reality equipment is often in a state
of disrepair.
Also, in half of the districts of the two cities, there is either
a small clinic or a medical center.
However, due to dilapidation or
lack of material and/or medication, which even affects the hospitals of
these two cities, the efficient operation of these infrastructures is
diminished.
Infrastructural Capacity in the two Zones
Kaolack
Tamba
Percentage of neighborhoods
having:- elementary school
81
8 3
- junior high school
3 8
3 3
- public water faucet
95
8 3
- clinic
4 7
5 0
- paved roads
57
6 6
- roads with electricity 95
- markets
33
ii;
- mosque/church
100
100
- cinema
14
17
Does the city have:
- a harbor?
yes
II0
- an airport?
yes
yes
- a hotel?
two?
two
- a hospital?
yes
yes
- a high school?
two?
one
Source: IFPRI/ISRA urban survey data (1990).
31
Each district of the two cities has a religious center, generally
a mosque. Kaolack has three cinemas, whereas Tamba only has one for the
entire city.
Caution must be used when evaluating data for markets. In looking
at the table, it would seem that Tamba is better off than Kaolack.
However, Kaolack,
being geographically larger and more populated,
actually has more markets than Tamba, which apart from the central
market only has smaller, less lively markets. Kaolack is known for its
commercial nature; the strength of this sector is due to the peanut
trade and to the proximity to The Gambia.
For the other infrastructure, Kaolack is better endowed than Tamba,
which has been affected by its isolated position.
Both cities have an
airport, but Kaolack, situated on the ocean, has a port (primarily used
for transporting salt products).
The telephone system is now automatic
in the two urban centers and each has regional broadcasting station.
There is much more industrial activity in Kaolack: processing of
peanuts,
salt harvesting and processing, textiles, etc.
In Tamba,
ginning of cotton by SODEFITEX remains the main industrial activity of
the city.
There is also a banking sector in the two cities where the
regional agencies of the principal banks of the country are located.
SUMMARY OF MAIN POINTS
Results presented in this chapter come from the village
reconnaissance
survey
conducted prior
to the household survey.
Reconnaissance data provides a broad picture of how respondents (village
chefs and elders) perceive the physical and socio-economic environment
in which they live.
Reconnaissance data were used to select
representative study zones and villages.
The most important points made in the chapter are:
a The IFPRI/ISRA sample is considered representative of
approximately 50 percent of the Senegalese rural population and 6
percent of the urban population.
o Fifteen to 20 percent of the rural population in the study zones
live in villages with commercial infrastructure (that is, market
villages and seats of the arrondissement or rural community).
a Access to markets is easiest in the central and southwestern
Peanut Basin (50 percent of villages less than 5 kilometers from a
market) and most difficult in the southeastern basin (only 12
percent of villages within 5 kilometers).
l Crop production is considered the principal source of income for
greater than 95 percent of villages in all zones.
.
32
l Cropping productivity is most hindered by land constraints in the
central Peanut Basin and by equipment constraints in
the
southeastern Peanut Basin and the Colobane zone of the central
basin.
l Labor was rarely mentioned as a constraint, but most cases were
in the zones of high migration (northern basin and Senegal
Oriental).
l Adequacy of cereal production
for household needs in the two
years prior to the IFPRI/ISRA study was best in the Niakhar zone of
the central basin and in the southwestern basin; it was poorest in
the northern basin and Senegal Oriental.
l Income from animal husbandry activities was considered the most
frequently used source of money to cover cereal deficits across all
zones.
l The highest rates of school attendance appear to be in the
Niakhar zone of the central basin and the southwestern Peanut
Basin.
l The lowest rates of school attendance appear to be in the
southeastern Peanut Basin and the Colobane zone of the central
basin.
l Access to water is a serious problem in 70 percent of villages in
the northern and southeastern Peanut Basin and in 40 to 50 percent
of villages in the central basin.
l Institutions providing formal credit are most commonly found in
Senegal Oriental and least common in the central Peanut Basin.
The extent to which the reconnaissance survey results are
consistent with data obtained at the household level is examined in the
next chapter.
33
REFERENCES
Bureau National du Recensement.
1988.
Donn6es preliminaires du
recensement demoqraphique.
September.
Jeune Afrique.
1983.
Atlas du Seneqal.
Serie de 1'Atl as Jeune
Afrique Editions, Second edition. Paris.
Le Borgne, Jean. 1988.
La pluviometrie au Seneqal et dans la Gambie.
Ministere Fransais de la Cooperation et ORSTOM. April. Paris
..;
.’
3.
SAMPLE
CHARACTERISTICS
INTRODUCTION
This chapter describes sample and household characteristics for
the six rural and two urban study zones covered by the report.
The
analysis is drawn from the full set of households that participated in
the study during the first year of data collection for each zone --
those interviewed from the outset as well as those brought in as
substitutes for households that dropped out.'
Both sample and household characteristics are generally based on
the situation prevailing when interviews began in each zone (September-
October 1988 for the rural sample and June 1990 for the urban sample).
The only exception is that the number of adult equivalents reported is
the average number present in sample households during the entire survey
period. This was done to adjust for migration and demographic changes.
First the demographic characteristics of the sample are described:
size,
age
and gender distributions,
and ethnic composition. A
description of the typical household in terms of size and structure
follows,
The next section describes socioeconomic characteristics of
the typical sample household -- age, gender, and educational profile of
the household head; children's education, religion, economic activities
(both local and migratory), rural assets, urban living standards, and
urban
settlement patterns.
The final section points out the
implications of selected findings for agricultural price policy.
As might be expected, a sample covering as many agroclimatic
regions and sociocultural situations as this sample does is bound to
exhibit a great deal of diversity among households within each zone as
well as across different zones.
The discussion in this chapter
generalizes rather than reporting the details of this diversity -- a
"typical" household is described.
In doing this, we relied primarily on
median rather than mean values for households within each zone. In
generalizing about the "typical" household across zones, we report the
modal value of the individual zone medians, pointing out zones where
characteristics differ substantially from the 'typical' picture being
presented.
' Sample households were selected randomly from a list of households
developed by project personnel.
The number of households originally s
selected was larger than the required sample size, thereby providing a
number of households available as substitutes for dropouts.
35
SAMPLE DEMOGRAPHICS
The study sample analyzed in this chapter includes 296 households
(226 rural and 70 urban) which contain 3,452 people or 2,636 adult
equivalents (AE).
Each of the six rural and two urban zones is
represented by 35-40 households (see Tables 3.1 and 3.2).
Aqe and Gender Distribution
The average age of the sample population is about 20 years while
the median is 15 (Table 3.1 and Table 3.2). Forty-five to fifty percent
of the population of each zone is considered "active" (between the ages
of 15 and 59, inclusive).*
Forty-five to forty-seven percent of the
population is under 15 while 4-7 percent is 60 years or more. Males
(51-52 percent) outnumber females slightly in most zones. There is no
difference in age or gender distribution between the urban and rural
zones.
The age/sex pyramid for the sample population was compared with
the theoretical population structure for each rural zone using 1986 data
(Table 3.3).3
This type of comparison allows one to verify whether the
population structure of the IFPRI/ISRA sample is similar to the
structure reported by the national statistical service.
There was only
one statistically significant difference between the sample and the
theoretical population.
The difference occurs in the southwestern
Peanut Basin (Passy zone) where only 40 percent (rather than 45-50
percent) of the sample falls in the "active" category.
Ethnic Composition
The ethnic composition of the sample reflects what one would
expect in each zone (see Chapter 2 for a detailed discussion of zone
characteristics) so details are not repeated here. Table 3.1 shows the
percent of the major ethnic groups represented in each zone.
HOUSEHOLD DEMOGRAPHICS
The typical rural household consists of 11 people or 8 adult
equivalents (Table 3.4).
This typical household is a synthesis of
households ranging in size from 2 to 43 people (2 to 36 adult
* We use the ISRA definition of "active" which includes all persons
215 or t60 years of age.
3 The data for the 1986 theoretical population is based on
information contained in Situation Economique 1986, issued by the
Direction de la Statistique, Senegal.
..
.
.:
36
Table 3.1 --
Characteristics of the rural sample by zone
Central Peanut Basin
Northern
Southwest
Southeast
Central
Peanut
Peanut
Peanut
Senegal
Characteristic
Basin
Niakhar Colobane
Basin
Basin
Oriental
Number of
40
37
37
38
36
38
households
Number of
519
378
438
437
469
431
people
Number of AE
394
290
346
325
350
331
Average age
2 0
21
2 2
2 0
19
2 1
Median age
1 5
1 5
1 8
12
15
1 6
Percent
Active'
46
4 5
5 2
40
48
49
< 15
4 7
47
4 2
54
48
47
2 60
7
8
6
6
4
4
Percent male
4 9
52
5 1
5 1
5 2
5 1
Percent female
5 1
48
49
49
48
49
E-t:~~~~ol;mp,~~~~~~~
.:.:.:; . . >:.:.::::.:::.:..,.:.: . . . . . . . . . . ..A.....
Dominant group
Wolof
Serer
Wolof
Serer
Wolof
Diakhanke
Percent
(63)
(100)
(77)
(47)
(86)
(45)
Second group
Fulani
Serer
Wolof
Fulani
Fulani
Percent
(31)
--
(19)
(33)
(13)
(38)
Third group
Toucouleur
Tout.
Fulani
Serer
Other
Percent
( 2)
--
( 3)
( 5)
( 3)
(17)
Source: ISRA/IFPRI project survey.
Notes:
Information presented in this table describe the sample at the beginning of the survey; changes
in household size and composition are taken into consideration in the analyses but not reported
here.
a)
We use the ISRA definition of active worker which is 115 years and
~60
37
Table 3.2 --
Characteristics of the urban sample by zone
Characteristics
Kaolack
Tamba
No. of households
3 5
3 5
Total population of the households
441
339
% active population
(>=15 ans et ~60 ans)
4 9
50
% population less than 15 years
4 6
4 6
% population 60 or more
years
5
4
% males
4 7
53
% females
53
47
No. of adult equivalents
340
260
Average age (years)
21
2 0
Median age
16
16
p~ijr"i"li":i;ic.ii~~:t-li;ic:i'i~~."'$'i~~bn::::I-~r
Bambaraf
Zone
Yol of
S&&e
Peul
Toucoul cur
Handi ngue
D i a k h a n k e
Sarakol e
Kaolack 51 18
9
15
5
2
Tamba 17 1
4 5
3
16
2
14
Source:
ISRA/IFPRI project survey (1990 - 91).
:
;.
.
.
. .
38
Table 3.3 --
Age and gender breakdowns of sample compared with theoretical population
Percent by age group
ZONE
Age
Percent
Percent
Group
Males
Females
Sample
Theoretical
Northern
< 15
24.3
25.2
49.5
46.6
Peanut Basin
15-54
21.1
22.4
43.5
46.3
>55
3.3
3.6
6.9
7.1
Percent
Sample
48.7
51.2
100.0
100.0
by gender
Theoretical
4 9
51
100.0
100.0
Central Basin
NIAKHAR
< 15
26.8
22.4
49.2
46.5
15-54
20.6
22.2
42.8
46.3
>55
4.3
3.7
a.0
7.1
Percent
Sample
51.7
48.3
100.0
100.0
by gender:
Theoretical
49
5 1
100
COLOEANE
< 15
22.6
19.4
42.0
46.5
15-54
24.2
25.6
49.8
46.3
>55
4.1
4.1
a.2
7.1
Percent
Sample
50.9
49.1
100.0
100.0
by gender:
Theoretical
49
5 1
100
Southwestern
< 15
28.3
25.9
54.2
46.5
Peanut Basin
15-54
17.3
20.0
37.3
46.3
>55
5.0
3.5
8.6
7.1
Percent
Sample
50.7
49.3
100.0
100.0
by gender:
Theoretical
4 9
51
100
Southeastern
cl5
25.4
2 3
48.4
46.5
Peanut Basin
15-54
23.9
21.9
45.0
46.3
>55
2.8
3
5.3
7.1
Percent
Sample
52.1
47.9
100
by gender:
Theoretical
50.2
49.8
Central
< 15
23.8
22.9
46.7
46.9
Senegal
15-54
24.0
22.2
46.2
47.2
Oriental
>55
3.1
4.0
7.1
5.9
Percent
Sample
50.9
49.1
100.0
100.0
by gender:
Theoretical
49.8
50.2
100
Source: Compiled from ISRA/IFPRI project survey and Situation Economique, 1986, Direction de
la Statistique, Senegal.
,
:.
,’
:
.,
.
.’
39
Table 3.4 --
Household characteristics for the rural zones
Central Peanut Basin
Northern
Southwest
Southeast
Central
Household
Peanut
Niakhar Colobane
Peanut
Peanut
Senegal
Characteristic
Basin
Basin
Basin
Oriental
Number of people
Average
13
10
12
12
13
11
Median
12
10
11
11
11
11
Nunber of AE
Average
10
a
9
9
10
9
Median
9
a
a
a
a
9
Composition'
Percent
Nuclear
61
60
60
61
54
56
Extended
38
40
36
39
31
39
Other
1
--
4
v-h
15
5
Education
Percent of HH with
children attending
French school
n.a.
68
26
71
0
52
Migration
Percent
Ory Season
All migrants
13
10
13
4
5
7
Navetane as percent
of all migrants
5
0
25
m-b
aa
38
Rainy Season
4
5
1
1
2
7
Source:
ISRA/IFPRI project survey.
Notes:
Information presented in this table describes the sample at the beginning of the survey; changes in
household size and composition are taken into consideration in the analyses but not reported here.
1
Nuclear family includes: household head, wives, children.
Extended family includes others related to the
household head or his wives (siblings, parents, grandparents, aunts, uncles, nieces, nephews,
grandchildren).
Others includes all not falling in the above categories; they are primarily navetane
(agricultural laborers), apprentices, and Koranic students.
b
This zone does have navetanes, however, none were present at the beginning of the survey so they were not
enumerated until they returned to the sample households in 1969.
Approximately 20 percent of unrelated
household members were navetanes in 1989.
..’
_,
.
:
*
.
.
*.
.
i
_’
4 0
equivalents).
Households in the Serer part of the central Peanut Basin
(Niakhar) tend to be somewhat smaller than households in other zones.
Urban households are smaller than their rural counterparts (Table
3.5); 8-9 people and 6-8 adult equivalents. Households in Tambacounda
are somewhat smaller than those in Kaolack.
Structure
Sixty to seventy percent of the people in the typical rural
household are members of the nuclear family (household head, wives, and
unmarried children) (Table 3.4).
Thirty to forty percent are members of
the extended family (relatives of the household head or his wives).
Five to ten percent of the people in the typical household are not
related to the household head.
In this latter category we find most
frequently navetanes (agricultural laborers), apprentices, and Koranic
students.
This type of household structure is fairly uniform across all
rural study zones.
In the Southeastern Peanut Basin and Senegal
Oriental, however, the nuclear family accounts for a smaller share of
the household population (54 and 56 percent).
In the southeast the
difference is taken up by the "other" category which is 88 percent
agricultural laborers.
In Senegal Oriental, the share of extended
family is larger than in other zones.
The Tambacounda urban sample exhibits the typical
rural
composition described above (Table 3.5).
The Kaolack urban sample,
however, has the smallest share of nuclear family (47 percent) and the
largest share of extended family (53 percent) found in any sample zone.
SOCIOECONOMIC CHARACTERISTICS OF THE HOUSEHOLD
Profile of the Household Head
The typical household head is a male in his late forties or early
fifties.4
More than 95 percent are married; 50-55 percent have one wife
and 30-40 percent have two wives.
Cases of three and four wives are
rare; these cases are more common in the southeastern Peanut Basin where
about 40 percent have one wife, 30 percent have two, and 20 percent have
three (Table 3.6).
Monogamy is more common in the Serer zone of the central Peanut
Basin (Niakhar study zone) where slightly more than 70 percent of
household heads have only one wife.
Formal education in the French system is rare among household
heads;
only 5 of the 226 rural household heads have attended school,
4 In rural zones less than 1 percent of household heads are female.
.
.
:
41
Table 3.5 --
Household characteristics for the urban zones
Characteristics
Kaolack
Tamba
s&$&#g&l$E~@;:( kii.&&)
‘:i.:...... ..,.:. I,. . . . . . :,.,........., . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . : ,.,.,...,.i,.,..,.,
.
Average population at start of study
12.0
10.0
Median population at end of study
9.0
8.0
Average AE at start of study
9.7
7.4
Median AE at end of study
7.5
5.7
Average annual population
12.0
9.0
Median annual population
6.0
8.0
Average annual AE
9.4
7.0
Median AE
6.5
5.9
~~~iSirjii’,oFi.r~~~&~:~PXt
..,..,.........,...,.,.,,..,
.,.. :..:.:.:,:
. . . . . :. ..;,.,,_,.
.,... .. ,.,
,.>>,.::.:..::.:::
,,........
“‘.....’
Nuclear family
46.5
6 1
Extended family
53.0
36.0
Others (not clear)
0.5
3.0
EnoeAri~.:i:fi’E~~~~~~~~E~,T
. . . . . . . . . ..I.. .
..,.. ..‘.,.. .......,, ,, .,,.:. . . . . . . . . .., .,.;. .., .,,.,
% of total household population who have:
- attended French school
23.0
30.0
- received formal vocational or technical
10.0
1.0
training
- served a traditional apprenticeship
6.0
9.0
- a current source of income
17.0
19.0
);Il&pJzf~
~qJ&jz ~‘:~&&~~~~@j~~
:.... : :.:. :. :::.::..:.:::.:.:..
: ..:.:.+: : ;:. .:.. . . .,.,
Dry season
0.7
1.0
Rainy season
4.0
4.0
Source:
IFPRI/ISRA project survey (1990-91).
42
Table 3.6 --
Characteristics of household heads for rural zones
Central Peanut Basin
Northern
Southwest
Southeast
Household
Peanut
Niakhar
Co1 obane
Peanut
Peanut
Characteristic
Basin
Basin
Basin
Oriental
&++g&.&
.:..
‘A6’e;as’e
.;..,.
47
53
49
51
--
48
Median
45
51
51
52
40
50
h&r
.:..:“‘Phrce”t ma, e
100
97
97
100
100
100
Percent female
0
3
3
0
0
0
..,
&j-&l i::.&$$.js
., . .:: . . . .’ .:.:..: . . . . . .,
Percent
Unmarried
0
a
5
3
0
1 0
One Wife
50
69
47
64
35
59
Two Wives
31
15
37
28
38
21
Three Wives
17
a
5
5
24
10
Four Wives
2
0
5
0
3
0
Number with
formal schooling
~6 years
n.a.
0
1
0
0
1
6-9 years
n.a.
1
0
1
0
0
9-12 years
n.a.
0
0
0
0
1
Number with
literacy training
n.a.
0
3
0
0
3
Source: ISRA/IFPRI project survey.
Notes:
Information presented in this table describes the sample at the beginning of the
survey; changes in household size and composition are taken into consideration in
the analyses but not reported here. n.a. means not available.
43
most for 6 or fewer years.
Literacy programs in local languages have
been no more popular;
only 6 household heads have attended classes.
There are some distinct differences between "typical" rural and
"typical" urban household heads.
Urban zones have a much larger share of female household
heads (26 percent in Kaolack and 14 percent in Tamba) than
rural zones (<l percent) (See Table 3.7).
About 30% of urban households are headed by unmarried
persons (t5 percent in rural zones).
Urban household heads have a greater exposure to formal
education than their rural counterparts, with 20-30 percent
having attended school at some time. Another 20-30 percent
have had some technical or apprenticeship training. This
leaves, however, about 60 percent with no type of education
or formal job training.
Profile of the Household
The rate of formal education among children is generally higher
than that of household heads; however, great disparity exists across
zones.
In two study zones (the Niakhar zone of the central Peanut Basin
and the southwestern Peanut Basin) about 70 percent of households have
at least one child attending school.
Thirty to thirty-five percent of
households are sending children to formal schools in Senegal Oriental
and the Colobane zone of the central Peanut Basin,
In the southeastern
Peanut Basin (Dioly) not a single child in the entire sample is
attending school.
Disaffection for formal education is typical of this
zone and is not a sampling anomaly.
In the urban sample, 20-30 percent of the total population has
some formal schooling (the rate is higher in Tambacounda).
About 40
percent of the sample in each urban zone has had some type of formal
schooling, apprenticeship, or formal job training.
Reliqion
Households are predominantly Muslim in both rural and urban
samples.
Only in the Niakhar zone of the Central Peanut Basin do we
find some Catholic households.
The presence of Catholics in this zone
undoubtedly contributes to the higher rate of monogamy noted above.
Educational infrastructure provided by the Catholic church in the zone
helps to encourage the higher rate of school attendance found in this
area.
.
. .
44
Local Economic Activities
Almost all rural households declared crop production as their
major source of income.
As Chapter 4 reveals, this was not found to be
true in all zones and all years covered by the ISRA/IFPRI survey.
Table 3.7 -- Characteristics of household heads in the urban sample.
Average (years)
52
4 6
Median (years)
51
42
$*:+,::.
% Male
74
86
% Female
26
14
.,
~~yAtp;ciii’$~j$~&:Qi(,X]:::
. . ,.,: . . . . :.:.: . ,.:.:.:.>,.: :.,.:.:.:.:.:.:.:.,.:.: ,,._.,.,., ,.
% Unmarried
31
28
% Married with 1 wife
4 9
54
2 wives
11
11
3 wives
6
3
4 wives
3
3
.,
,. . . .
EnucAT~a#AL.::.sErr~~~,~~~,
: ,.,., ,.:... ,. .):.:.~,.:., ,.. :.: ,., ..:.:.: :,,
- Formal education
Primary school
9
12
Secondary school
3
6
University
6
12
Other
82
70
- Formal and formal vocational or professional training
3
6
- Formal instruction and traditional apprenticeship
9
- Traditional apprenticeship only
24
12
- Formal instruction only
15
14
- None of the above
58
59
Source:
IFPRI/ISRA project survey (1990 - 91).
4 5
Most rural households have several other sources of income in
addition to their crop production; the importance of these activities
varies across households and across zones.
Very few urban households draw income from cropping and livestock
activities.
The most common activities are in the service and commerce
sectors.
Only 17 percent of the Kaolack and 19 percent of the Tambacounda
population declared that they were earning some type of income when the
survey began; this means that just 35-40 percent of the "active"
population considered themselves to be employed at that time.
Miqration
Migration is a phenomenon of varying importance across zones.
The percent of household members 115 years of age leaving the village
during the dry season ranges from 4 to 13 percent. The highest rates
are found in the northern Peanut Basin and the Colobane zone of the
central Peanut Basin.
The lowest rates are in the southwestern and
southeastern Peanut Basin.
Outmigration during the dry season in three zones includes a large
number of 'navetanesn
returning to their own homes (25 percent of
outmigrants in the Colobane zone of the central Peanut Basin, 38 percent
in Senegal Oriental, and 88 percent in the southeastern Peanut Basin).
The remainder of the outmigrants are predominantly males, both married
and single; many are household heads.
In terms of gender, the Niakhar zone of the central Peanut Basin
presents an exception to the "typical" migration profile. Slightly more
than 50 percent of the dry season outmigrants from this zone are
unmarried women going to Dakar to work as maids.
Rainy season
migration is generally lower than that during the
dry season.
It is more common in the northern Peanut Basin and Niakhar
in the Central Peanut Basin where poor rainfall and land constraints
make it increasingly difficult to farm.
Senegal Oriental, with its low population density and relatively
good rainfall, surprisingly has the highest rainy season outmigration.
In this zone there are several households with adult males on long-term
migration to Europe or to other African countries.
These individuals
are considered to be household members in those cases where they have
left spouses and children in the village or if they are single but
regularly send remittances to their parents.
Migration from urban households during the dry season is about 1
percent of total population; this increases to 4 percent during the
rainy season when some individuals return to rural areas to participate
in agricultural production activities.
46
Household Assets in Rural Zones
Rural assets considered in this discussion are land and
livestock.5
Farm size varies considerably across households.
The
sample contains farms as small as 1 hectare and as large as 60 hectares.
In the Peanut Basin zones, farms are typically 5-10 hectares; they are
somewhat smaller (3-5 hectares) in Senegal Oriental, where traction
equipment is not as common.
Access to household land by different members of the household is
subject to some variation across zones (see Table 3.8). Across-zone
comparison shows a range of 62 to 90 percent of males 215 years having
their own field(s).
The lowest access is in central Basin (Niakhar),
which has a serious land constraint.
The northern Basin also exhibits
a relatively low rate of "own fields" for males.
This may be due more
to the low productivity of agriculture in this zone than to the
availability of land.
Men most likely to not have fields are unmarried
sons, brothers, or uncles of the household head.
Participation by women ranged from a low of 44 percent in the
northern Peanut Basin to a high of 82 percent in Senegal Oriental,
Access to land was quite high in the southeastern Peanut Basin (77
percent) and in both zones of the central Peanut Basin (60-66 percent).
This simple picture of access to land without specifying the crop
or the size of the plots farmed fails to provide a true picture of the
economic
returns generated
by different groups having access.
Examination of net returns to agricultural production during the survey
shows that in the zones where the largest percent of women have access
to land, women's cropping income is a larger share of agricultural
income for the zone.
The typical share of net cropping income going to
women is about 10 percent.
It is highest (23 percent) in Senegal
Oriental, where a relatively large share of women stated that they had
access to land.
It is lowest in the northern Peanut Basin - a zone
where only a small share of women had access to land.
The typical rural household owns about 20 animals (including
traction animals and poultry) with an approximate value of 120,000 to
220,000 francs (CFA) per household or 13,000 to 21,000 FCFA per adult
equivalent.
Livestock holdings are extremely variable across households within
each zone (see Table 3.9).
For example, the range between the minimum
and maximum number of animals owned in a given zone varies between 55
and 107.
The range in value of holdings per household is more than
5 Although most rural households own animal traction equipment, it
was generally purchased in the sixties and seventies.
After
depreciation of this equipment, the net present value would be zero or
very close to it.
Only in Senegal Oriental, where traction equipment
has been available more recently through the cotton parastatal, is
there relatively new equipment. For this reason, traction equipment
has not been analyzed.
.
.
47
4,000,OOO FCFA in one zone (Senegal Oriental) and there are no zones
where it goes below l,OOO,OOO FCFA.
Even standardized values in FCFA/AE
continue to exhibit broad ranges between the smallest owners, who have
less than 2,000 FCFA/AE and the biggest owners, who
have 175,000 to
500,000 FCFA/AE.
Table 3.8 -- Access to land for persons 2 15 years in rural zones
Percent faming own field 1988/89 season
Zone/vi 11 age
Ma1 es
Females
Northern
70.1
44.4
Peanut Basin
Sagatta
68.6
36.0
Khelcom
83.8
41.0
Darou Cisse
55.2
60.0
Central Basin
62.1
61.2
(Niakhar)
Niakhar
58.1
54.1
Mboltogne
64.5
69.0
Mboyene
63.6
62.5
Central Basin
90.0
66.0
(Colobane)
Colobane
84.6
57.5
Teck Mar
94.6
62.2
Khayane
91.9
80.0
Southwest Peanut
75.5
54.2
Basin
Passy
70.3
27.8
Keur Aly Gueye
76.7
57.1
Ndiathiang
80.0
77.8
Southeast Peanut
83.0
77.0
Basin
Dioly Mandat
78.0
74.0
T. Thieckene/M. DiocKoul
77.0
83.0
Kouthia
95.0
77.0
Central Senegal
86.0
82.1
Oriental
Missirah
73.0
74.0
NiaoultYTouba
80.0
76.0
Medina AlylMoukouty
77.0
70.0
Source:
ISRA/IFPRI project survey.
..!
:”
..
48
Table 3.9 -- Livestock assets in rural zones
Central
Characteristic
North
Southwest
Southeast
Senegal
Basin
Niakhar
Colobane
Basin
Basin
Oriental
Number of animals
Mean
22
17
15
20
22
26
Median
Minimum
1
1
1
1
1
1
Maximum
86
92
5 5
70
6 7
107
@$g&ijitii
.:‘. a
Me.;i...
294
567
212
283
461
533
Quartile values
Lower 25% own
less than
this amount:
49
4 7
7 5
65
a7
2 2
Median value:
129
154
122
191
339
221
Upper 25% own
more than
this amount:
274
645
278
423
649
566
Minimum
6
6
2
10
14
.5
Maximum
1895
4065
1334
1433
2866
3102
gJ&y&@ g&y&&<
.j:.: .;.: . . . . . . “.tiGa;, ,,.,., ,.,,.,.j .,.,.,,,,
3 2
7 1
26
30
5 3
46
Quartile values
Lower 25% own
less than
this amount:
7
9
11
12
11
5
Median value:
13
25
1 8
20
28
21
Upper 25% own
more than
this amount:
39
96
2 7
40
70
7 6
Minimum
.9
.9
.6
1.9
1.4
.1
Maximum
177
509
175
172
312
249
ource:
ISRA/IFPRI project survey.
otes:
Information presented on this table is based on livestock inventory data collected shortly before the beginning of the
1990/91 cropping season (Feb.-April, 1990). Animals are valued by category (cattle, sheep, goats, etc. without respect
to age, sex. or weight) using producer prices obtained from ISRAIIFPRI transactions data and average over the entire
survey period.
Note that values are reported in thousands of FCFA. Nia. Niakhar; Colo. Colobane.
:
.
:
.
.’
~
49
When comparing the typical value of holdings per AE across zones,
the southeastern Peanut Basin comes out on top, followed by the Niakhar
zone of the central Peanut Basin (median values of 28,000 and 25,000
FCFA/AE).
The zone with the lowest value of holdings is the northern
Peanut Basin
( d
me ian value of 13,000 FCFA/AE), where livestock
contributes substantially to overall household income (see Chapters 4
and 7).
Details concerning the mean and median size of livestock herds and
the mean, median, and quartile values of holdings by zone are presented
in Table 3.9.
Income from livestock transactions is discussed in
Chapter 7.
Indicators of Urban Livinq Standards"
Tables 3.10 and 3.11 present characteristics of urban households
giving the proportion of households in the two towns owning real estate,
household appliances or a means of transportation. The proportion with
access to water and electricity is also shown.
Habitat: Type and ownership
More than 75 percent of the households in both towns own the house
that they occupy (79 and 76 percent respectively in Kaolack and Tamba).
About half of these houses
are built of durable materials
(cement/concrete) in Kaolack and about one fourth in Tamba where housing
is typically made of "banco" (clay mixed with straw), crintin (straw
mats) or straw. In Kaolack, 42 percent own empty lots in the same or
other towns compared to 12 percent in Tamba.
Means of Transport
The Tamba households seem better equipped to meet their transport
needs: 41 percent of them own at least a 2 wheeler (bicycle, moped,
motorcycle) and 12 percent own a car or a truck.
In Kaolack, the
proportion is only 9 and 3 percent respectively for both types of
vehicles.
Household appliances
In general, urban households own few appliances: less than 33
percent of households in both zones combined own an appliance. Two times
as many households have a sewing machine in Kaolack as in Tamba (18 and
9 percent); however, the opposite is true in the case of fans and air
conditioners (12 and 24 percent). Fifteen percent of the households in
the sample in each town have at least a fan.
6 The following paragraphs, as well as any other sections of this
draft report presented in French, will be translated to English for
the final report.
:
.
..:
:
50
Table 3.10 -- Standard of living indicators for households in urban sample
Characteristics
X of households possessing
Kaol ack
Tamba
., . . . . . . ,.,. ., : : : ,. ,. ./ ,.
A. ~$j$;~“pi:~Pa
P....ngDorr+tt-ari’!’
Bicycle/motorcycle
9
4 1
Automobile/truck
3
12
Cart
3
6
B .
j$.$K;gtiGIB'i$ .:':?? ..>:.:pki
earn
Sewing machine
18
9
Refrigerator/freezer
15
1 5
Fan/air conditioner
12
2 4
Others: camping stove,
gas
- '
3 5
peanut mill
c .
@&ond cs
Radio/cassette player
9 7
88
TV/Video
9
2 6
Telephone
3
0
0.
m
Owner occupies house
79
7 6
Owner of another house
6
0
Building constructed in
4 8
21
cement
Owns parcel of bare land
4 2
12
Have electricity
3 6
3 5
Have water faucet in compound
4 8
6
Have well in compound
3
8 8
Source: IFPRI/ISRA project survey.
' not available for Kaolack.
.’
:
51
Table 3.11 --Settlement profile of households in the urban sample
Characteristics
Kaol ack
Tamba
&G!ige. . .
-.... ::dur$$ .I,... . . . .
,. f.tjii.‘ii:j ,,.,.,.,.,.................
,.I.~.~..~/,,j ..:, ,:,.. ,.
:.:., :
..,.:. . .
,:
‘.,,.:.: ::, ..:I..:.:.):.:.),.~
n’.rilfo~.:iilb~~~,~~~.~: ,..:.:.:...- ,.:.>::.:.:.: .,.,.,.,..,,,...,.....,.,.,,.
30
24
_,., . . . . . . . ., .. . . . . . . . .
P’liac~:..t3f.:i:arr,g:n:l,:~~~:~~~:~a~~~.~~:~~~~,~~ha~
.::.,...:..:
.,.,.
.,.,.,..;:.:..:.y
y::,,.. .
:;.:.:.:..:. :
,:,:,:,,
,,,,.:
:,:,.::,:,:.:,:,..:,,.,, d
.:,:,:,.
_,.,.,..... ,, :.:.:...:.:.:.:.:.:.:.:.:.:.:.:.:
. ., ,.
:.:.:
.,.,.:
. . . . . . . . . .
- Same town (% of households)
9
3
- Other urban semiurban centers (%)
22
30
- Rural zone in the same region (%)
33
37
- Rural zone in another region (%)
33
10
- Outside the country (X)
3
20
- Born there (%)
9
3
- Work (%)
54
71
- Rejoin family/marriage (%)
34
26
Source:
IFPRI/ISRA project survey (1990-91).
Almost all households in Kaolack have a radio and/or tape recorder
compared to 88 percent in Tamba.
However, it is surprising to note that
many more households in Tamba than in Kaolack own a television set or a
video: 26 percent compared to 9 percent.
A telephone is a luxury in
both zones; only one household in Kaolack subscribes to SONATEL.
Direct access to water and to electricity
Slightly more than one third of the households in each zone has
direct access to electricity through SENELEC. Kaolack has more difficult
access to water than Tamba; only half of the households from Kaolack
have a faucet or a well inside their compounds while the other half has
access to water mostly from public water fountains.
Almost all
households in Tamba (94 percent) have a private water source;
subscribers to SONEES are not numerous (6 percent) because wells
predominate as the main source of household water (88 percent) in Tamba.
Settlement of urban households in towns
The average length of urban settlement of survey households is 30
years in Kaolack and 24 years in Tamba (Table 3.11). In both towns, the
proportion of household heads born in the town of current residence is
low: 3 percent in Tamba and 9 percent in Kaolack. Households from rural
zones in the same or other regions constitute almost two thirds of the
Kaolack sample and 38 percent in Tamba. However, those who are native of
other urban or semi-urban centers within the country make up almost one
.
.
.
.
:
.
.!
”
.’
.
52
fifth of the sample in each town. There is also an important share of
households in Tamba coming from outside Senegal (about 16 percent
compared with 3 percent in Kaolack), especially from Guinea and from
Mali.
The search for work, commercial factors or transfers initiated by
one's employer are the reasons most often mentioned to explain the
settlement of a household in town: 51 percent in Kaolack and 59 percent
in Tamba.
Other motivations
are
strictly personal
("to join
family/husband or "to be entrusted to a family member") and make up 34
percent of the cases in Kaolack and 22 percent in Tamba.
RELEVANCE OF HOUSEHOLD CHARACTERISTICS FOR AGRICULTURAL PRICE POLICIES
Two important reasons for looking at sample demographic and
socioeconomic characteristics are (1) to ensure that our sample
resembles the general population and (2) to examine factors that might
cause households or zones to respond differently to agricultural price
policies.
From the above discussion,
we conclude that the sample is
generally representative of the population it is to reflect in terms of
age, gender, and ethnic distribution.
The large share of the population
(15 years of age has important economic implications because this puts
a greater burden on the "active" population.
Data for the urban zone
show that 17-19 percent of the entire sample population had a source of
income at the time the survey began.
This means that each "employed"
individual must earn enough to support more than five people.
The data confirm what has been observed elsewhere -- that
education and literacy levels in both rural and urban Senegal are
extremely low.
This finding has important implications for economic
modeling exercises that assume farmers are able to use available price
information
and
their
knowledge of agricultural
input/output
relationships to make rational, profit-maximizing or cost-minimizing
production and marketing decisions.
When farmers are unable to keep
records of economic returns to different input combinations across years
in an environment as climatically fickle as Senegal, it is extremely
difficult to sort out the relative profitability of different
activities.
Village reconnaissance results reported in Chapter 2 suggest that
cropping activities provide the principal source of women's income in at
least 75 percent of villages in each zone. Information provided in this
chapter shows, however, that women's fields produce a relatively small
share of total household crop output.
These results imply that while
income from personally managed plots accounts for an important share of
women's total income,
the value of women's crop production is a
relatively small share of total household production.
. .:
53
Much of the women in development literature on Africa portrays
African women as the primary agriculturalists responsible for producing
most of the food crops.
Information provided above shows that this is
not true
for Senegal, where
household heads have the primary
responsibility for providing household cereals and female headed
households are rare.
Although Senegalese women in the IFPRI/ISRA study
zones manage only a small share of the area cultivated, they play a very
important supporting role by providing labor to fields managed by other
members of the household.
Understanding the extent of female
participation in agricultural production is extremely important when
designing agricultural policies to increase production and marketed
surplus.
In the urban zones, there is a significantly higher level of
female-headed households (14-26 percent). The implications of this for
household well-being (levels of income in particular) and consumption
habits (preferences among cereals,
for example) warrant further
examination.
The wide range of local and migratory activities practiced by
sample households reflects the transition from a purely agricultural to
a mixed economy that is taking place throughout Senegal. In many cases,
this is a transition that is being forced on an agrarian population that
is becoming less and less able to provide its basic needs through
agricultural production due to short and unreliable rainy seasons, a
deteriorating natural resource base, and a growing population. In other
cases,
it can be the result of poor infrastructure or inappropriate
agricultural policies.
The role played by the diversification of income sources in
providing food security and liquidity for improving agricultural
production is one of the most important issues to be addressed by this
study.
The data suggest that agricultural policies in Senegal cannot
be adequately designed without paying greater heed to the broader policy
issues of rural development that take into account the links between
agricultural and nonagricultural sectors.
Chapter 4 deals with these
issues in depth.
One generally anticipates that households with larger asset bases
enjoy better standards of living and tend to be more economically
productive.
We use information on access to land and livestock holdings
to examine the extent to which these anticipated relationships can be
corroborated.
Assets are frequently used to classify households. Knowing the
effect that agricultural policies have on households with different
asset bases and the extent to which asset levels determine the type of
response a household makes to a particular policy can greatly improve
policy design.
Our survey is unique in that it contains data on assets
as well as income and expenditure.
We are able to use all of these
variables to classify households, examine the relationship between the
different indicators,
and the relationship between levels of the
indicators and impact from or response to policy interventions.
.
.’
:
54
SUMMARY OF MAIN POINTS
Chapter 3 presents descriptive information about the sample
households and household heads.
Both demographic and socioeconomic
characteristics are discussed.
The main points to retain about the sample are:
Average age is 20 years while median age is 15.
The active population is about 45-50 percent of the total
sample.
The main points to retain about the sample households are:
Household size is generally 11-13 people and 9-10 adult
equivalents;
it is smaller in the urban zone of Tamba (9
persons and 7 AE) and the Niakhar zone of the central Peanut
Basin (10 persons and 8 AE).
The nuclear family accounts for 55-60 percent of the
household population except in Kaolack where it is 47
percent.
Households in the southwest Peanut Basin and the Niakhar
zone of the central basin have the highest rates of school
enrollment (about 70 percent of households with children in
school) while the southeastern basin households have no
children in school at all.
Zones with the highest outmigration during the dry season
are the northern basin and Colobane in the central basin;
Senegal Oriental has the highest rates during the rainy
season.
Seventy to 90 percent of adult males have access to farm
land in all zones but Niakhar in the central basin where
only 62 percent have access;
45 to 60 percent of females
have access in the north, center, and southwestern basin
while 75 to 80 percent have access in the southeastern basin
and Senegal Oriental.
Households in Niakhar in the central basin, the southeastern
basin and Senegal Oriental have the largest value of
livestock holdings per AE (roughly 45 to 70,000 FCFA/AE),
while households in the northern and southwestern basin, and
Colobane in the center have the smallest values (25 to
30,000 FCFA/AE).
.!
.
:
.
55
The main points to retain about household heads are:
Rural household heads are almost exclusively males (only two
females in the entire sample).
Urban household heads are 26 percent female in Kaolack and
14 percent female in Tambacounda.
Only 2 percent of rural household heads have any formal
education (French system) while 20 percent of the Kaolack
sample and 30 percent of the Tambacounda sample have some
formal education.
In most cases, the characteristics of the sample reflect the
characteristics that the reconnaissance survey suggested one would find
in each zone.
:
-,
,
4. DESCRIPTION AND ANALYSIS OF RURAL AND URBAN INCOME
INTRODUCTION
The level of income is an important determinant of the level of
household expenditures in general and the composition of the food
basket in particular.
It can also influence production behavior,
particularly a household's ability to make capital investments or to
purchase costly inputs when credit markets do not function well.
The share of income earned in different activities can also
influence consumption and production behavior.
In most studies, data
on total household expenditures are used as a proxy for income because
it is so difficult to enumerate income data directly.
Unfortunately,
data sets based on expenditures cannot address questions about income
diversification and what this implies for household response to policy
changes.
For example, assuming equal levels of income, a Senegalese
household earning 30 percent of income from cropping and purchasing 80
percent of cereal needs will have a different reaction to a rice price
increase than a household that earns 80 percent of its income from
cropping and only purchases 10 percent of its cereals.
These two
households could also be expected to respond differently to programs
designed to improve
agricultural productivity by decreasing soil
erosion
if such programs
competed
with nonfarm activities for
household labor.
The IFPRI/ISRA data set provides a full enumeration
of household income by source,
thereby permitting one to better
understand household consumption and production behavior.
Objectives
Levels,
distribution,
and sources of income in rural Senegal
differ across zones due to differences in the natural resource
endowment (climate, soils), the infrastructure (transportation,
communication),
and the institutional environment (savings, credit,
markets).
The overriding objective of this chapter is to show how
levels, distribution, and sources, of income differ across study zones
and to understand what this
implies for policy design and
implementation.
The chapter responds to the following specific objectives:
1.
Compute the level and distribution of rural and urban income for
each study zone.
2.
Examine the degree
of disparity in incomes
using frequency
distributions and Gini coefficients.
.
I’
:
.’
*
:
:
.
.
57
3.
Analyze the nature and degree of income diversification by zone
and by income strata.
4.
Develop a topology of diversification patterns.
5.
Model the relative importance of factors that explain higher
shares of nonagricultural incomes.
6.
Discuss the relevance of income levels and diversification
patterns for design of rural development policies.
Research questions
The objectives mentioned above can be summarized in the form of
research questions that we attempt to answer in this chapter. The
questions are as follow:
1.
What is the importance in absolute terms (level) and relative
terms (share) of cropping and non cropping incomes in the total
income of rural households by agroclimatic zones?
2.
What is the total level of income in urban sectors by study zone
and by strata?
3.
What different activities contribute to the composition of rural
and urban income,
and what is the share of each sector of
activity in the total income of the household?
4.
What are the most common income diversification patterns?
5 .
What are the factors that determine which type of diversification
pattern evolves in a particular zone?
6.
What is the policy relevance of the different levels and
diversification patterns found?
RESEARCH METHODS
Zones and time periods covered
Incomes for the rural sample are reported for harvest years
1988/89 and 1989/90 for five study zones (Niakhar and Colobane in the
central Peanut Basin, the southwestern and southeastern basin, and
Senegal Oriental).
For the northern Peanut Basin, only harvest year
1988/89 is covered.
A harvest year begins with the cereal harvest
(October 1) and ends just before the following harvest begins
(September 30).
We use a harvest year rather than a calendar year or
a cropping year to analyze income.
The harvest season provides an
excellent point of departure for the income analyses, because the
success of the harvest tends to influence household expenditure
.
”
.:
.
:
.,
.
:
58
patterns as well as decisions on allocating labor and capital to both
cropping and non-cropping activities throughout the rest of the year.
The urban analysis covers the towns of Kaolack and Tambacounda
for the period July 1, 1990 through June 30, 1991. These dates were
dictated by funding considerations.
As urban incomes are not subject
to the pronounced seasonal patterns of rural incomes, the beginning
and ending months of the survey period do not pose methodological
problems.
The lack of overlapping periods for the rural and urban
surveys does, however, limit the possibility for modeling rural and
urban demand concurrently.
Data collection methods
Fall, Kelly, and Reardon (1989) describes in detail the types of
questionnaires and procedures used to collect the income data.
The
following paragraphs summarize the most important aspects of data
collection procedures used.
l Cropping income was collected using a plot-level questionnaire
that enumerated all inputs and outputs for each field.
@ Data for on farm livestock income come from a questionnaire
that enumerated all purchases, sales, and home consumption of
livestock.
The questionnaire was administered monthly.
l Data on locally earned off-farm wages, cash transfers, and
borrowing were collected monthly.
l Data on migration remittances were collected by enumerating all
cash and inking remittances received by household members and
corroborating this information through in-depth interviews with
migrants when they returned home.
l Data on off-farm activities were collected using a variety of
techniques.
For local off-farm activities, interviewers used a variable
interview
schedule with more frequent interviews for persons
exercising activities that were difficult to enumerate (for example,
shop owners or petty traders with rapidly changing inventories) and
longer recall periods for less complex or less frequently practiced
activities (for example, preparing and selling donuts once a week at
the market).
Questionnaires on income from off-farm activities were
introduced about six months later than other questionnaires and a long
recall used to cover the first six months.
The recall was usually
conducted by a researcher rather than by the interviewer.
This was
done because we felt that project personnel had to develop the
necessary rapport with respondents before asking detailed questions
about a topic as sensitive as personal income.
.*
:
”
:
59
Analysis procedures
The activities are organized into the following nine production
sectors: own cropping activities
livestock production
wages from off-farm cropping
commerce
services
transport
food preparation
gathering
handicrafts
Table 4.1 provides a list of the activities classified in each of
these nine sectors.
Three additional 'Isectors" are used to report
income from cash transfers, borrowing, and remittances from household
members on migration.'
The methodology used for the analysis is based on the computation
of all sources of remunerative activities (cash and/or in-kind) taking
place locally/or in migration by members of a household during the
year. The income of each activity is calculated by deducting variable
and fixed costs (depreciation) from the gross receipts.
Cropping income is estimated in terms of net returns to family
labor following methods described in Chapter 5.
Livestock income includes the households' net receipts from
transactions of animal products and by-products and the income from
activities related to animal husbandry (shepherding).
Capital gains
and losses associated with changes in the size of the household's
total livestock holdings are not considered.* Details of estimation
methods used are described in Chapter 7.
Net transfers and borrowing reported in this chapter cover only
t r a
cash n s a c t i o n s .
No attempt has been made to value and incorporate
inking transfers and borrowing into the income analysis.
The household income for a given sector constitutes the sum of
net income earned by all household numbers from activities in the
sector.
Total household income is the sum of income across all
sectors plus net transfers and borrowing. Net transfers and borrowing
' The definition of all the sectors appears in Fall (1991).
2 See Ndiaye 1993 for a more complete analysis of the IFPRI/ISRA
livestock data for Niakhar in the central basin and the southwestern
basin.
Ndiaye combines the stock and transactions data to estimate
capital gains and loses.
.’
1
.’
Table 4.1 -- Classification of income generating activities by sector
CM-I
Livestock
Agricultural
Cannerce/
Service
Transportation
Food
Gathering
Cottage
:
Cropping
Labor’
Trade
Processing
I&tries
Crop
Herding
Cropping labor
Shop owner
Tire repair
By cart
Peanut processing Gathering/sale
Tailoring
Production
to produce oil
of wood
Cattle sales/
Fruit/vegetable
Petty-trade
Bicycle repair
By motorized
Errbroidery/
Fruit/
trading of
gardening, labor
vehicle
Sale of Koumpa/
Gathering/sale
Hand-
Vegetable
household
Livestock
Apprenticeship
beignets
of fruit
sewing
Gardening
animals
trade’
Various
Maid
transportation
Sale of roasted
Gathering/sake
BLacksmith
Sale/processing
Fish trade
Peanuts
of birds
animal by-
Civil Servant
Basketry
products (mi lk,
Kola nut trade
Bread making
leather)
Guard
Leatherwork
Drink trade
Restauration
Village Chief
Cloth
Tools, parts,
weaving
t i r e r e p a i r
Marabout
Jewelry-
Hardware trade
Praise singer
making
Tobacco trade
Veterinary
Fence
Assistant
construction
Assistant trader
Hechani c
Peanut trade
Hairdresser
Blacksmith
Retirement
Pension
Source:
IFPRI/ISRA survey data.
’ For a third party outside the household.
’ This involves the buying and reselling of cattle that are not considered part of household herds.
61
do not originate in a specific sector, but constitute supplemental
sources of income.
Total household income was standardized on the basis of household
size (income per capita) and on the basis of household composition by
age and gender (income per adult equivalent). These results were then
compared to estimates of household expenditure per adult equivalent as
a control.
Only those households where income per AE was at least 70
percent of expenditure per AE were retained for further analysis.
Given the difficulties inherent in collecting income data the
assumption was that the expenditure data were more reliable than the
income data.
The average income per adult equivalent was then calculated for
each village,
permitting one to examine inter-village differences
within each zone as well as the influence of market infrastructure
within and across zones.
Individual household observations were
weighted by household size (AE) in calculating village averages.
Tables with the village and zone averages per AE are presented in
Appendix tables A4.
Most of the discussion in this chapter draws on estimates of
average income per AE or average income per capita for each zone. In
these estimates individual household observations were weighted by
household size (AE) and a coefficient to correct for intentional over
sampling of market villages.3
For the 1988/89 harvest year data for 189 rural households were
analyzed.
In 1989/90, the number of households with complete income
data was 158.
Some analyses (primarily ones making inter-annual
comparisons) use only those rural households for which two full years
of income data are available; for these analyses, the sample size is
141 households.
The urban analysis covers 60 households (31 in
Kaolack and 29 in Tambacounda).
LEVELS, DISTRIBUTION, AND SOURCES OF INCOME
Levels of Income
Table 4.2 shows that average household income in rural zones
ranges from 27000 FCFA ($95) per AE in the central Peanut Basin zone
of Niakhar in 1988/89 to 72000 FCFA ($250) per AE in the southeastern
Peanut Basin during harvest year 1989/90. Average income for urban
3 See Kelly and Reardon, 1989, for a full discussion of sampling
methods and stratification procedures.
.
,.
:
.
,’
-.
Table 4.2 --
Level and distribution of net household income per AE (FCFA) by zone and year
~
Northern
Central Basin
Central Basin
Southwest Peanut
Southeast Peanut
Central Senegal
Peanut Basin
Ni akhar
Co1 obane
Basin
Basin
Oriental
a
c
Kaol
k
Tambacounda
Y e a r
a8189
88189
89190
88189
89190
88189
89/90
88189
89190
88189
89190
go/91
go/91
1. Rev. per AE
42509
26975
32696
33319
56246
43459
67208
49806
72051
39163
42487
107760
102625
N.Obs.
-29
25
3 1
33
35
30
2 7
36
36
36
29
3 1
29
C.V. (%)
70
4 1
37
60
49
60
a4
41
68
7 1
145
125
64
2. Rev. per Capita
32072
20298
20297
25744
42190
32675
49309
36685
51897
29340
30893
83625
79950
N.Obs.
2 9
25
3 1
33
35
30
27
36
36
36
2 9
3 1
29
C.V.(%)
7 3
42
39
6 2
5 1
57
a2
39
67
69
7 5
140
67
.
3. Gini
Coefficient
.34
.30
.23
.30
.26
.30
.43
.I8
.26
.38
.49
.48
.35
.,
Source:
ISRA/IFPRI Project survey, 1988-1991.
Notes:
In rural zones, the data are presented by harvest year (from October 1 - September 30) for 1988/89 and for 1989/90. However, in urban zones,
the period is from July 1 - June 30, 1991.
..(I.
:i
Sagatta was only covered in 1988/89.
The Gini Coefficient is calculated using income per capita.
6 3
households is higher -- about 105,000 FCFA per AE ($365) in both
Kaolack and Tambacounda in 1990/91.4
The level of average income per AE varies significantly across
zones.
Two of the three study zones with high agricultural potential,
the southwestern and southeastern Peanut Basin, have the highest
levels of income both years.
Average incomes in the third zone with
high potential (Senegal Oriental) are, however, at the lower end of
the spectrum both years.
The Niakhar zone in the central Peanut Basin
exhibits the lowest average incomes across all zones in both survey
years.
Average urban incomes in 1990/91 are about double most of the
averages for rural zones in the year following the above average
1989/90 harvest.
The highest net income per capita found in rural areas during the
survey period (52,000 FCFA in the southeast of the Peanut Basin} is
about one fourth of the value of gross domestic product per capita
estimated from macroeconomic datae5
The urban incomes are about half
of the GDP estimates.
This illustrates, as one would expect, that GDP
figures fail to provide an accurate picture of purchasing power and
living standards for an average Senegalese household.
A comparison of household income and the cost of minimum needs in
each zone helps one evaluate the adequacy of incomes earned by rural
households."
Following the poor 1988/89 harvest, income did not
cover minimum needs for 50 percent of households in the Niakhar zone
of the central basin.
In the Colobane zone of the central basin, the
southwestern basin and Senegal Oriental, 20 to 30 percent of
households had incomes less than minimum needs.
Following generally better 1989/90 harvests, Niakhar (central
Peanut Basin) and Senegal Oriental still had a substantial share of
households with combined income from all sources less than "minimum
needs" levels (35 and 21 percent, respectively).
Interannual differences in levels of rural income are large.
Following the better 1989/90 harvest, average household incomes (from
4 The per capita equivalents of the Niakhar and southeastern
Peanut Basin figures cited above are 20,000 and 52,000 FCFA,
respectively ($70 and $180); the urban per capita figure is 82,000
FCFA ($280).
During the survey period, US$l.OO was worth
approximately 290 FCFA.
5 See, for example, DPS, 1988, or IMF, 1992, both of which report
GDP per capita of about 180,000 FCFA.
6 Minium needs is defined as the cost (using zone-specific
prices) of 1900 kcal/AE/day of coarse grains plus 20 percent to cover.
other essential food and nonfood expenditures (see Kelly, 1993 for
details)
..!
:
64
all sources) increased by 20 to 70 percent throughout the Peanut
Basin.
The central Peanut Basin zones had both the lowest (Niakhar)
and the highest (Colobane) increases.
Most of the increase came from
better cropping outcomes which increased by 50 to 100 percent between
the two years, but some also came from higher off-farm incomes.
Only Senegal Oriental failed to experience an important increase
in income in 1989/90.
This was due in part to poor rainfall but also
because farmers in the zone boycotted cotton production the second
year.
Distribution of Income
The degree of concentration of income in the households for each
study zone is estimated using the Gini coefficient.
This measurement
provides information on the equality (coefficient nearing 0) or the
lack of equality (coefficient nearing 1) of income distribution for a
given population.
The Gini coefficients calculated on the basis of
average income per capita are shown for each zone and year in Table
4.2.
Gini coefficients range from .18 to .49, but are generally in
the .25 to .35 range.
Most zone estimates are similar to those found
in other Sahelian countriese7
The distribution of income in central Senegal Oriental and in the
southwest Peanut Basin is more unequal (Gini about .4) than in the
other rural zones.
This is due in part to the level of income of
certain migrants (in Senegal Oriental) and large merchants (southwest)
which are considerably higher than the average
income
in their
respective zones.
The urban distribution appears to be more skewed
than the rural with Gini coefficients of .35 in Tambacounda and .48 in
Kaolack.
One also notes that in the zones with greater agricultural
potential,
Gini coefficients increased following the better 1989/90
harvest.
This result suggests that something is preventing the poorer
households from fully realizing the benefits that accrue from better
rainfall.
Tables 4.3 and 4.4 show the frequency distribution for average
net household income by zone using seven categories.
The modal
category for rural zones in both survey years is 30,000 to 60,000 FCFA
per AE.*
This category accounts for approximately 45 percent of
rural households in the sample both years.
Table 4.3 shows that the
two central Peanut Basin zones and Senegal Oriental have the highest
number of households with incomes below 20,000 FCFA per AE.
Incomes
at this level generally fail to cover minimum needs.
7 Nigeria , Niger and Burkina Faso, for example, are .28, .30 and
.33 (Matlon, 1979; Reardon, 1989; Hopkins and Reardon 1993).
a This is approximately $100 to $200 per AE per year.
.
.
:
I
I
65
Table 4.3 --
Frequency distribution of average net household incomes (FCFA/AE) by zone and
year for rural zones
Part A: Harvest Year 19BB/B9
Incaiie
Northern
Central Basin
Southwest
Southeast
Central
Category
Peanut
Peanut
Peanut
Senegal
Basin
Niakhar
Colobane
Basin
Basin
Oriental
(number of households)
1100,00
5
0
1
1
1
2
60-99,999
0
0
2
6
1 3
5
30-59,999
1 7
10
1 5
13
1 9
1 3
20-29,999
5
7
9
a
3
8
15-19.999
2
6
6
1
0
4
lo-14,999
0
2
0
1
0
2
<
10,000
4
Number of
29
25
3 3
30
36
38
Households
Part B: Harvest Year 1989/90
Income
Northern
Central Basin
Southwest
Southeast
Central
Category
Peanut
Peanut
Peanut
Senegal
Basin
Niakhar
Colobane
Basin
Basin
Oriental
(number of households)
I100,OO
na
0
3
5
9
3
60-99,999
na
1
1 2
4
1 5
2
3659,999
na
1 9
1 7
1 2
1 1
1 0
20-29,999
na
6
3
4
0
9
1s19,999
na
4
0
2
1
3
lo-14,999
na
1
0
0
0
2
Number of
na
31
35
2 7
36
29
Households
.
.
_’
. :
66
Table 4.4 -- Frequency distribution of
average net household
incomes (FCFA/AE) in
urban zones
Incune
Kaol ack
Tarnbacounda
Category
(number of households)
I100,00
13
12
60-99,339
11
9
30-59,999
5
6
20-29,999
1
1
1s19,999
1
1
10-14.999
Number of
31
23
Households
The modal group for urban households in 1990/91 is the highest
income category --
2 100,000 FCFA per AE.'
The modal group includes
about 40 percent of the urban sample.
Sources of Income
One of the major strengths of the IFPRI/ISRA data set is the
manner in which it permits household income to be disaggregated by
sector in both rural and urban zones.
Table 4.1 at the beginning of
the chapter illustrated how income generating activities were mapped
into 9 economic sectors.
Table 4.5 shows average income by sector for
each rural zone and year of the study while Table 4.6 presents the
urban results.
Table 4.5 reveals that cropping activities contributed from 20 to
70 percent of rural household income in 1988/89 and 40 to 80 percent
in 1989/90.
An extreme case of only 20 percent cropping income was
found following the locust invasion of the northern Peanut Basin in
1988/89.
Consistently large shares of cropping income (greater than
70 percent) were found only in the southeastern Peanut Basin.
Most
zones were in the 40 to 50 percent range in 1988/89 and 70 to 80
percent range in 1989/90.
' This is approximately $345 per AE per year.
67
Table 4.5 - Net household incom per AE (FCFA) by zone, year, and sector for rural sample.
No
Own
Ag
Food
cot
Ttl
Obs
Crop
Lvsk
labor
Cm
Serv
Tv
Prep
Gath
Ind
Trf
Bar
Mgrt
Unk
Rev
NORTHERN PEANUT BASIN
88-89
2 9
7340
7899
578
4004
3919
0
1766
0
8810
1640
491
2762
3300 42509
Pet
17
19
1
9
9
0
4
0
2 1
4
1
6
8
100
c.v
.92
2.00
1.40
4.55
4.01
3.21
2.91
2.03
3.27
1.62
1.61
.70
CENTRAL BASIN - NIAKHAR
88-89
25
12642
1505
7 5
4699
4427
0
119
0
2386
87
0
805
232 26977
Pet
47
6
0
1 7
1 6
0
0
0
9
0
0
3
1
100
c.v
.68
2.10
1.06
1.69
1.94
4.72
2.32
3.52
2.42
3.69
.41
89-90
31
25931
1125
44
2260
1086
0
40
0
987
-20
119
929
1 9 5 3 2 6 9 6
Pet
79
3
0
7
3
0
0
0
3
0
0
3
1
100
c.v
.51
3.81
1.81
2.49
3.29
5.84
4.37 -11.90
6.03
2.75
3.98
.37
CENTRAL BASIN - COLOBANE
88-89
33
18154
7512
168
1190
1250
814
1597
1 2
1178
-88
123
1409
0 33319
Pet
5 4
2 3
1
4
4
2
5
0
4
0
0
4
0
100
c.v
.74
1.19
1.15
3.64
2.94
4.74
3.76
5.38
3 . 8 2 - 1 3 . 0 8
3.42
1.82
.60
89-90
35
42624
2615
169
864
2716
2521
1192
6 1
1469
5 2
180
1782
0
56246
Pet
7 6
5
0
2
5
4
2
0
3
0
0
3
0
100
c.v
.69
2.55
1.20
3.82
3.65
4.75
1.40
4.52
3.61
33.49
8.60
1.76
.49
SOUTHWESTERN PEANUT BASIN
88-89
3 0
20738
1810
1180
4027
10292
1250
323
724
1686
6 3
1350
0
15
43459
Pet
48
4
3
9
2 4
3
1
2
4
0
3
0
0
100
c . v
.53
1.82
2.85
2.47
2.15
3.82
7.42
5.08
3.50
4.19
1.36
4.91
.60
89-90
27
33426
2380
1257
16746
9117
1176
833
593
1772
1234
-1474
140
7 67208
Pet
5 0
4
2
25
14
2
1
1
3
2
-2
0
0
100
c.v
.49
2.68
3.00
3.08
2.24
4.27
3.54
4.60
3.50
3.57
-2.89
3.07
4.19
.84
SOUTHEASTERN PEANUT BASIN
88-89
36
33954
6465
1 5 1
2551
2699
54
253
182
677
-119
2660
114
165
49806
Pet
6 8
1 3
0
5
5
0
1
0
1
0
5
0
0
100
c.v
.48
1.23
1.85
2.85
1.54
5.17
4.80
5.63
2.97
-7.43
1.74
6.15
4.47
.41
89-90
36
51003
12865
126
5652
2827
414
189
155
1216
-1343
-1436
2
380
72051
Pet
71
1 8
0
8
4
1
0
0
2
-2
-2
0
1
100
c.v
.73
1.12
1.87
2.62
1.96
6.12
5.01
5.66
3.25
-3.85
-4.06
25.94
3.84
.68
SENEGAL ORIENTAL
88-89
36
18688
5931
687
2579
2147
9
1447
276
764
2209
331
4115
0
39163
Pet
48
1 5
2
7
5
0
4
1
2
6
1
11
Cl
100
c . v
.70
1.56
1.39
2.80
2.42
4.07
3.43
3.56
3.64
2.02
2.16
3.43
.71
89-90
2 9
15673
9326
656
967
1536
7
-92
20
2250
2670
-39
9504
0
42487
Pet
37
22
2
2
4
0
0
0
5
6
0
22
0
100
c.v
.64
2.09
1.69
3.59
2.11
4.98
-14.97
4.77
3.13
2.89
-48.60
3.43
.81
Source:
IFPRI/ISRA survey data.
Note:
See Table 4.6 for legend to sector abbreviations.
:
,
.’ 2
,’
:
.
Table 4.6 --
Net household income per AE (FCFA) by tom and sector for urban sample.
Z O N E
SECTOR
Own
Lvsk
corn
Serv
Tw
Food
Gath
cot
Trf
t30r
Mgrt
Total/
Crop
Prep
Ind
AE
KAOLACK
Rev/se
660
300
24085
49380
0
14570
0
4570
8740
640
2865
107760
%
1
2 2
46
0
1 4
0
4
8
1
3
100
c v (%)
368
1353
270
146
0
379
0
277
192
1752
416
129
Rev/se
485
0
11875
77115
175
5285
0
5910
2785
575
-440
102625
%
0
1 2
7 5
5
0
6
3
-1
100
c v (X)
673
0
180
102
244
282
0
260
374
914
314
6 4
Source: ISRAIIFPRI Project survey, 1990/91.
Legend for sector abbreviations:
Own crop = onfarm cropping income
Lvsk
= livestock
colwl
= commerce and trade
Tv
= transportation
Food Prep= Food processing/restaurant
Gath
= gathering plant/forest products for resale
Cot Ind = cottage industries
Trf
= cash transfers/gifts
B0r
= cash borrowing
Mgrt
= migration
,Unk
= unknown
69
Interannual differences in levels of cropping income are quite
large.
In the central Peanut Basin, levels of cropping income doubled
between 1988/89 and 1989/90.
In the southern Peanut Basin zones, the
level increased by 50 percent.
In Senegal Oriental the level of
cropping income dropped between the two years by about 20 percent (due
largely to the cotton boycott and poor rains in 1989/90).
Net income from livestock transactions was an important share of
total income (13 to 23 percent) in all zones but Niakhar in the
central basin and the southwestern basin (5 percent) following poor
1988/89 harvests."
The following year, livestock income accounted
for less than 5 percent of income except in the southeastern basin and
Senegal Oriental where it represented about 20 percent. The tendency
for livestock income to vary inversely with cropping income and,
therefore, smooth total income is clearly evident.
The final columns of Table 4.5 that concern income earned in the
agricultural sector (broadly defined as all cropping, forestry, and
livestock activities) are agricultural wage labor and gathering
activities.
Income is classified as agricultural wages only when it
is earned by working off-farm.
Both cash and inking payments are
included.
Gathering includes all activities where forest products
were gathered and sold; gathering for home consumption was not covered
by the survey.
It is noteworthy that only a very small share of total
household income (0 to 5 percent) is generated in agricultural wages
and gathering activities.
Low agricultural wage income is anticipated
in situations where there is not a large class of landless households.
When the demand for agricultural labor is high, everyone is working
their own fields and not hiring out their labor".
Gathering (particularly leaves for sauces and firewood) for home
consumption is common, but the resale of gathered products is not a
major source of household income in any of the IFPRI/ISRA study zones.
In summing the four columns that report income earned in the
agricultural sector (broadly defined), we find that households in most
zones earn from 60 to 90 percent of total income in these activities
except in the northern Peanut Basin where less than 40 percent of
income was earned in agriculture and animal husbandry.
Haggblade,
Hazel1 and Brown (1989) found that 25 to 30 percent of rural household
income in
selected countries
of sub-saharan Africa came from
" The extreme reluctance of the Serer households in the Niakhar
zone of the central Peanut Basin to sell their livestock is evident in
the IFPRI/ISRA data.
This zone has the largest livestock herds, the
lowest income levels, and the lowest caloric intake in 1988/89 yet
they sold very little livestock to supplement income.
" In most of the study zones, supplementary laborers
ar,e
temporary migrants from Guinea.
This is particularly true at harvest
time.
.
.
I
.
.‘.
:
:
:
.
.
I
. .
70
nonagricultural activities, while 70 to 75 percent came from cropping
and
animal
husbandry.
This suggests
that the northern and
southwestern basins and Senegal Oriental -- zones with less than 70
percent of income earned in agriculture (broadly defined) -- are
somewhat atypical of the sub-saharan region in general.
Although agricultural incomes remain an important share of total
income, many households still fail to meet minimum needs with incomes
earned directly in agriculture,
livestock and forestry activities.
Nonagricultural incomes, therefore, play a critical role in assuring
household food security and well being.
For this reason, it is
important to understand how sources of income vary by income group.
Nonagricultural income has a variety of sources; commerce and
service activities account for the largest share in most zones while
handicrafts (northern Peanut Basin) and migratory remittances
(northern basin and Senegal Oriental) tend to be more zone specific.
Household income from net transfers (gifts) and borrowing account
for less than 1 percent of income in both the Niakhar and Colobane
zones of the central Peanut Basin.
In the remaining zones, transfers
and borrowing together usually account for 5 to 7 percent of total
income.
Net transfers and borrowing are generally positive for sample
households suggesting that funds may be moving into sample zones from
other areas of Senegal or beyond national borders.
The only exception
was found in the southeastern Peanut Basin following the good 1989/90
harvest when sample households' net transfers and borrowing reduced
total income by almost 3000 FCFA per AE on average.
Relatively low shares of income from borrowing (either positive
or negative) suggest that although credit markets do exist, they are
not
p;oviding
important sums
of money over extended periods of
time.
Table 4.6 shows that the service and trade sectors take first and
second positions,
respectively in the two urban zones studied in
1990/91. The service sector represents 75% of the total income per AE
in Tambacounda, while trade accounts for 12%. The other sectors are
relatively less important.
In Kaolack, however, the service sector generates almost half of
the total income by AE (46%), and trade represents 22%. It is
important to note that Kaolack is reputed to be a trade center as a
result of its dynamic peanut trade and its proximity to Gambia which
stimulates cross-border trade.
Due to the role that Kaolack plays as
12
" Matt Warning, a Ph.D. student at UC Berkeley, is
currently doing an independent study of credit using the ISRA/IFPRI
sample in Niakhar and Passy.
He will collect additional data on
access in Niakhar and Passy.
This data will then be combined with
existing data and used to model demand and supply of credit.
,
.
‘:
.’
.’
,
71
a transport and commercial crossroads, the town registers an
impressive number of snack bars and restaurants. Results show that
income from the food preparation sector is quite important (14% versus
5% for Tambacounda).
Cottage industries provide about 5 percent of urban income. This
is comparable to the share in most rural areas.
The migratory sector provides 3% of the total household income in
Kaolack.
It is negative
in Tambacounda suggesting that sample
households sent more money to members on migration than they received.
This occurs when a migrant fails to find employment quickly.
Transfers are more important in Kaolack than in Tambacounda (8%
and 3%).
The fact that transfers are positive in urban as well as
rural areas - and account for a larger share of income in urban areas
- fails to support the conventional wisdom that cash transfers are
flowing from urban to rural areas.
More in-depth analysis of the data
is required to sort this out.
Tables 4.7 and 4.8 illustrate the manner in which the share of
income from different sectors varies across income strata.
For each
study zone and year households were ranked by total net income per AE.
Each household was then assigned to an income tercile according to its
ranking in the zone.13
Due to large differences in levels of income
across zones and nonproportional sampling procedures, we have made no
attempt to lump households together as the aggregate results could be
misleading.
Table 4.7 reveals that following the poorer 1988/89 harvest, the
share of cropping income is neqatively correlated with the level of
total income in all zones but Colobane (central basin). Households in
the lowest tercile have the highest shares of cropping income while
households in the highest tercile have the lowest share.
This
supports the hypothesis that households with more diversified incomes
are better able to compensate for crop shortfalls following poor
harvests because their total incomes are higher.
Following the above average 1989/90 harvest, the negative
correlation between the share of cropping income and the level of
total income remains unchanged in the two zones of the southern basin.
In the central basin, however, the middle and higher income households
have a larger share of cropping income than the poor.
One also notes (data not shown) that the absolute value of net
cropping income tends to increase more between the two harvests for
l3 In other words,
the third of households with the lowest
incomes in the zone were placed in the lowest income tercile, the
middle third of households in the zone in the middle tercile, and the
highest third din the top tercile.
.-,
;
.
I
,
.1’
:
:
72
Table 4.7 --
Average net total incane per AE (FCFA) and shares by sector and tercile for each rural
zone and year.
S H A R E O F
I N C O M E B Y
S E C T O R :
TOTAL
FOOO
COTTAGE
ZONE
TER INCOHE/AE
CROPS AGLAB LVSTK COI4 SEW TRPT PREP GATH INDUST MIGRA TRNFS BOR UNK
Northern
low
24839
36
4
24
0
2 0 2
0
1
11 2
1
1 8
Peanut
mid
35492
2 9
1
2 1
2
1 0 14
0
7
9 7
2 7
Basin
top
75336
1 1
0
1 4
1 7
17
0
0
0
3 1
3
1
1 4
Central
low
15481
5 7
1
7
6
16 0 0
0
1
9 1
0 3
Basin
mid
26909
5 3
0
a
2 7
8 0 0
0
3
1
0
0 0
Niakhar
top
41219
38
0
4
1 3
21
0 1
0
18
3 0
0 0
Central
low
20799
56
1
1 7
1 5
10 3
0
1
4 1
0 0
Basin
mid
30838
5 9
1
28
0
60 1
0
3
3
-1
0
0
Colobane
top
56919
6 1
0
17
2
3 4 6
0
4
4 0
1
0
Southwest
low
23721
7 0
6
a
5
0 5 0
1
0
0 0
5 0
Peanut
mid
37130
5 9
1
7
6
14 3 3
0
4
0 0
3 0
Basin
too
77613
42
2
0
1 3
31 2 0
2
6
0
0
2 0
Southeast
low
34002
7 6
7
3
6 0 0
0
1 0
2
1
Peanut
mid
55025
75
9
3
5 0 0
1
0 0
7 0
Basin
tOD
77331
70
15
6
20 1
0
0 0
5 0
Central
low
17883
66
9
0
8
0
0
1
1
7
2 0
Senegal
mid
34193
5 9
1 3
6
5
0
7
0
5 2
1
0
Oriental
top
72789
5 0
1 3
a
30
1
1
16 5
0 0
S H A R E O F
I N C O M E B Y
S E C T O R :
TOTAL
FOOD
COTTAGE
ZONE
TER INCOME/AE
CROPS AGLAB LVSTK COM SERV TRPT PREP GATH INDUST MIGRA TRNFS BOR UNK
Central
low
20960
68
0
12
1 0
6
0
1
0
2
2
0
0
0
Basin
mid
34205
aa
0
2
1
3
0
0
0
0
5
0
1
1
Niakhar
top
48636
7 8
0
-1
1 2
2
0
0
0
8
2
0
0
0
Central
low 3 5 3 8 0
68 1
3
8
60 3 1
4
7
1
-
l
0
Basin
mid 5 6 8 6 2
77 0
3
1
8 0 1 0
6
3
1 0
0
Colobane
top 100114
a3 0
4
0
2 7 2
0
0
1
1 1
0
Southwest
l o w 27988
a5
i
i
i
a
i 3 1
0
0
2 -2
0
Peanut
mid 48642
56 1
10
3
12 5 5 0
2
1
5
0
0
Basin
top
140647
40 2
2
39
14 1 0 1
4
0
0 -3
0
Southeast
low 42285
a2 0
6
3
5
0
0
1
5
0
-1 0
0
Peanut
mid 71963
a0
0
20
2
4 2 1
0
2
0
-5 -6
0
Basin
top 126319
74 0
15
9
2 0 0
0
a
0
0 0
1
Central
low 18951
47 8
1 9
0
4 0 0
1
1 2
8 0
0
Senegal
mid 3 0 6 2 2
57 1
27
2
6 0 -1
0
:
0
3
2
0
Oriental
top 141131
64 0
8
2
10
0
o
3
20
2 0
0
Source:
IFPRI/ISRA
survey
1988-90.
.
Note:
See Table 4.6 for legend to sector abbreviations.
73
Table 4.8 --
Average net total incane per AE (FCFA) and shares by tercile and sector for
each urban zone
S H A R E O F
I N C O M E B Y S E C T O R :
TOTAL
FOOD
COTTAGE
ZONE
TER INCOHE/AE
CROPS AGLAB LVSTK COH SERV TRPT PREP GATH INDUST HIGRA TRNFS BOR UNK
Kaolack
low
46142
0
0
0
I5
53
0
6
0
6
1
18
-2
3
mid
92319
0
0
1
1 2
5 6
0
8
0
7
8
6
-2
3
top
245092
0
1
0
30
38
0
1 9
0
2
1
5
2
1
Tabma
1OW
36046
0
3
0
32
30
1
2 1
0
10
0
4
-1
0
mid
95407
0
0
0
22
62
0
9
0
6
0
1
0
0
top
177737
0
0
0
3
90
0
0
0
5
-1
3
-1
0
Source:
IFPRI/ISRA survey 1988/89
Note:
See Table 4.6 for legend to sector abbreviations
households in the higher income terciles than for those in the lower
terciles.
This suggests that the wealthier households are better able
to realize the benefits of good rainfall than the poorer households.
The mechanism by which this works is not clear, but it is reasonable
to hypothesize that the wealthier households have greater liquidity to
purchase inputs (peanut seed, in particular) and are better able to
hire labor to get around bottleneck periods.
There is no clear pattern across zones and terciles concerning
the sectors in which off-farm incomes are earned.
For example,
livestock and migration are important sources of income for the poor
in the northern Peanut Basin and an important source of income for the
rich in Senegal Oriental.
Cottage industries are important to the
wealthy in the Niakhar zone of the central basin and in the north.
Poor households in the lowest tercile account for most of the
agricultural wage income earned in the north and the southwest.
The
commerce and service sectors provide an important share of off-farm
income across all terciles.
As Fall (1991) has noted, it is the
capital intensity of different activities within each sector that
tends to differentiate the lower and upper income terciles.
For
example, the poor in the commerce sector tend to be petty traders
while the rich are shop owners.
The above discussion has' stressed the critical role that
nonagricultural incomes play in rural Senegal.
It must be stressed,
however, that most of the nonagricultural sources of income are not
independent of the agricultural sector.
The amount of income earned
in nonagricultural activities is often determined by how successful
agriculture is in a particular zone.
This is true because many
nonagricultural activities are tied to agriculture through production
and consumption linkages.
.I
:
”
:
:‘.
I
,
74
INTERSECTORAL LINKS OF INCOME GENERATING ACTIVITIES IN RURAL AREAS
This section describes in greater detail the nature of activities
by sector and by zone to facilitate the examination of intersectoral
linkages in rural areas.
We are particularly interested in
identifying upstream and downstream flows between activities in the
agricultural and the nonagricultural sector.
As was mentioned in the previous section, cottage industry, trade
and services constitute the principal sectors of nonagricultural
activity in rural zones. One notes, however, that the nature of these
activities differs by zone.
Cottage industry in the north of the Peanut Basin includes
embroidery, basketry and shoe making.
In the other zones where cottage
industry is
less important,
the sector is dominated by sewing,
blacksmithing,
weaving and jewelry making. Some of these activities
are closely related to those upstream of agriculture (for example,
with blacksmithing as an input). Other activities of the sector are
likely to exhibit strong consumption linkages with the agricultural
sector.
For example, growth of the agricultural sector would lead to
an increase in the demand for cottage industry goods.
Trade is oriented toward the marketing of condiments, fish and
goods of first necessity in the northern Peanut Basin and the Niakhar
zone of the central Peanut Basin.
These products are, for the most
part, not directly linked to the local agricultural economy.
On the
other hand, in the Colobane zone of the central basin and the southern
Peanut Basin zones trade in cola nuts, cereals and peanuts is more
common. These activities exhibit downstream linkages with agriculture.
Animal trading ("tefanke") dominates in the northern basin and the
center of Senegal Oriental where animal husbandry is very important,
This activity has primarily downstream links to agriculture, but also
exhibits some upstream links.14
The service sector in all zones is made up primarily of "griot"
(praise singers) and "marabout" services, security and domestic work.
The latter activity is usually conducted in Dakar or in other regional
capitals. The income from these activities does not have production
links to the local agricultural economy. In the center of the Peanut
Basin and in Senegal Oriental, the service activity "repair of
equipment" does exhibit production linkages (as an input) to local
agriculture.
l4 Traders use agricultural by-products such as peanut hay to
feed animals, which in turn are potential inputs for agriculture (as
traction animals and sources of manure)
:
;
75
Food preparation is tightly tied to agriculture in all the zones.
This concerns production and sale of grilled peanuts and doughnuts15;
and extracting oil from peanuts (principally, in the Colobane zone of
the central basin). In the southeastern Peanut Basin, women living in
the market .village earn a large share of their personal income by
selling water to thirsty people and animals on market day.
Therefore, it is possible to conclude that many of the
nonagricultural activities in the zones covered by the study exhibit
some upstream or downstream production linkages to the agricultural
sector.
The nonagricultural activities upstream of agriculture are
limited almost exclusively to animal husbandry and blacksmithing
(production and repair of tools, for example). This reflects the
relatively low use of inputs in Senegalese agriculture.
Downstream
production links are primary food processing or marketing activities.
In addition to the production links, one must also consider the
indirect consumption links.
As agricultural incomes increase, the
demand for many locally produced goods and services will also increase
(tailors, crafts, and "griots", for example).
In
conclusion,
the
increasing
importance of
income
diversification in the rural areas doesn't exclude the need to improve
agricultural productivity.
Results presented above highlight the fact
that many nonagricultural
activities are tied directly (through
production linkages) or indirectly (through consumption linkages to
the agricultural sector.
The intersectorial links described here
require
additional
study to better understand and quantify the
interaction of the agricultural and nonagricultural sectors in the
IFPRI/ISRA study zones.'"
IMPORTANCE OF THE INFORMAL SECTOR IN URBAN ZONES
In urban zones, formal activities are defined as originating from
a public or private enterprise, with a legal statute, that allows the
State to obtain information on the firm's operating budget. These
activities are therefore more easily accounted for in the national
accounting system and estimates of national income. On the other hand,
many informal activities escape the national accounting system, due to
the fact that they are not legally structured and can easily avoid
reporting income and paying taxes.
l5 In town, doughnuts are generally prepared with wheat flou r;
but in rural zones,
it is possible to find doughnuts made from mill et
flour.
l6 More systematic analysis of indirect consumption linkages
the IFPRI/ISRA study zones is currently being conducted w Enh
supplemental funding obtained from USAID/Washington.
7 6
As was mentioned in the preceding section, services and trade are
key sectors of activity in the composition of urban income. However,
the nature of activities in the two towns in the study present
variations.
It is possible to note that the service sector in Kaolack is
composed essentially of room rental, "marabout", "griot", tailors,
electricians,
guard services,
and transport (driver) services. The
majority of these activities are in the informal sector (76% of the
service sector). Trade activities in Kaolack are dominated by shop
tending,
sale of food, fish, ice, cloth and used clothing. The trade
activities in Kaolack are almost exclusively informal (95%). This
shows clearly that in Kaolack the informal sector dominates income
generating activities of households.
In
Tambacounda,
the
service
sector is
represented by
administration activities,17 transport (drivers),
"marabout", maids,
and room rental. More than 70% of these activities are from the
informal sector.
Trade is composed especially of activities in the
informal sector such as trade of condiments, charcoal,
fruit and
tobacco.
PATTERNS OF HOUSEHOLD INCOME DIVERSIFICATION
Knowing what factors are driving households to diversify income
and the long-run
economic
implications
of existing patterns of
diversification is an important prerequisite to designing rural
development policies that will foster growth in rural incomes and
increased agricultural production.
Furthermore,
understanding income
diversification strategies of rural households permits one to better
judge household-level response to proposed policy changes.
To better
grasp these elements, a topology of income diversification patterns
and a
model
of factors driving
households to diversify were
developed."
" Receipt of retirement benefits, teachers, and other civil
servants.
'* There is an extensive literature on income diversification in
Asia,
but evidence from Nigeria (Matlon), Burkina Faso (Reardon,
Delgado, and Matlon), and Niger (Hopkins and Reardon), suggests that
there are some important differences
between Asian and Sahelian
diversification patterns.
The topology and model presented here are
specific to
Senegal
and
based on
survey
findings;
the
conceptualization of the different factors to consider in the topology
and model drew heavily on the analyses of the other West African
research cited above.
‘I
.
*
I
.
:
.’
:
.
.
:
.:
*
77
Toooloqv of diversification patterns
The topology provides a framework that facilitates the tasks of
(1) diagnosing the underlying causes of the diversification in each
zone, (2) understanding what factors are motivating production and
consumption behavior in each zone, and (3) anticipating the type of
household response that different policy interventions are likely to
elicit.
Three major styles of diversification have been identified:
(1) an outward-oriented
"copinq" pattern characterized by:
declining levels of cropping income over time due, in
large part, to a poor, deteriorating natural resource
base;
a dearth of activities with production and consumption
linkages to agriculture
an important share of household income coming from off-
farm activities only in emergency situations following
severe crop shortfalls
consistently low total household income due to poor
cropping outcomes and limited off-farm opportunities;
(2) an outward-oriented, lonq-run, adaptive pattern characterized
by:
declining levels of cropping income over time (as in the
"coping"
case) or low levels of cropping income because
technological
transformation has not yet taken place
and/or price risk is high due to poorly developed markets
and physical isolation from centers of demand.
nonfarm income often earned outside of the zone of
residence;
little reinvestment, if any, of nonfarm income into
cropping activities;
total incomes that are low but generally adequate (higher
than those for the "coping" case)
in the long-run, households may move out of crop
production entirely;
(3) an inward-oriented, lonq-run, adaptive pattern, characterized
by:
good
cropping
performance
stimulating
growth of
production and consumption linkages;
*
..’
*.
,
.
:.
.
.
.
:
.
‘.
_’
78
off-farm
activities
that
complement
rather
than
substitute for cropping income;
income from off-farm activities that is reinvested in
crop production;
the presence of well developed transportation and market
infrastructure
a policy environment that fosters development of factor
markets (particularly credit and labor).
The "copinq" pattern is found in the Niakhar zone of the central
Peanut Basin.
Households in this zone had the lowest average incomes
in the study during both survey years.
Following the extremely poor
1988/89 harvest, off-farm income was 53 percent of total income. It
was earned primarily from petty commerce of nonlocal products (sugar,
tobacco, beer) and migration to Dakar.
Off-farm activities in this
zone do not have many production linkages to local
cropping
activities.
Following the better 1989/90 harvest, cropping income
doubled and off-farm income dropped to only 20 percent of total
income.
Even with the doubling of cropping income, however, total
income remained low.
Households in this "coping"
zone have not yet developed an
adaptive strategy.
Crop land is a serious constraint preventing
households from expanding crop production through extensive practices
and also limiting the size of livestock holdings that can be kept in
the zone year-round.
Rainfall has declined over time and soil
degradation has increased, eroding the zone's agricultural resource
base.
Overall low levels of income do not permit productivity
enhancing investments in either farm or off-farm activities.
If nothing is done to improve the physical environment in this
zone, households will be forced to develop a more aggressive type of
outward-oriented diversification through migration to urban areas or
other zones.
The extent to which future policies should encourage the
continuation or intensification of cropping and livestock activities
in this zone cannot be answered by the IFPRI/ISRA data base as one
needs to evaluate the capacity of the zone's natural resource base to
support agricultural intensification.
An example of the outwardly-oriented, lonq-run, adaptive pattern
is found in the northern Peanut Basin, where cropping potential is
extremely
low
due to
poor
rainfall
and
degraded
soils.
Diversification is based to a large extent on migration and provision
of products and services to Dakar.
There are, however, some cases of
"inwardly-oriented"
activities that appear to have production and
consumption links to the local livestock sector.
An important share
of income comes from sales of household livestock.
Income from sales
of animal byproducts is also important as are commissions to sales
agents who transport and market local livestock in Dakar.
79
Incomes are generally adequate in this type of zone and most
households manage to assure food security,
even in years of severe
crop failure.
Another zone that appears to be moving toward an outwardlv-
oriented strateqy is Senegal Oriental.
In this case, however, the
strategy does not appear to be well adapted to the zone.
Senegal
Oriental has promising cropping and livestock potential given high
average rainfall,
less degraded soils than the Peanut Basin, and
abundant supplies of crop and pasture land due to low population
density.
Nevertheless, households earn a relatively low cropping
income ,per AE and a high share of "outwardly-oriented" off-farm
income.
There are many indications that market failures and lack of
infrastructure
are
constraining
agricultural
productivity
and
unnecessarily pushing Senegal Oriental's households into migration (22
percent of household income in 1989/90).
The low 1989/90 cropping
income was due in part to a cotton boycott.
The boycott was to
protest high input prices and low net returns for cotton, suggesting
the existence of some market or price policy failure.
*
Cereal sufficiency ratios in Senegal Oriental were higher in
1988/89 than those in several zones where larger shares of total
cereal production were sold,
This suggests the possibility of market
failure or inadequate transportation infrastructure.
Because migration remittances are being invested mostly in
housing and mosques and rarely in agricultural production,
one
surmises that households do not view agriculture as a good investment.
The lowest overall returns to family cropping labor were found in
Senegal Oriental.
This implies that migration is probably the best
income strategy at present.
Senegal Oriental clearly has agricultural potential. The policy
issue here is how to design policies that will encourage households to
invest in the intensification of agriculture rather than in outward-
oriented off-farm activities, particularly migration.
An inwardly-oriented adapted strateqy is currently found in the
two zones in the southern Peanut Basin and Colobane in the central
basin.
These zones have relatively low levels of migration income.
Much of the off farm income in these three zones is linked to crop
production activities
(peanut marketing
and processing, cereal
marketing, restaurants).
Animal husbandry is also an important source
of revenue in Colobane and the southeast but not in the southwest.
" The zone had the lowest average cropping income in 1989/90 and
the third lowest in 1988/89.
1
.I
80
The challenge
in these zones is to provide the incentives
necessary to improve agricultural productivity.
Without growth in
agriculture the impetus for growth in the local nonfarm sector will be
lost.
If this happens, the prevailing credit constraints will be
magnified because reinvestment of off-farm income in crop production
will decline.
Determinants of off-farm income shares
Understanding why diversification off-farm is taking place is
important to policy
analysts because households will
respond
differently if the factors pushing them to diversify are different.
Information about factors driving diversification can be used to
develop nonprice policies that will speed up and increase market
response to price signals.
It is relatively easy to understand the role of climate and
natural resource constraints in pushing households into "outwardly-
oriented" off-farm activities in zones such as the northern basin. It
is more difficult to assess the role of inadequate infrastructure,
credit constraints and other market failures.
Regression analysis was
used to gain a better understanding of the relative importance of
these factors.
Households located in market villages earned 12 percent more of
their income from off-farm sources than households in other villages
(approximately a 60 percent increase over average shares of 18 percent
for nonmarket villages).
The larger share for market villages can be
explained by "pull" factors -- access to market infrastructure draws
households into off-farm activities because the infrastructure makes
it easier to find employment and easier to sell goods and services.
Factors associated with the agroclimatic zone also determine the
share of off-farm income.
Households in the southwestern Peanut
Basin, earned a statistically significant greater share (-t 14 percent)
of their total income from off-farm activities than other zones. This
represents a 70 percent increase over average non-farm shares of 20
percent for other zones.
Higher nonfarm income in the southwest is explained by a number
of "pull" factors.
Proximity to the very dynamic markets in Kaolack
and the Gambia encourage individuals to engage in petty commerce
activities.
The relatively dynamic agricultural sector in the zone
increases the opportunity for employment at peanut collection points
and in cereal marketing activities.
This zone also has the second
highest level of average household income in the IFPRI/ISRA study; the
higher income is no doubt stimulating the demand for nonagricultural
goods and services."
" Coefficients for market and zone were significant at, the .01
and .06 level, respectively.
._
.
.
.
.
.
:
81
"Push" factors such as poor access to land and credit, shortfalls
in cereal production,
and inadequate levels of household savings
(represented by livestock assets) were also statistically significant
variables explaining the share of off farm income in total household
income."
Elasticities estimated for these variables suggest that:
*
A 10 percent increase in the cereal sufficiency ratio decreases
the share of off-farm income by 7 percent;
*
A 10 percent increase in land assets (proxied by cultivated area)
decreases the off farm share by 5 percent;
*
A 10 percent increase in livestock assets decreases the off-farm
share by 3 percent;
*
A 10 percent increase in agricultural and food credit decreases
the off-farm share by 1 percent.
The elasticity for the cereal sufficiency ratio is large and
negative,
as expected.
Households better able to meet food security
needs from home production have
less incentive to seek off-farm
income,
all else being equal.** This suggests
that programs or
policies which reduce food insecurity (food aid, food banks, or
credit,
for example) could encourage households in zones with good
agricultural potential to concentrate more on agricultural production.
The same types
of programs in zones
with an outwardly-oriented
adaptive pattern of diversification or a coping pattern would be
unlikely to elicit greater agricultural productivity.
The elasticity for land assets is also large and negative as
anticipated.
All else equal, more land permits households to earn
more cropping income by producing the desired mix of cereals and cash
crops; higher cropping income reduces the pressure to move off-farm to
supplement income.
This result is not easy to translate into policy
because it involves land tenure issues that were not covered by the
study.
The implication is, however, that larger farms should be
encouraged if the objective is to reduce unnecessary diversification
that reduces productive efficiency.
Larger herd size also decreases the share of off-farm income
because livestock provide savings that are easily converted to cash
when households need liquidity for food or input purchases, reducing
the need to move off-farm.
Policies to improve animal health and
water supplies (a major constraint to animal husbandry in several
zones)
when the carrying capacity permits larger herd sizes would be
2' These effects were all significant at the .Ol level or better
except the credit variable which was significant at the .08 level.
22 Cereal sufficiency is defined relative to beginning stocks of
cereals, avoiding problems of reverse causality.
.
.
_ .
.
i
I
*
*
,’
82
most likely to diminish off-farm diversification and promote more
integrated cropping/livestock systems in zones with inwardly-oriented
diversification patterns.
Such policies would be unlikely to improve
crop productivity in zones with coping or outwardly-oriented, adapted
strategies.
Although the size of the elasticity for credit is smaller than
that for other factors, it is an extremely important variable from a
policy perspective.
The credit elasticity implies that better access
to credit permits households to increase the share of cropping income
in total income (that is, it diminishes the need to seek off-farm
income).
This suggests that improving access to credit might permit
households to use labor which is now engaged in off-farm activities to
improve farm productivity -- planting trees or building windbreaks,
for example.
In sum,
the significant coefficients on three of the above
variables -- cereal sufficiency ratios, livestock holdings, and access
to credit -- provide ample support for the hypothesis that growth in
agricultural productivity is being constrained by liquidity problems
that push farmers into greater income diversification than would be
the case were institutions for credit and savings better developed.
Although these regression results provide some insights into the
nature of rural income diversification in the Peanut Basin, the model
did not directly address the links between risk and the share of off
farm income in total income.
Additional work on the risk aspects of
diversification is clearly needed.
SUMMARY OF MAIN POINTS
The discussion in this chapter has stressed the importance of
both the agricultural and nonagricultural sectors in the income
generating strategies of rural households.
While the agricultural
sector continues to provide the majority of income in all zones but
the northern Peanut Basin, income from cropping alone is frequently
not sufficient to meet minimum needs.
The food security and general well being of rural households in
the IFPRI/ISRA study zones depends on a balanced program of rural
development that stimulates economic activity in both the agricultural
and nonagricultural sectors.
The extent of production and consumption
linkages between the two sectors found in most zones suggests that
improved agricultural productivity can have positive repercussions on
incomes earned in nonagricultural activities.
Inadequate credit and
infrastructure (commercial and transport) have been identified as
factors that are preventing households from realizing their full
agricultural potential, particularly in Senegal Oriental.
Household income generating strategies were found to differ
significantly
across
zones.
The
existence of
different
*.,j
.
.‘
.j
:
.
.
:
‘_
83
diversification strategies implies that successful rural development
policies will need to respond to the particular circumstances found in
each zone.
This point will be developed further in the concluding
chapter of this report.
The most important quantitative findings presented in this
chapter follow:
*
Across all zones, rural household income is most often (i.e., the
modal range) in the 30,000 to 60,000 FCFA/AE ($100-200) range
following both good and poor harvests; 36 percent of households
in HY88 and 22 percent in HY89 earned less than this amount -
most of these households failed to cover minimum needs.
*
The modal income category for urban households was 2 100,000
FCFA/AE; approximately 40 percent of households were at this
level; only 7 percent of urban households earned less than 30,000
FCFA/AE.
*
Equality of income distribution across households within a zone
is typical of other West African cases
analyzed (Gini
coefficients in
.30 to .35 range) except in the southwestern
basin,
Senegal
Oriental,
and
Kaolack where
the
income
distribution is more skewed (Gini > .40).
*
Highest rural household incomes are found in the southern Peanut
Basin during both survey years.
*
Lowest rural household incomes are found in the Niakhar zone of
the central basin and in Senegal Oriental.
*
The share of nonagricultural (noncrop, nonlivestock) income in
total income is highest in the northern Peanut Basin (about 70
percent) and in the southwestern basin (about 50 percent);
*
The share of agricultural (cropping and livestock sectors) income
in total income is highest in the southeastern Peanut Basin and
the Colobane zone of the central basin (about 80 percent).
84
REFERENCES
Fall, A.A. 1991.
Composition multisectorielle et distribution du
revenu dans le bassin arachidier: Remuneration du capital et du
travail, Memoire de confirmation a l'ISRA, et document no. V du
projet ISRA/IFPRI, janvier, Dakar.
Fall, A.A., V. Kelly, and T.A. Reardon.
1989. Household-level survey
methods used in the IFPRI/ISRA study of consumption and supply
impacts of agricultural price policies in Senegal. Document 2,
IFPRI, Washington, D.C.
Haggblade, S., P. Hazell, and J. Brown. 1989. Farm-nonfarm linkages in
rural Sub-Saharan Africa, World Deve?opment, Vol. 17, No. 8.
Hopkins, Jane and Thomas Reardon. Aqricultural Price Po1ic.v Reform
Impacts and Food Aid Tarqetinq in Niqer. Washington, DC, IFPRI,
April 1993.
Reardon, Thomas.
“Income Diversification in Burkina Faso, unpublished
mimeo". 1988.
Ndiaye, 0. 1993.
Micro-level Small Ruminant Transaction Anal.vsis and
Livestock Policv in Seneqal.
Masters Paper,
Michigan State
University, E. Lansing.
Kelly,
v. 1993.
"Aspects Economiques de la Production et de la
Commercialisation des produits Agricoles au Niveau des Menages du
Bassin Arachidier et du Centre du Senegal Oriental".
Workshop
paper presented at l'atelier du Project IFPRI/ISRA/BAME sur
"Politiques Agricoles et Comportements des Menages dans le Bassin
Arachidier et au Senegal Oriental", LNERV/Dakar Hann, 29-30 Mars.
Reardon,
Thomas,
Christopher L.
Delgado,
and
Peter
Matlon.
"Determinants and Effects of Income Diversification amongst Farm
Households in Burkina Faso".
Journal of Development Studies.
January 1992.
I
.,
:
.
I
:
.:
.
.
85
5. CROP PRODUCTION
INTRODUCTION
Chapter 4 described
income-generating patterns
for sample
households.
Particular attention was paid to the relative importance of
net cropping income in total income.
In this chapter we look more
closely at crop production
activities,
addressing the following
questions:
What is the relationship between gross and net cropping income?
To what extent does cropping income cover minimum needs?
What are the land and input use patterns and costs?
What are the returns to labor?
How much does crop production contribute to women's personal
income?
To what extent is coarse grain production satisfying household
cereal needs?
In answering these questions, differences across zones, and crops
are discussed.
METHODOLOGY
Each field cultivated by sample households for the production years
1989/90 was enumerated'.
Data were collected on all purchased inputs
(usually fertilizer, insecticide, or herbicide), all inputs coming from
home production (usually seed and manure), and all labor inputs (family
labor and hired labor).
The quantity of inputs used and the price (if
purchased) were enumerated (except for manure).
Labor inputs were recorded for each of the major activity periods
(clearing,
land
preparation,
planting,
weeding,
harvesting,
transporting).
If labor was not household labor, amounts paid in cash
or in kind were recorded.
All labor inputs were divided into broad
gender and age categories.
' Although crop production data were calculated for HY90, there
is no corresponding income or expenditure data for this harvest year %
because the rural survey ended in December 1990, consequently, no
analysis of HY90 crop production data is presented in this report.
1
..;
.
:
*
8 6
The number of harvest units was enumerated for each field. Total
harvest was estimated by weighing two to three samples of the harvest
unit (a bundle for cereals , a sack for peanuts) and multiplying this by
the total number of units harvested. All fields were measured.
Gross income is calculated for each field by multiplying total
quantities harvested by the average producer price for each product and
zone.
Net income is calculated for each field by deducting the cost of
all purchased inputs and the imputed value of all home-provided inputs.
Returns to family labor are calculated by dividing the total number of
hours worked on a field (all activities combined) into the net revenue.
No allowance has been made for depreciation on animal traction
equipment.
An analysis of depreciation costs for the northern Peanut
Basin (Dakono 1991) showed that most equipment was so old that it had
already been fully depreciated.
For this reason, the depreciation
analysis has been dropped.
In the future, it may be appropriate to
examine depreciation in the southeastern Peanut Basin and Senegal
Oriental where traction equipment has been introduced more recently than
in the older part of the Peanut Basin.
Total household income per adult equivalent is obtained by summing
the income across fields within the household and dividing by the
average number of AEs present in the household for the harvest year.
Average crop income per AE for each zone is obtained by averaging across
all households and weighting by household size and a coefficient used to
adjust for nonproportional sampling of market villages.
Although data on total quantities harvest in HY88 are available, no
input data were collected because the survey began at harvest time
(October/November 1988).
Net cropping incomes for HY88 presented in
Chapter 4 were estimated by assuming that the HY88 ratio of input cost
to output value was the same as the one we calculated for HY89. This
provides only a rough approximation of net cropping income for HY88.
Given that all other income data are reported as net income, it was
deemed appropriate to present at least an approximation of net HY88
cropping income.
Because the analysis of input use and returns to labor presented in
this chapter relies entirely on HY 89 data, the results present more
rosy picture of economic returns than would be the case if both HY88 and
HY89 input data were available because the 1989 harvest was unusually
good while the 1988 one was relatively bad.
The reader is reminded that the income per AE data presented in
this chapter
will differ somewhat from that presented in Chapter 4
because we use, data for 177 households that had complete cropping
information rather than the more limited set of households that had
complete income information.
Table 5.1 summarizes data presented in other tables throughout the
chapter.
.
..;
:
.
‘:
,_
.’
Table 5.1 -
Cereal/pulse production sufficiency ratios by zone and harvest year
n?g
E;pBR
:F?!!?
SwEg
sEJ?J3
. . . ..,.,
w
Average Gross Incane (FCFA/AR)
16760
29550
59750
40870
55320
20805
Average A& Income (FCFA/AE)
13170
23800
48030
35010
46410
17440
Input Costs as Share of Gross Income
(2)
2 1
2 0
20
1 4
1 5
1 6
share of Ret Income by crop
Peanuts (2)
65
39
65
43
68
52
Cereals (X)
30
61
3 5
2 9
40
other (X)
5
0
0
0
3
8
Share of Uininnsn Reeds Covered by Ret Crop Income
(I)
50
82
249
173
244
78
Median Amber of Hectares cultivated
Per Household
4.3
4.5
1 1
7
14
5
Per AE
-49
-5
1.21
-75
1.1
.49
Madian Returns to Family Labor (FCFA Per 8 hr person-da7)
Peanuts
870
530
1460
1380
720
610
Cereals
160
1010
1100
900
640
510
Share of Croppin(; Income by Gender (X)
wale
9 0
85
86
88
84
68
Female
8
13
12
1 0
1 2
23
Cereal Production Sufficiency Ratios
1988/89
.14
-47
.50
-58
.83
.70
1989/90
-25
.76
.98
1.03
1.02
-69
. . .
Ave. kilos of Cereal Surplus par AI?
1988/89
(1
(1
1
2
1 3
56
198&90
(1
15
6 9
4a
58
31
88
GROSS AND NET CROPPING INCOMES
Table 5.2 illustrates the difference between gross and net income
and the relative importance of peanuts, cereals, and other products
(cowpeas in the north and cotton in the southeastern Peanut Basin) in
total cropping income.
Despite higher input costs (due primarily to
high seeding ratios), peanuts account for about 65 percent of net income
in three zones (the northern basin, the Colobane zone in the central
basin, and the southeastern basin).
Cereals are dominant in two zones -
- Niakhar in the central basin and in the southwestern basin. In
Senegal Oriental households boycotted cotton in 1989/90 giving both
peanuts and cereals a larger role than usual (about 50 percent of income
comes from each crop).
Table 5.2 -
Canparison of average gross and net cropping income by crop for 1989/W
(FCFA/AE)
ZCMES:
NPB
C P B N
CPBC
SWB
SEPB
so
Peanuts:
Gross income
11900
14890
42630
20210
39700
11900
Net income
8590
9320
31070
14960
31690
9111
Cereals:
Gross income
4150
14660
16750
20130
14060
7425
Net income
3940
14480
16600
19890
13440
7024
Other crops:
Gross income
710
0
370
320
1560
1478
Net income
630
0
360
170
1270
1305
Share of peanuts
in net income:
.65
.39
.65
.43
.68
.52
Source :
IFPRI/ISRA Survey
.i
‘:
.
‘.
.’
89
It is useful to examine the extent to which net cropping income
covers minimum needs.
Minimum needs are defined as the cost (using
zone-specific prices) of 1900 kcal/AE of coarse grains plus a margin of
20 percent to cover other food and non food essentials. The cost of
minimum needs, defined in this manner, provide us with some measure of
the relative well being of households across zones with different costs
of living.
Table 5.3 shows that following the above average 1989/90 harvest,
only half of the zones (Colobane in the central basin, and the two zones
in the southern basin) earned net cropping incomes that covered the
estimated cost of minimum needs. In the northern Peanut Basin and in
Senegal Oriental, cropping income at the 75th percentile was below the
IFPRI/ISRA estimate of minimum needs (see Table A5.1 for a full
breakdown by percentile).
Note:
Cereal price data for 1989/90 is not available for the northern Peanut Basin so
we usa the 1968/89 minimum needs for 1989/90.
Table 5.3 - Levels of net cropping income and the cost of minimm needs
ZONES:
(FcFA/AE)
NPB
CFBN
CPNC
SWPB
SEPB
So
cost of
minimum
26055
29310
19245
20135
18950
22230
needslaelyr
Average net
cropping
income/ae/yr
13170
23800
48030
35010
46410
17440
Share of minimum
needs covered (2) 50
82
249
173
244
78
Source :
IFPRI/ISRA survey data.
9 0
LAND CULTIVATION AND INPUT USE PATTERNS
The average amount of land cultivated per household in 1989/90
varies significantly across zones.
The smallest farms (median size of
4 to 5 hectares) are in the northern basin, central basin (Niakhar), and
Senegal Oriental.
The largest median farm sizes are in the central
basin (Colobane) and southeastern basin (11 and 14 hectares,
respectively).
In the southwestern basin the median size is about 7
hectares.'
Cultivated area per adult equivalent is less variable than total
farm size, averaging from .49 hectares in the northern basin to 1.21 in
the Colobane zone of the central basin.3
Results from village
reconnaissance surveys (see Chapter 2) suggest that the relatively small
farm size in Niakhar is due to a serious land constraint, while in
Senegal Oriental, the lack of equipment and heavier soils make farming
large areas more difficult than in the Peanut Basin.
Peanut seed is clearly the most important variable input in terms
of value.
In general, about 50 percent of its total value comes from
household stocks while the rest is purchased. The central basin zone of
Colobane exhibits the highest per adult equivalent level of purchased
peanut seed (over 5000 FCFA/AE vs. less than 2000 FCFA/AE for other
zones).
Very little cereal seed is purchasede4
Use of fertilizer is limited primarily to contract cotton and
confectionery peanut production where credit can be obtained; 80 percent
of cotton fields and 100 percent of confectionery peanut fields received
fertilizer.
No fertilizer was used by sample households in the
northern and the central Peanut Basin.
In the southern Peanut Basin
zones, 4 to 5 percent of cereal fields were fertilized while the share
increased to 9 percent in Senegal Oriental. Greater fertilizer use in
these zones is related to better rainfall and, in some cases, proximity
to the Gambia.'
* Median farm size is used in lieu of means because the
distributions are skewed due to a few exceptionally large farms in
each zone. Table A5.2 presents more detail on the distribution of farm
size across zones and households.
3 Mean and median values of area per AE are quite similar in all
zones but the southeastern basin.
4 The seed purchase data used to calculate the shares of seed
value purchased cover only cash purchases, seed purchased on credit
through the confectionery peanut and maize contracts is not factored
in.
5 These zones are close enough to the Gambia to be able to
.
purchase lower priced Gambian fertilizer. Price data suggest that
this was frequently, but not always, the case.
.
’
‘.
:
91
Use of manure is somewhat more common but still not very important
in terms of total cultivated surface areas.
Manure was used on peanuts
in the northern and central basin (1-2 percent of fields) and on
confectionery peanuts in the southwest (14 percent) of fields. No
manure was used on cereal fields in the north while 5-8 percent of
cereal fields received some manure in the central and southern zones.
In Senegal Oriental 31 percent of millet/sorghum fields and 48 percent
of maize fields were manured."
With the exception of cotton production (which demands substantial
f amounts of herbicide and insecticide), the only other chemical input
used in large quantities was fungicide for peanut seed.
Hired labor falls into two categories: temporary (hired on a daily
basis) and seasonal ("navetane" employed for an entire season).
"Navetanes"
are most common in the Southeastern basin and Senegal
Oriental where about 25 percent of households use them. Households that
employ "navetanes" in these zones tend to hire several at a time. In
other zones, fewer than 10 percent of households contracted a seasonal
laborer during the survey period.
Use of day laborers is not substantial in most zones. Most hired
labor was used at harvest time with smaller amounts for land clearing,
planting, and weeding.
Payments for hired labor are extremely variable
making it very difficult to calculate an average wage rate for each
zone.
The low levels of hired labor and the absence of a consistent
wage rate suggest that rural labor markets are not functioning very
well.
Table A5.3 in the appendix provides supplementary details on
average and median expenditures by crop for the three major input
categories (seeds, fertilizer and chemical products, and hired labor).
No attempt has been made to impute costs to seasonal laborers (usually
paid inking with room and board) and manure.
RETURNS TO FAMILY LABOR AND LAND
"Returns to labor" is defined simply as net cropping income per field
divided by the number of 8-hour labor days worked on a field. No
distinction is made for different categories of laborers (male, female,
child) and no allowance is made for returns to capital or management.
Only the principal crop has been valued. Because the value of crop by-
products (particularly peanut hay) has not been taken into account the
returns to peanut production are really higher than those shown.
6 Percentages are percent of fields and not percent of
cultivated areas.
As the smaller household plots tend to be the ones.
receiving manure, the share of cultivated area receiving manure is
most likely much lower.
: ;
1
.
92
Table 5.4 presents the median values per 8-hour work day while Table
A5.4 in the appendix prov ides details on returns at the 25th, 5Oth, and
75th percentiles7.
Table 5.4 -
Median returns t .o family labor in cropping activities
for 1989/90
(FCFAD-hour person day)
Cereals
160
1010
1100
900
640
510
Other crops
2.50
0
245
70
90
360
Source:
IFPRI/ISRA survey
Median returns to an 8-hour workday in household crop production
ranged from 150 to 1500 FCFA/day, depending on crop and zone, during the
relatively good 1989/90 season.
The higher end of these returns
compares favorably to the SMIG (official minimum wage in the formal
sector) of 1500 FCFA/day.
Analysis of returns to family labor shows that there is little
incentive for most farmers to increase cereal production given current
technology and price relationships.
Cereal production was significantly
more remunerative than peanut production (almost double) in only one
zone -- Niakhar in the central Peanut Basin.* In all other zones median
returns to labor in peanut cultivation were higher (12 to 500 percent)
than those for cereals.
In three zones (the northern Peanut Basin,
Niakhar in the central Peanut Basin, and the southwestern basin) the
difference in mean returns to labor for the two crops was statistically
significant at the .05 level.'
All zones but the southwestern basin
7 Numbers in this table supersede those presented in Kelly
(1993).
Some programming errors that resulted in miscalculation of
total household hours worked have been corrected.
a Numerous explanations for the higher returns to millet in this
zone can be hypothesized: climatic and pest patterns during the study
period that were more favorable to millet than to peanuts, poor
quality peanut seed, or farming techniques that favor millet over
peanuts.
Further analysis of the IFPRI/ISRA crop production data and
additional field work are necessary to fully understand why this zone
appears to be different.
' In the Colobane zone of the central basin, the difference was '
significant only at the .15 level.
93
appear to be giving more attention to the crop that provides the highest
returns to labor (see Table 5.2)".
Although we have not yet analyzed returns to land, evidence from
other sources suggests that peanut crops are more profitable per unit of
land than cereals (Martin 1988, Kelly 1988). This is due primarily to
much greater yields per hectare for peanuts grown without fertilizer
(800-1000 kilos) than for cereals (400-600 kilos).
DISTRIBUTION OF CROP INCOME BY GENDER
There is a great deal of interest in the distribution of income
by gender.
It is difficult to talk very precisely about distribution of
cropping revenue by gender in Senegal because there are fields that are
said to belong to the household head (usually male) but that are clearly
farmed by all family members and whose products are consumed by all
family members --
these are the cereal fields grown under the
supervision of the household head.
Ignoring this difficulty, Table 5.5
assumes that the cropping income for each field goes to the person
supervising the cultivation of that field. The table shows the share of
gross income earned by gender for each zone". Although the crop
Table 5.5 -- Distribution of crop income by gender
Northern
Central Basin
Southwest Southeast
Senegal
Basin
Niakhar
Colobane
Basin
Basin
Oriental
Year
1989/90
Percent of total cropping income
Male
90
a5
86
88
84
68
Female
a
13
12
10
12
23
Source:
IFPRI/ISRA project survey.
Note:
When percents do not sum to 100, it is because there are some cases of incorrect or
missing codes for the gender of the person supervising the field.
" We have not yet had the time to examine in-depth factors that
are causing the returns to labor to be different across zones. It
seems important to determine whether differences are a result of
factors such as soil quality and rainfall or because patterns of labor
allocation differ from zone to zone.
" Numbers in Table 5.5 supersede those in the draft report which
incorrectly showed the share of cereal income rather than share of
total cropping income.
_’
*
.
.’
.
94
breakdown is not presented, over 90 percent of women's fields in the
Peanut Basin are peanut fields; cowpeas are usually the second most
common crop.
In Senegal Oriental, women cultivate rice and fonio in
addition to peanuts and cowpeas.
Chapter 6 discusses the share of different crops marketed by
gender.
Although Table 5.5 shows women's share in total crop production
to be quite small, a fairly large share of what women produce is
marketed, giving them a more important role in crop sales than one would
suspect from the production data alone.
CEREAL PRODUCTION SUFFICIENCY LEVELS
As calculation of cereal production sufficiency levels requires
only harvest data, this section of Chapter 5 presents results for both
HY88 and HY89.
In evaluating cereal production sufficiency we use a
level of 1900 kcal/AE as the standard requirement'*. Given fairly high
levels of pulse consumption in Senegal and the not insignificant role
that vegetable oil, milk and vegetables play in rural diets, we use 1900
kcal rather than minimum caloric requirements (2400 kcal) because it
better reflects real demand for cereals,
Table 5.6 illustrates that in 1988/89 cereal production fell very
far short of the 1900 kcal standard in all zones.
Production
sufficiently ratios13 were less than .5 in three zones and between .6
and .8 in the remaining three.
Table 5.6 -- Average cereal product ion sufficiency rat ios for 1988/89
and 1989/90 by zone
zone
Northern
Central Basin
southwest
Southeast
S e n e g a l
Peanut
N i a k h a r Colcbane Peanut
Peanut
Oriental
Basin
Basin
Basin
1988189
.I4
.47
.50
.5a
.a3
.70
1989190
.25
.76
.9a
1.03
1.02
.69
Source:
IFPRI/ISRA survey.
Note:
The cereal production sufficiency ratio is the amount of cereal produced by the household
divided by the amount required to provide 1900 kcal/AE for the entire household.
'* The draft version of this report used a variety of assumptions
for calculation production sufficiency levels. More detailed analysis
has convinced us that the 1900 kcal level is the most appropriate one
to use so references to other levels have been eliminated.
l3 The cereal production sufficiency ration (CPSR) is the
quantity of cereals produced divided by the quantity of cereal
required to provide each adult equivalent in the household with 1900
kcal of cereal.
.
:
.,
.
:.
.
.’
:
9 5
The same table shows that 1989/90 cereal production improved
greatly.
Three of the six zones had an average cereal production
sufficiency ratio (CPSR) in the neighborhood of 1 (Colobane in the
center and the two zones in the southern basin) while two others were in
the .6 to .8 range.
The northern basin was the only one still
exhibiting a very large deficit.
Table A5.5 in the appendix provides
details of CPSRs at different percentiles.
Table 5.7 shows estimates of the cereal surplus per AE averaged
across all sample households for each of the two study years. In 1988/89
amounts were insignificant except in the southeast (13 kilos/AE) and
Senegal Oriental (56 kilos/AE).
The situation improved in 1989/90 where
amounts available were in the 50 to 70 kilo range for three zones
(Colobane in the central basin and the two zones in the southern basin)
and in the 15-30 kilo range for two (Niakhar in the central basin and
Senegal Oriental).
Table 5.7 -- Average kilos of surplus cereals per AE:
1988/89 and 1989/90
ZONE:
NPB
CPBN
CPBC
SYPB
SEPB SO
198ala9
MEAN
~1
<l
1
2
13
5 6
1989/90
MEAN
~1 15
69
48
5 8
31
Source : IFPRI/ISRA survey.
Information is provided in appendix Table A5.6 for cereal surplus
available at the 25th, 50th, and 75th percentile.
The surplus in both
years was zero up through the 50th percentile -- i.e., at least half of
sample households had no surplus, even following the better 1989/90
harvest.
The bulk of surplus production is coming from a small set of
households.
Using two by two contingency tables and chi-square tests of
significance,
we attempted to identify factors that distinguish
households with large surpluses from others.
In general, the 18
households with surpluses in 1988/89 were more likely than others to
exhibit the following characteristics:14
- total cultivated area > 6 hectares
- cultivated area per AE > 1 hectare
- in the highest income category (tercile)
- earn greater than 40 percent of income from cropping
l4 "More likely" means that there was a statistically significant
difference -- usually at the .Ol level of significance or better but
always better than .lO..
96
- reside in Southeastern basin or Senegal Oriental
- reside in a nonmarket village
- rice provides less than 10 percent of cereal calories
Factors that did not seem to explain the level of surplus in
1988/89 were level of caloric adequacy, dependency ratios, and ethnic
group.
Following the better 1989/90 harvest 52 of the sample households
had a surplus.
These households were more likely to have the following
characteristics:
total cultivated area > 6 hectares
cultivated area per AE > 1 hectare
*
be in the middle and highest income category (tercile)
earn greater than 40 percent of income from cropping
*
reside in the southeastern basin and to a less extent in
the southwestern basin and the Colobane zone of the
central Peanut Basin
reside in a nonmarket village
*
rice provides less than 25 percent of cereal calories
*
achieve caloric adequacy levels (2400 kcal/AE)
An li*' marks those characteristics that changed between the two
years.
Caloric adequacy is a factor that becomes significant with more
than 65 percent of those with surplus consuming more than 2400 kcal/ae
per day.
Despite the significance, there are still 18 households with
surplus that do not appear to have adequate levels of caloric intake.
Income, rice consumption, and zone are factors that appear in both
years, but following the better harvest those with surpluses are less
concentrated in the extreme categories.
For example, higher income is
associated with surplus; but unlike 1988/89 those with surplus are found
in larger than expected numbers in both the middle and upper income
groups while those with no surplus are concentrated in the poorest
income category.
SUMMARY OF MAIN POINTS:
The important points to retain about crop production are:
e
Net cropping income fails to cover the costs of minimum househol‘d
needs in 3 of the 6 zones studied, even in a better than average
year.
In the zones where minimum needs were not covered (northern
Peanut Basin, Niakhar in the central Peanut Basin, and Senegal
97
Oriental), from 50 to 75 percent of households had cropping income
below the minimum needs level.
In 4 zones (the northern basin, Colobane in the central basin, the
southeastern basin and Senegal Oriental) than half of cropping
income comes from peanut production while in the 2 other zones
(Niakhar in the central basin and the southwestern Peanut Basin)
cereal income was greater.
Use of modern inputs is extremely low, even in zones where
households have access to confectionery peanut and cotton
contracts.
Rural markets for agricultural labor do not appear to be very
active; most cropping labor is supplied by household members or
through reciprocal arrangements with neighbors.
Despite relatively low use of modern inputs such as fertilizer,
input costs represent from 15 to 20 percent of gross cropping
income, depending on zone.
The bulk of input costs are for peanut
seed.
Returns to labor in peanut production are higher than returns in
cereal production in 5 of 6 zones; the difference is statistically
significant in 2 zones.
All zones but one produce more of the crop providing higher
returns to labor;
Households producing more than adequate levels of cereals tend to
have larger farms and more land per AE, specialize more in crop
production than other households, and be in the upper income
categories;
The policy implications of these relatively low levels of cropping
income and productivity were discussed in the previous chapter in our
analysis of income diversification patterns so they will not be
discussed here.
The links between crop production patterns and
marketing behavior are examined in the next chapter.
98
REFERENCES
Dakono, A. 1991.
Comparaison des revenus au capital et an travail dans
les activites aqricoles et non aqricoles au nord du bassin
arachidier.
mimeo, Dakar.
Kelly, V.A.
1988.
Factors affectinq the demand for fertilizer in
Seneqal's Peanut Basin.
Ph.D. dissertation, Michigan State
University, E. Lansing.
Kelly, V.A. 1993.
Aspects Economiques de la Production et de la
Commercialisation des Produits Aqricoles au Niveau dese Menaqes du
Bassin Arachidier et du Centre du Seneqal Oriental.
Paper
presented at IFPRI/ISRA Project Workshop, Dakar, March 1993.
Martin, F. 1988.
Food security and comparative advantaqe in Seneqal:
A micro-macro approach.
Ph.D. dissertation, Michigan State
University, E. Lansing.
.,
:
:
,.
‘.
9 9
6. ANALYSIS OF HOUSEHOLD AGRICULTURAL TRANSACTIONS
INTRODUCTION
This chapter examines the nature of sales, purchases, and gifts of
the principal agricultural products produced and consumed by households
in the IFPRI/ISRA survey.
The value of transactions in FCFA per AE are
examined by zone, harvest year,
and product.
The discussion is
organized around the following questions:
0
How much (in FCFA and kilos) of each product are households selling
and purchasing?
How does this differ by study zone and by year?
What does this imply for aggregate demand and supply in the study
zones?
0
Who are the household members selling and buying? What is their
gender? What is their status in the household?
l
Why are the transactions taking place? need for food? need for
cash? normal postharvest marketing?
l
Where and with whom are the transactions taking place? What does
this imply about the role of the marketing infrastructure?
One of the main objectives of this analysis is to understand the
factors that influence the quantities of cereals marketed by rural
households and the quantities of local and imported cereals purchased.
Following
an overview of transactions,
the rural zones are
discussed first and then the urban.
METHODOLOGY
The analysis is based on data collected in fortnightly interviews
during the entire survey period -- October 1988 - September 1990 for
rural zones and July 1990 - June 1991 for urban zones.
A full recall
technique was used by the interviewer -- all purchases and sales made
between two sequential interview periods were enumerated even if the
interview period was longer than two weeks.' The household head as well
1
Due to interviewer illness or nonavailability of household
members,
it was not always possible to adhere to a strict fortnightly
schedule.
” 1’
)’
..
.
.
..
100
as all other members of the family who made transactions were
interviewed directly.
Respondents were asked about the products, forms (whole grain,
flour), quantities, units (kilos, calabashes, bowls), unit prices, total
amount paid, nature of payment (cash, credit, in-kind), source of
financing, place of transaction, type of buyer/seller, and reason for
making the transaction.
Fall et al. (1989) provide details of the
survey methods used.
All of these variables except prices are dealt
with in this chapter; prices are discussed in Chapter 9.
Estimates of average FCFA per AE are calculated by using actual
prices paid or received.
Gifts are valued using transaction-derived
producer or consumer prices drawn from the survey data.'
Kilos
transacted are obtained by converting all cereals to grain equivalent
form.
Most transactions were made in grain form so there is not a great
deal of measurement error introduced by making these conversions.
Products covered on the sales side of the questionnaire were
cereals, pulses,
home-produced fruits and vegetables, and crop by-
products.
On the purchases side, only cereals and pulses are covered by
this analysis.
Purchased fruits and vegetables were enumerated less
frequently and with less detail.3
Crop by-products are purchased
primarily for animal feed and are analyzed as part of the animal
maintenance costs in Chapter 7.
The discussion in this chapter is based on detailed information in
the appendixes to this report.
Tables A6.1-A6.55 in Appendix 6.1
present the value of sales, purchases, and gifts for each rural zone and
village in FCFA per AE.
Tables A6.80 and A6.81 cover the urban zones.
Tables A6.56-A6.72 list the characteristics of the transactions (gender
and status in household of person making the transaction, location of
transaction,
source of financing,
source of product, reasons for
transactions, and so on). Appendix 6.2 provides some additional
information on particularities of each zone (for example, weather or
interviewer errors) that influence the patterns observed,
' Because of extremely thin producer price data in zones which had
very poor harvests, we have not used the usual procedure of valuing
gifts given at the price prevailing for the month of the transaction.
We have instead used the annual price (averaged across the entire 12
months of the survey year).
3 These data are being analyzed in conjunction with the demand
modeling.
There are no descriptive results to report at present.
.
.’
. .
101
Preliminary results of analyses performed on data for 1988/89 were
reported in Kelly, Reardon and McNeilly (1991). Results reported here
will differ from earlier versions because weighting for market villages
has been incorporated and changes in the number of adult equivalents per
household during the survey period have been taken into account.
Data on purchases of cereals and pulses in this chapter will differ
from that presented in Chapter 8.
The differences are due to the types
of questionnaires providing the data.
In Chapter 6 we use a full recall
of all purchases and sales.
In Chapter 8 we estimate kilos of different
products purchased and consumed by extrapolating from a 24 hour recall
of foods consumed that was conducted only two times in each month.
OVERVIEW OF CROP SALES AND PURCHASES BY SAMPLE HOUSEHOLDS
Relationship Between Sales and Purchases
In 1988/89, half of the zones spent more on purchases than they
received from sales.
The northern Peanut Basin represents an extreme
case where purchases were almost five times greater than sales.
In the
central Peanut Basin, purchases were double sales for Niakhar and about
25 percent greater in Colobane.
The implication is that crop revenues
alone were unable to cover the full extent of cereal and pulse purchases
made by households in these zones.
Households in the northern Peanut
Basin needed almost 20,000 FCFA per adult equivalent from other sources
of income while those in the central Peanut Basin needed about 5,000.
The 1988/89 situation for the remaining three zones was more
favorable, with the value of sales minus purchases ranging from about
2,500 FCFA/AE in the southwestern Peanut Basin to 9,000 in Senegal
Oriental and about 12,000 in the southeastern Basin.
Total value of
1988/89 sales was highest in the southeastern Basin (about 22,000
FCFA/AE) and Senegal Oriental (about 17,000).
It was lowest in the
Niakhar zone of the central Peanut Basin (2,500) and the northern Peanut
Basin (4,500).
Of the five rural zones retained for 1989/90, both Niakhar in the
central Peanut Basin and Senegal Oriental had negative net transactions,
but the amount was less than 2,000 FCFA/AE. The total value of 1989/90
sales was three times the previous year's level in the Colobane zone of
the central Basin and more than double that in the southwestern Basin
(Table 6.1).
In the southeastern Peanut Basin, the increase in sales
was about 50 percent.
Only in Senegal Oriental did the value of sales
decline in the second year.
This occurred primarily because farmers in
the zone boycotted cotton.
The implication here is that in the Colobane
zone of the central basin and in the two southern basin zones, crop
sales minus cereal/pulse purchases left households with 15-25,000
102
FCFA/AE to be used for repaying credit, purchasing consumer goods,
savings,
or investing in other productive activities; households in
Niakhar of the central basin and in Senegal Oriental needed to find
other sources of income to simply cover food purchases.
Product Composition of Sales and Sources of Supply
Table 6.1 shows that peanut production provides the largest share
of cash income coming from crop production for most rural households.4
The only exception is Senegal Oriental in 1988/89, where 88 percent of
sales revenue came from cotton.
In a show of their discontent about
high input costs and low (frequently negative) net returns to cotton
production,
the farmers in the Senegal Oriental sample completely
boycotted the 1989/90 cotton program.
The boycott led to a sharp
increase in peanut production and sales (91 percent of sales revenue)
and increased marketing of cereals (8 percent in 1989/90 versus only 1
percent the previous year).
Table 6.1 --
Share of sales revenue from sales of home production by zone,
harvest year, and crop purchased
Part A -- Harvest year 1988/89
North
Central Basin
Southwest
Southeast
Senegal
crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Peanuts
91’
98
91
9ab
88
13
Cotton
0
0
0
0
5
86
Cereals
0
2
a
2
7
1 ’
Cowpeas
9
0
0
0
0
0
Total value per
AE of all sales
4041
1819
10645
a552
22210
15309
4 Readers should not confuse cash income with total real income
from agricultural production.
The latter includes production consumed
by households as well as cash sales (see Chapter 4 and 5).
103
Part 5 -- Harvest year 1989/90
North
Central Basin
Southwest
Southeast
Senegal
crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Peanuts
n.a.
91d
87
94
86
91
Cotton
n.a.
0
0
0
7
0
Cereals
n.a.
9
12
6
7
8
Cowpeas
n.a.
0
0
0
0
1
Total value per
AE of all sales
n.a.
3020
32338
15298
34772
5428
Source: ISRAIIFPRI
projeci survey.
Note: ".a. mmn~ not available.
’ Fourteen percent in shelled form.
’ Five percent are confectionery peanuts.
’ local rice.
’ Five percent in shelled form.
As a share of crop income, cereals were most important in the
Colobane zone of the central Peanut Basin in both years (8 and 12
percent of sales revenue, respectively).
In 1988/89 the southeastern
Peanut Basin followed Colobane, with 7 percent of cropping income coming
from cereals.
In the following year, the Niakhar zone of the central
Basin was in second place (9 percent), while the higher potential zones
in the southwest, southeast, and Senegal Oriental drew 6-8 percent of
sales revenue from cereals.
The share of cereal sales in total gross receipts from crop sales
illustrates the degree to which households depend on cereal production
for their cash incomes, but it does not improve understanding of coarse
grain supply.
Table 6.2 shows that per adult equivalent, the largest
supplier (29 kilos per AE sold) in 1988 was the southeastern Basin,
followed by Colobane in the central basin (15 kilos); quantities
marketed by other zones were insignificant.
In 1989/90, the Colobane
zone provided the market with the greatest quantity (89 kilos per AE),
followed by the southeastern Peanut Basin (60 kilos) and then the
southwestern Basin (15 kilos).
The fact that the Colobane zone was an
important supplier of cereals both years is worth noting because it is
the north-central part of the Peanut Basin in an agroclimatic area that
is not considered very conducive to production.
During the survey
period, the zone was considered a deficit area by the "Commissariat de
Securite Alimentaire" (CSA).
:
.’
:
”
I
:
104
Given the different population densities of the major cereal
marketing zones,
we extrapolated from survey households to the
production zones in general in Table 6.3 to show the total quantities of
marketed cereals that probably came from each study zone during the
survey period.
Table 6.2 -- Volume of cereal sales by zone and year
Harvest
North
Central Basin
Southwest
Southeast
Senegal
Year
Basin
Niakhar
Colobane
Basin
Basin
Oriental
(kilos per AE)
1988/89
0
1
15
3
29
1
1989/90
n.a.
5
a9
15
60
a
Source:
ISRA/IFPRI project survey.
Note:
n.a. means not available.
Table 6.3 -- Estimated volume of coarse grain sales by zone and year
Harvest
North
Central Basin
Southwest
Southeast
Senegal
Year
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(metric tons per zone)
1988/89
0
287
11362
886
3137
63
i989/90
n.a.
1437
67415
4432
6450
507
Source:
ISRA/IFPRI project survey.
Note:
n.a. means not available.
The Colobane zone of the central Peanut Basin is clearly the
biggest supplier of marketed cereals in the survey. Estimated marketing
from Colobane is almost four times greater than the next most important
zone in 1988/89 and ten times greater in 1989/90.
Product Composition of Purchases and Sources of Demand
Table 6.4 shows the share of purchases for home consumption by
zone, harvest year, and product.
Of greatest interest are the purchases
of rice versus coarse grains and the extent to which these differ by
zone and year.
.
;
:
:
:
s
‘.
105
Table 6.4 -- Share of purchases for home consumption by zone, harvest
year, and product
Part A -- Harvest vear 1988/89
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent of total FCFA/AE purchased)
Millet/sorghum
23
32
29
23
42
23
Maize
4
4
2
20
Imported rice
67
63
67
77
45
40
Cowpeas
3
Peanuts
3
1
3
13
1 1
Part B -- Harvest year 1989/90
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent of total FCFA/AE purchased)
Millet/sorghum
n.a.
23
29
7
46
22
Maize
n.a.
1
1
36
Imported rice
n.a.
71
6 4
91
47
3 4
Cowpeas
n.a.
1
Peanuts
n.a.
5
6
1
6
1
Source:
ISRA/IFPRI project survey.
Note: n.a. means not available.
In all zones but Senegal Oriental (1989/90), imported rice accounts
for the largest single share of expenditure on cereals and pulses,
representing 60 to 90 percent of purchases in the north, center, and
southwest and 30 to 45 percent in the southeast and Senegal Oriental.
In four zones (Niakhar in the center, both zones in the southern basin,
and Senegal
Oriental') the share of rice in purchases
increases
following the better harvest, suggesting that increased income and a
desire for diversity in the cereal diet are influencing demand. In the
Colobane zone of the central Basin, where the relative share of rice
declines following the better harvest, the drop is only 3 percent.
5 The better harvest for the first three zones was 1989/90; in
Senegal Oriental it was 1988/89.
.r
‘
.
.!
106
Purchases of pulses account for a relatively small share of
expenditures (less than 6 percent) except in 1988/89 when peanuts
accounted for 13 percent of purchases in the southeastern Basin and 11
percent in Senegal Oriental.
Cowpeas are purchased in such small
quantities that they are an insignificant share of expenditures in all
zones but the north (3 percent).
Table 6.5 presents data on cereal purchases in terms of kilos
purchased per adult equivalent.
With the exception of the northern
Peanut Basin in 1988/89, Colobane exhibits the largest cereal purchases
per adult equivalent for both years.
This is surprising given that the
zone also has one of the largest volumes of cereal sales in both years.
The production sufficiency ratio of 0.83 for Colobane in 1989/90
(Chapter 5) also suggests that the zone should not be purchasing 174
kilos of grain per adult equivalent in the same year. The possibility
of an error in data entry or the inclusion of commercial transactions
with household consumption transactions is being investigated. Assuming
there are no data errors, the implication is that households in the zone
are selling coarse grains and purchasing imported rice for their own
consumption.
Table 6.5 -- Volume of cereals purchased by zone, year, and product
North
Central Basin
Southwest
Southeast
Senegal
Year/Product
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(kilos per AE)
1988/89
Rice'
89
20
46
35
22
15
Coarse grains
52
17
41
16
38
28
Totalb
178
45
109
66
69
49
1989/90
Rice'
n.a.
22
60
43
23
16
Coarse grains
n.a.
10
87
7
64
52
Totalb
n.a.
41
173
68
97
75
Source: ISRA/IFPRI project survey.
Note:
n.a. means not available.
' Kilos of processed, imported rice.
' Rice transformed to whole grain equivalent using a coefficient of 1.43,
which implies a 70 percent processing loss.
Table 6.6 compares the level and composition of purchases in market
and nonmarket villages.
In both survey years, market villages spent
more per adult equivalent on cereal/pulse purchases than nonmarkqt
villages,
irrespective
of the zone.
The differences were more
,*
:
I
.’
.
107
pronounced
in the central Peanut Basin for both years,
in the
southeaste rn Basin for 1989/90, and in Senegal Oriental for 1988/89.
Table 6.6 -- Share of purchases for home consumption by zone, harvest
year, and product
Part A -- Market villages 1988/&X9
North
Central Basin
Southwest
Southeast
Senegal
crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
19
28
27
14
49
12b
Maize
1
7
1
22
Imported rice
70
63
67
86
48’
57
Cowpeas
5
Part B -- Nonmarket villages 1988/89
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
24
35
26
23
40
36
Maize
5
1
2
21
Imported rice
67
64
70
78
45
30
Cowpeas
3
Peanuts
3
2
15
13
:
,’
,.‘,
_.
‘;
.
I
.
,:
.
:
.
. .
.
108
Table 6.6 -- Continued
Part C -- Market villages
1989/90
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
n.a.
12 :
16
17
49
gb
Maize
n.a.
4
3
1
1
2 1
Imported rice
n.a.
a3
7 6
8 0
4 6
6 0
Cowpeas
n.a.
1
Peanuts
n.a.
2
4
2
4
9
Part D -- Nonmarket villages 1989/9o
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
n.a.
31
2 2
1 2
4 5
2 8
Maize
n.a.
3 3
Imported rice
n.a.
6 4
68
93
48
3 1
Cowpeas
n.a.
Peanuts
n.a.
6
9
1
5
9
Source: ISRAIIFPRI project survey.
Note: n.a. meene no1 available.
' One percent local rice included
b
All sorghum.
The only large differences in the share of rice purchased in
1988/89 occurred in the southwestern Peanut Basin and Senegal Oriental,
where the share of rice for market village households was 8 and 27
percent more than that in nonmarket villages.
In 1989/90, market
villages in three zones (Niakhar and Colobane in the central basin, and
Senegal Oriental) purchased a larger share of rice than the other
villages while the nonmarket villages in the two southern zones
purchased more rice than their market village neighbors.
Table 6.7 presents a picture of how total coarse grain demand for
the study zones looks when the population density of each zone is taken
into account.
Demand in Colobane alone is much greater than the sum of
demand across all other zones.
109
Table 6.7 -- Estimated volume of coarse grain purchases by zone and year
Harvest
North
Central Basin
Southwest
Southeast
Senegal
Year
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(metric tons per zone)
1988/89
10881
4887
31056
4727
4111
1775
1989/90
n.a.
2875
65900
2068
6923
3297
Source:
ISRA/IFPRI project survey.
Note:
n-a. means not available.
In comparing results from Table 6.3 with results from Table 6.7, it
is apparent that the Colobane zone is the greatest supplier and the
greatest demander of coarse grains.
Table 6.8 uses the estimated total
demand and supply data from Tables 6.3 and 6.7 to calculate net market
supply (supply minus demand) for each zone.
In harvest year 1988/89
demand exceeded supply in the zones covered by 41,703 tons; not a single
zone marketed more coarse grains than it purchased.
In harvest year
1989/90, the Colobane zone in the central basin and the southwestern
basin marketed 3,879 tons more than they purchased. In all other zones
purchases exceeded marketing.
The net effect across all survey zones
was a positive supply of 1,726 tons of coarse grains which would have
been available to other zones or urban areas.
Table 6.8 -- Estimated net volume of coarse grain transactions by zone
and year (quantity marketed minus quantity purchased)
Harvest
North
Central Basin
Southwest
Southeast
Senegal
Year
Basin
Naikhar
Colobane
Basin
Basin
Oriental
Total
(kilos)
1988189
-10881
-4600
-19694
-3841
-975
-1712
-41703
1989/90
n.a.
-1441
+1515
+2364
-433
-2790
+1726
Source:
ISRA/IFPRI project survey.
Note:
n.a. means not available.
Table 6.9 summarizes the value of seed purchases per adult
equivalent (175 to 5408 FCFA).
The Colobane zone of the central Peanut
Basin spends more per adult equivalent than other zones. Peanut seed is
clearly the dominant item, accounting for 91 to 300 percent of all seed
purchases.
It is informative to note that cowpeas are a larger share of
seed purchases than cereals despite the very small role cowpeas play in
”
.i
.
.
,’
110
total agricultural production.
This is most likely due to the
difficulty in storing cowpeas.
Table 6.9 -- Share and total value of seed purchases by zone, year, and
product
Part A -- Harvest year 1988/89
%
1
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Cowpeas
2
0
0.5
0
0
1
Cereals
1
1
0.5
0
0
0
Peanuts
97
99
99
100
100
99
Total value FCFA/AE
1590
400
5408
233
1486
621
Part B -- Harvest year 1989/90
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Cowpeas
n.a.
1
1
0
0
3
Cereals
n.a.
0
0
0
0
6
Peanuts
n.a.
99
98
99
100
9 1
Total value FCFA/AE
n.a.
340
2547
571
1661
175
Source:
ISRA/IFPRI project survey.
Note:
n.a. means not available.
Role of Gifts in Total Value of Crop Transaction
It is sometimes argued that traditional mechanisms for transferring
cereals from surplus to deficit households provide rural areas with a
food security buffer that functions as well as any government program
might be expected to function.
Data on gifts given and received
presented in Tables 6.10 and 6.11 show that transfers do occur but that
they are a very small share of total transactions. The total value per
adult equivalent of gifts given ranges from 67 to 3446 FCFA across zones
. . . ,,
,,:
.
.’
.
.
:
;
111
and years.
For gifts received, the range is 28 to 2203 FCFA per adult
equivalent.6
Table 6.10 -- Share and total value of gifts out by zone, year, and
product
Part A -- Harvest vear 1988/89
North
Central Basin
Southwest
Southeast
Senegal
Product
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
27
88
68
99
86
51
Maize
0
0
0
0
1
26
Imported rice
4
2
2
0
2
18’
Cotton
0
0
0
0
0
Cowpeas
14
0
0
0
0
Peanuts
56
10
31
1
11
4
Total value FCFA/AE
of all gifts given
516
194
1237
1394
1887
1326
Part B -- Harvest year 1989/90
North
Central Basin
Southwest
Southeast
Senegal
Product
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
n.a.
100
56
98
91
63
Maize
n.a.
0
0
0
2
24
Imported rice
n.a.
0
0
1
4
6b
Cotton
n.a.
0
0
0
0
0
Cowpeas
n.a.
0
0
0
0
0
Peanuts
n.a.
0
44
1
3
a
Total value FCFA/AE
of all gifts given
n.a.
67
407
1749
3446
704
Source:
ISRA/IFPRI project survey.
Note: n.a. means not available.
' Includes 7 percent local rice.
b Includes 4 percent local rice.
6 The estimates of net transfers presented in Chapter 4 did not
consider the value of in-kind transfers.
Adding values reported in
Tables 6.7 and 6.8 to those in the "transfer" column of Table 4.5
provides a more complete picture of net transfer income.
-
1
I’.
.
‘
.’
.
,,
:
.
.
*.
:
112
Table 6.11 -- Share and total value of gifts received by zone, year, and
product
Part A -- Harvest year 1988/89
North
Central Basin
Southwest
Southeast
Senegal
Product
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
,,,
10
3 5
72
88
8 8
2 4
Maize
3
0
0
0
0
2 9
Imported rice
42
59
21
1 2
9
22'
Cotton
0
0
0
0
0
0
Cowpeas
26
0
1
0
0
1
Peanuts
19
5
6
0
3
23
Total value FCFA/AE
of all gifts given
9 9
28
174
1 4
756
716
Part 6 -- Harvest year 1989/90
North
Central Basin
Southwest
Southeast
Senegal
Product
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
n.a.
7 2
3 1
4 1
46
7
Maize
n.a.
0
0
0
0
3 1
Imported rice
n.a.
2 1
36
45
5
52’
Cotton
n.a.
0
0
0
0
0
Cowpeas
n.a.
1
2
0
0
4
Peanuts
n.a.
6
3 2
1 3
49
5
Total value FCFA/AE
of all gifts given
n.a.
174
208
241
2203
1188
Source:
ISRA/IFPRI project survey.
Note: n.a. means not available.
' Includes 1 percent local rice.
b The large share of peanuts may be due to a case where in-kind payments
for assisting neighbors with their peanut harvest were coded as gifts.
Gifts received by rural households are composed to a large extent
of imported rice and gifts given are primarily coarse grains.
These
results provide some basis for inferring that gifts are exchanged
between rural (giving coarse grains and receiving rice) and urban
(giving rice and receiving coarse grains) households.
.
.
I
.’
.(
.’
:
113
CHARACTERISTICS OF CROP SALES TRANSACTIONS
Identity of Household Members Sellinq
Figures 6.1A-E shows the share by gender of agricultural products
sold.
The shaded part of the bars shows the share of total sales made
by women.
The left axis shows the total value per AE of all sales of
the product concerned,
helping one keep in perspective the relative
monetary value of each crop.
Figure 6.1
Relative shares of sales by gender for (A) millet, (6) peanuts, (C) cotton, (0) cowpeas,
and (E) other products for harvest years 1988/89 and 1989/90. Key: SAG, Sagatta; NIA, Niakhar; COL,
Colobane: PAS, Passy; 010, Nganda-Dioly; MIS, Missirah.
Part A
Y
-2500-
F
\\
v-l 2000 -
w
2
IT
crl rsooj
-
11 II
t;
j lOOO-
2
II
500 -
rl n ! NA
i
I
I
I
I
I
I
1
3HU IYIh LUL I ,-Ta U,”
MIS
SAG NIA COLPAS DIO
1988/89
1989/90
0 MALE
m FEMALE:
114
Part B
35 (
SAG NIA COL PAS DIO iIS
SAG NIA COL PAS
1988/89
1989,'90
0 MALE m FEMALE
Part C
1988/89
1989,'90
0 MALE @FEMALE
115
Part D
400
NA
I I
I
I
I
I
I
I
m
I
1
SAG NIA COI PAS DIO MIS
I
I
SAG NIA COL PAS DIO MIS
1988/89
1989/90
0 MALE @ FEMALE
Part E
0.5
20.45
g 0.4
E 0.35
\\
z 0.3
-J
2 0.25
c
SAG NIA COL PAS DIO MIS
SAG NIA COL PAS DIO MIS
1988/89
1989/90
0 MALE m FEMALE
,’
.
::.
._
.,
;
_‘.
.
.’
.’
.
::’
.
.
116
Crop sales are dominated by males.
Women are more active in peanut
sales than in other crops.
They accounted for 14-17 percent of peanut
sales in the central, southwestern and southeastern Peanut Basin in
1988/89 and 14 to 29 percent in 1989/90. Women's share in the northern
Peanut Basin was unusually low (4 percent) and unusually high in Senegal
Oriental (33 percent in 1988/89 and 29 percent in 1989/90). Women's
share of peanut sales may have increased in 1989/90 because household
heads were obljged to concentrate on cereal production after the very
poor 1988/89 harvest.
These results suggest that the tendency of women to shift to cereal
crops immediately after the government seed distribution program was
dropped in the mid-1980s (see Ndoye, unpublished) has not been
sustained.
As time has passed, women have found the means to acquire
peanut seed and are now back in the peanut business.
A comparison of coarse grain sales and production data (Chapter 5)
for women suggests that many are selling small amounts of cereal
received from the household head for daily meal preparation and using
the receipts to purchase condiments.
This is particularly true in the
southwestern Peanut Basin where the women felt free enough to reveal
their "theft" of household cereals to interviewers. It probably occurs
in other zones to an extent greater than that revealed in the data.7
Women dominated cowpea sales in the central basin -- Niakhar (100
percent) and Colobane (50 percent) in 1989/90. In the northern Basin,
where cowpeas have been promoted as a cash crop alternative to peanuts,
men dominated but women had a 13 percent share.
Women had an important share of sales revenue from "other' products
in the central basin zone of Colobane, in the southeastern basin and in
Senegal Oriental in 1988/89 and in both zones of the central basin in
1989/90.
Only in Colobane in 1989/90, however, was the value of sales
of any importance.
Products sold were sorrel and gumbo.
In most zones and years, the household head sells 50 percent or
more of the principal cereal and cash crops.
Reasons for Sellinq
Because 50-95 percent of peanuts are usually sold during the
official marketing campaign, respondents did not usually elaborate on
reasons for selling.
In 1988/90 the Colobane zone in the central Peanut
Basin appears to have sold only 20 percent during the official campaign
(these data are being checked for entry errors because the table
concerning partners shows 71 percent of sales made to institutional
purchasers).
Since Colobane is a zone where it is common to process
7
The zone of Colobane in the central basin presents an exception
because women did produce most of the cereals they marketed in this
zone.
‘.
.:
.
.
.
i
I
:
.
:
%.
‘.
117
peanuts to produce oil, it would not be surprising to find a large share
of peanut production marketed unofficially, but 80 percent seems too
high.
The second most important reason for selling peanuts is to
purchase food.
Cereals are sold primarily to purchase other foods, finance
ceremonies,
or finance other economic activities.
Each zone and year
has its own particularities and there are no clear trends.
Location of Sales
Most peanuts appear to be sold to official marketing agents (71 to
100 percent,
with Colobane at the lower end).
Senegal Oriental
presented an exception to this rule in 1989/90 when a project by a
nongovernmental organization appears to have purchased about 90 percent
of the sales.
Local private traders are important as a secondary
purchaser in the central Peanut Basin, as are neighbors.
In the
northern Peanut Basin, private traders from outside the immediate area
were the most important secondary purchasers.
The type of persons to whom cereals were sold changed in some zones
between the two survey years, probably due to the increase in cereal
production in 1989/90.
In the Niakhar zone of the central basin, local
traders became more important in the second year, diminishing the share
of interhousehold transactions.
In the Colobane zone of the central
basin, traders from outside the village became more important as the
share of village traders declined.
In the southwest the share of
village traders declined as the share of local ambulant traders
increased.
Traders from outside the village increased in importance in
the southeastern basin while interhousehold transactions continued to
account for 25 percent of transactions.
In Senegal Oriental, most
cereal transactions in both years were interhousehold, with about 75
percent taking place in the interviewer's own village and the rest in
neighboring villages.
The market village was the most important location for sales of
cereals in the central Peanut Basin (Niakhar), the southwestern basin,
and the southeastern basin.
In Colobane, the arrondissement is most
important, but the arrondissement is also the site of the weekly market
for Colobane.
In Senegal Oriental most cereal sales are made in the
village of the seller, but cowpea sales are concentrated in the market
village.
Given the importance of market villages in marketing cereals,
programs to increase the number of market villages in zones which are
not well endowed (the southeast and Senegal Oriental) should encourage
greater cereal sales.
The locations of peanut sales are varied and determined by the
location of the official marketing services. In most cases, these are
in nearby villages or rural communities.
118
CHARACTERISTICS OF PURCHASE TRANSACTIONS
Identitv of Household Purchasers
Figure 6.2 A-E presents a breakdown of purchases by gender. Males
dominate cereal purchases, as is the tradition throughout Senegal.
Dependent household heads and wives of the household head are the other
major categories of cereal purchasers.
Women's participation was
greatest
in the northern Peanut Basin, where the harvest was
particularly bad i% 1988/89.
There appears to be a somewhat greater
female contribution to rice purchases than to coarse grains.
These
could be cases of women too busy to prepare the usual coarse grain meals
resorting to buying rice to shorten preparation time.
Figure 6.2
Relative shares of purchases by gender for (A) millet, (6) peanuts, (C) cotton, (D)
cowpeas,
and (E) other products for harvest years 1988/89 and 1989/90.Key:
SAG, Sagatta; NIA,
Niakhar; COL, Colobane; PAS, Passy; DIO. Nganda-Dioly; MIS, Missirah.
Part A
4500
2. 4000
2 3500
L
F 3000
\\
g 2500
LA
; 2000
5
& 1500
L3 1000
5 500
0 T
,
I
4
8
SAG MIA 'COL PAS DIO MIS
SAG NIA COL PAS DIO MIS
1988/89
1989/90
n MALE m FEMALE
119
Part B
NA
I
I
I
n
-
-
I
1
I
I
I
I
I
I
SAG NIA COL PAS DIO MIS
SAG NIA COL PAS DIO MIS
1988/89
989/90
0 MALE m FEMALE
Part C
14
SAN NIA COL PAS DIO MIS
SAG NIA COL PAS DIO MIS
1988/89
1989/90
n MALE m FEMALE
.‘_
I
:
.’
.
.
.
120
Part D
rfY
SAG NIA COI PAS DI( MIS
SAG NIA COL PAS DIO MIS
i988/89
i989/90
u MALE m FEMALE
Part E
600CI--
2
b
b5OOC )-
5
>4000 I -
t;f
IQ0 3000
5
a
r
g 2000
ii
2 1000
n
s
k!
NA
0 t-
J
l-l
I
I
1
I
1
SAG NIA COL PAS DIO MIS
SAG NIA CDL PAS DIO MIS
1988/89
1989/90
(MALE m FEMALE
1
‘.
.
+
,
.
:.
.
:’
:
.
..
.
*
.
:
121
Peanuts are purchased primarily by household heads and their wives
and dependent household heads.
Peanuts for consumption are the only
product where women's share is noticeable, although this is not true in
all zones.
Women in both zones of the central basin and in Senegal
Oriental generally purchased more peanuts than women in the southern
Peanut Basin.
This behavior reflects the traditional obligation that
women provide ingredients for the sauce.
Reasons Products Are Purchased
Most cereals are purchased for home consumption. In the northern
Peanut Basin and Senegal Oriental, the pattern of buying and stocking
rather than purchasing for immediate consumption is more common than in
other zones.
During the difficult 1988/89 harvest year, households in
the central basin did not purchase and stock but they did following the
better harvest in 1989/90.
Peanut purchases are primarily for seed (more than 50 percent of
all peanut purchases except in Senegal Oriental, where seed accounts for
30 to 50 percent).
Sources of Financinq for Purchases
Money used for cereal purchases comes from a wide variety of
sources.
It is clear that income from the sale of cash crops
contributes only a very small share to cereal purchases across all zones
(less than 20 percent). Wages, migration, commerce, and livestock are
the primary sources of income used to purchase cereals.
Livestock is
more important (26 percent) in the southeastern basin and Senegal
Oriental while wages and commerce dominate in other zones. Migration
income is the biggest single source of cereal money in Senegal Oriental.
Location and Sources of Purchases
Cereals are usually bought from traders; Senegal Oriental is the
exception where interhousehold transactions are more common. The market
village is the most common place to purchase cereals; Senegal Oriental
is again the exception, with nearby villages playing an important role.
AGRICULTURAL PRODUCT TRANSACTIONS IN URBAN ZONES
Town of Kaolack
Tables A6.55a-c show the average amount (in FCFA) of purchases,
sales
and transfers by adult equivalent of agricultural products
undertaken in Kaolack by households in the sample during the period from
July 1990 to June 1991.
Level of transactions.
The global level of agricultural products
purchased for all uses is 18,383 FCFA/AE.
This is above the level
registered in the rural zones.
The value of the sales, as anticipated,
is very small:
188 FCFA/AE.
.
.
;.
‘_
.’
122
As predicted,
transactions for home consumption are far more
important than purchases for trade or seed.
In fact, purchases of
coarse grain, for home consumption,
represent fourteen times the value
of those undertaken for trade reasons:
17,206 and 1,178 FCFA. The
outflows are primarily gifts (90 percent) with some sales (10 percent).
Composition'bv product.
Most of the purchases are imported rice. This
product represents 71 percent of the total value of cereal/pulse and 73
percent of cereal purchases and gifts received.
Millet, the second cereal consumed in Kaolack represents 19 percent
of cereal/pulse purchases and almost half of gifts received or made by
households.
Originating from harvests of relatives living in the
countryside, millet gifts reflect the links between the rural and urban
zones.
Locally grown products make up two thirds of gifts received;
the rest is imported rice.
The importance of household purchases of cowpeas is relatively
small compared to peanuts.
Of all purchases and gifts received, peanuts
follow rice and millet in importance.
In general, peanuts are bought
shelled but are given unshelled.
Finally,
maize makes up
a marginal
share of agricultural
transactions, especially in the form of purchases. The other products
such as sorghum, fonio and cotton are not found at all.
Commercial transactions are made up in large part of purchases of
millet for processing and resale.
The sale of millet cereal, especially
for breakfast, is a common activity in Kaolack.
Rice, vegetables,
cowpea and peanuts are also purchased for resale.
It is important to
note that these products purchased for trade are transformed (prepared
for meals) before they are resold.
Hence, the importance of their
resale doesn't appear in the sales which include only unprepared
products.
Seed transactions are non existant at the level of households being
studied because urban households in the Kaolack sample do not farm.
Town of Tambacounda
The average value (in FCFA/AE) of agricultural purchases, sales,
and gifts received and given by households in the Tamba sample from July
1990 to June 1991 is presented in tables A6.56a-c.
Level of transactions.
The total value of all agricultural products
purchased or received as gifts is 19,692 FCFA/AE.
It is higher than
that observed for Kaolack.
The very low value of sales and gifts given
(275 FCFA/AE) exceeds that of Kaolack.
Here, as in Kaolack, transactions for home consumption dominate the
transactions for trade and seed.
The inflows of coarse grains, for
consumption,
account for 19,626 FCFA/AE while only 41 FCFA are
..<I
.*
.J
.
.
:
.’
123
attributed to trade and 28 FCFA to seed. The outflows are made up, as
in Kaolack, of gifts.
There are no other reasons for grain outflows.
Composition b.y product. The product composition of transactions is very
diversified in Tamba. Almost all products, with the exception of fonio,
cotton, and local rice, play a part in household transactions.
Imported rice dominates transactions:
75 percent of the total
value of inflows of agricultural products for Tamba households. In
addition, rice makes up 78 percent of household cereal purchases+
Rice
also accounts for a few gifts received by households (only 6 percent)
but constitutes 24 percent of gifts made by these same households.
Maize, sorghum, and then millet are much less important than rice
in household purchases.
Their individual share is about 6 to 7 percent.
On the other hand, more than half of gifts received by households are
maize,
followed by sorghum.
Gifts made are exclusively cereals,
especially traditional local cereals.
This larger diversification of
gifts received or given can be explained by the fact that there are
relatively more households active in agriculture in Tamba than in
Kaolack.
Peanuts come in first place among pulses, but far behind
cereals as a share of purchases.
Commercial transactions are essentially millet purchases (94
percent).
As in Kaolack, the sale of millet cereal for breakfast can
explain the predominant place of millet.
Seed purchases (primarily
peanuts) also take place, even though at a low level.
SUMMARY OF FACTORS INFLUENCING THE QUANTITY OF CEREALS MARKETED
One of the major objectives of the IFPRI/ISRA study is to
understand the factors that influence the quantity of cereals marketed
by rural households in Senegal.
Economic theory tells us that producer
price should be one of the most important variables. This chapter deals
with factors other than price, leaving the discussion of price for
Chapter 9.
At this stage of the analysis we have simply noted correlations
between large quantities marketed and particular zones, types of
households (market or nonmarket village location), and types of
individuals within households (gender and status). The discussion ends
with presentation of results of chi square tests used to identify
characteristics distinguishing households selling large quantitities of
cereals from other households.
Crop production levels and the production sufficiency ratios
explain a great deal about cereal marketing behavior.
In general, zones
with the better production sufficiency ratios marketed a larger share of
their cereals.
124
Cereal marketinq ratios
Tables 6.12 and 6.13 present the ratio of cereals and peanuts
marketed to total production by zone and year.
Data on marketing are
available for harvest years 1988/89 and 1989/90 only. A comparison of
the two tables makes the point that peanuts play a much more important
commercial role than cereals.
Table 6.12 -- Share of total cereal production marketed by zone and year
Northern
Central Basin
Southwest Southeast Senegal
Basin
Niakhar Colobane Basin
Basin
Oriental
Year
1988189
0.03
0.08
0.14
0.16
0.21
0.18
1989/90
n. av.
0.04
0.12
0.23
0.28
0.15
Table 6.13 -- Share of total peanut production marketed by zone and year
Northern
Central Basin
Southwest Southeast Senegal
Basin
Niakhar Colobane Basin
Basin
Oriental
Year
1988/89
0.62
0.30
0.62
0.34
0.72
0.21
1989/90
n. av.
0.16
0.72
0.49
0.68
0.39
In general, one notes a correspondence between the propensity to
market and the higher CPSRs.
The only notable exception is Senegal
Oriental, which exhibited the lowest CPSR in 1989/90 (except for the
northern basin where deficits are chronic) but a relatively large share
of cereals sold (15 percent).
This behavior, coupled with the low
levels of caloric intake reported in Chapter 8, prompts one to ask if
consumption levels are not being pushed lower by households selling
cereals from stocks that were inadequate from the beginning.
The correlation between CPSRs and marketing behavior underscores
that importance of increasing cereal productivity as a means of
increasing quantities marketed.
Productivity was extremely low in 1988/89 when rural households in
the zones surveyed purchased almost 42,000 tons more coarse grains than
they sold.
Following the better-than-average harvest of 1989/90, the
net supply of cereals (sales minus purchases) in the zones covered is
estimated to be only 1726 tons.
125
Peanuts remain the major source of cash income in rural areas.
When peanut income is adequate, households are less likely to sell
cereals. The higher 1989/90 cereal sales in Senegal Oriental -- despite
a relatively low production sufficiency ratio -- are partly due to a
drop in cash crop revenue following the cotton boycott.
The Colobane zone of the central Peanut Basin marketed the greatest
quantity of cereal in both years (Table 6.3). Much of this zone falls
in areas of unreliable rainfall and rapidly deteriorating soils,
therefore it,is surprising that the production sufficiency ratio was so
favorable.
The zone is characterized by a very active commercial sector with
several weekly markets in close proximity (Colobane, Mbar, Gnibi) and
the Mecca of Peanut Basin commerce--Touba--not very far away.
Many
households in this zone paid attention to price information coming from
Touba, and sold their products in Touba when prices there were more
favorable than locally.
Two of the three villages in the zone are located on a paved road;
the third village is about 7 kilometers from the paved road via a very
sandy road not conducive to truck traffic.
Dispite the disparity in
road infrastructure, the village with more difficult access sold a
larger share of its cereal production and about the same share of its
peanut production as the other nonmarket village in the zone. We have
not analyzed sales patterns (to whom and where sold) on a village-by-
village basis. For the zone as a whole, 72 percent of cereals were sold
to a village trader.
It would be useful to know if households in the
more isolated village were transporting their production to market or if
the commercial network in the zone was present at the farmgate.
Household heads are clearly the most important marketers of cereals
in all zones.
Women in the Colobane zone, however, played an unusually
large role in the marketing (37 and 27 percent of sales, depending on
the year).
A quick review of crop production data for Colobane reveals
that the wives of household heads produced about 3600 kilos of cereals
during 1989/90--this is approximately the amount marketed by them. The
implication here is that more attention to women in extension and
marketing interventions might bring about a greater marketed surplus
than one gets when dealing with men who have the obligation of providing
household cereals.
The extent to which this would be true in other zones remains to be
examined.
Women in the southwestern Peanut Basin also sold a large
share of cereals (45 percent) in 1989/90; however the crop production
analysis shows only 103 kilos of cereals produced by women (less than 3
percent of total marketings).
This is a case where they were clearly
selling millet from the daily food allotment given to them by the
household head.
126
Table 6.14 shows the distribution of households by quantities of
cereals marketed in both 1988/89 and 1989/90. Households selling more
than 500 kilos almost doubled between the two years. The tendency was
for those in the 50 to 200 kilo range in 1988/89 to move into the higher
up categories.
We find very little change in households that were
selling less than 50 kilos; approximately the same number of households
are found in this category both years.
In general, they tend to be the
same households,
suggesting that something is preventing these
households from reaping the commercial benefits of the better rainfall
in the same way that the bigger sellers do.
We suspect that these
households are ones that do not significantly increase levels of cereal
production during the year of better rain; more work needs to be done
here to identify the exact nature of the problem.
Table 6.14 -- Distribution of households by kilos of cereals marketed in
1988/89 and 1989/90
Category
Number of households
Number of households
1988/89
1989/90
no sales
12
10
<50
25
2 7
<200
5 7
36
<500
3 1
37
>50
1 6
3 1
Source:
IFPRI/ISRA survey data.
The entire sample had 47 households that were selling more than 200
kilos of cereals per adult equivalent in 1988/89 and 68 in 1989/90.
Again using two by two contingency tables and chi-square significance
tests we looked for characteristics that differentiated these big
sellers from other households.
We found that big sellers were more
likely to:
- be in the southeastern basin in 1988/89 and in both southern
basin zones in 1989/90
- be living in nonmarket villages
- be Wolof or Fulani
- be in the upper income categories
- earn more than 40 % of total household income from
cropping activities
- have CPSRs >.75
- cultivate > 6 hectares and > 1 hectar/ae
- derive ~10 percent of cereal calories from imported rice
,
:
.
.
‘.
127
Although big sellers are more likely to exhibit the above set of
characteristics,
it is important to note that there are numerous
exceptions to the general tendency.
It is of interest that some large
sellers are in the low income categories
(13 percent in 1988/89 but
only 4 percent in 1989/90) and also ranked as food insecure households
(57 percent in 1988/89 and 40 percent in 1989/90).* These food insecure
households warrant further investigation to determine why they are
selling large quantities of cereals rather than retaining production for
their own consumption.
Somewhat disconcerting is the fact that Senegal Oriental had the
largest marketable surplus in 1988/89 but was not among the zones where
we found a larger than expected share of big cereal sales.
Is this due
to lack of demand for their production (primarily maize and sorghum) by
other zones? Transport costs that make their production too expensive
to market in other zones?
Access to land appears to be an important factor in encouraging
cereal sales and production.
The effect of farm size being greater than
6 hectares appears to be more important than farm size per AE.
Proximity to market infrastructure is normally expected to
encourage
increased crop production and marketing as
it reduces
transportation and handling costs for obtaining inputs and selling
production.
In the rural areas surveyed, however, proximity to
marketing infrastructure is not serving as an impetus to greater crop
production and marketing.
This behavior is in sharp contrast to the
very important influence that proximity to markets appears to have on
the consumption side.
Information presented in Chapters 5 and 6 makes it quite clear that
most farmers are not producing enough cereal for themselves -- even in
an above average year.
This provides a relatively pessimisitic
appraisal
of the potential to achieve national
cereal production
objectives.
Although the level of analysis conducted to date does not provide
precise guidance on how to increase cereal production, a number of
related issues warrant further consideration.
The fact that returns to labor in peanut cultivation are greater
than those in cereal cultivation in four of the six zones, clearly
diminishes the incentive for farmers to produce and market cereals.
Unequal progress in liberalizing cereal and peanut markets may also
be discouraging cereal production and marketing. The peanut marketing
a The food insecure classification is based on analysis of food
consumption data and a determination that average caloric intake for the
household was less than 2400 kcal/ae.
128
structure (guaranteed purchases at guaranteed prices) provides added
incentive for farmers to favor peanut production over cereal
cultivation.
Cereal producers market their produce in a coarse grain
market that is liberalized so producers do not benefit from guaranteed
purchases and prices.
Furthermore, they are faced with an imposed
"ceiling" price because imported rice prices are controled at about 135
FCFA/kilo.
SUMMARY OF FACTORS INFLUENCING THE QUANTITIES OF CEREALS PURCHASED
Chapter 8 presents a more detailed analysis of food consumption
patterns and the link to cereal purchasing.
The major factor
influencing purchases of cereals to be identified to date is living in
a market village--the quantity of cereal purchases per adult equivalent
for market village residents is higher across all zones.
There is no clear pattern suggesting that living in a market
village increases the share of rice purchased compared with the share of
coarse grains.
In 1989/90, three zones had a larger share of rice in
purchases made by market village residents; in two zones the share is
lower.
Purchases appear to be lower in zones where nonagricultural incomes
are lower (Niakhar,
in the central basin, the southeastern basin, and
Senegal Oriental).
This suggests that households lacking access to
nonagricultural
incomes have difficulty finding the liquidity to
purchase cereals.
This hypothesis is supported by relatively low levels
of caloric intake for these zones (see Chapter 8).
Men, particularly household heads, are the primary purchasers of
cereals.
Women purchase a greater share of rice than of other cereals,
but it is still a small share,
..;
.
.j
:
.
.
:
..
129
REFERENCES
Fall, A.A., V. Kelly, and T. Reardon.
1989.
Household-level survey
methods used in the IFPRI/ISRA study of consumption and supply
impacts of agricultural price policies in Senegal.
Document 2.
IFPRI, Washington, D.C.
Kelly, V., T. Reardon, and L. McNeilly, 1991. Sales, purchases, and
gifts of cereals and pulses in the Peanut Basin and Senegal
Oriental.
Document 6.
IFPRI, Washington, D.C.
Ndiaye, 0.
Unpublished Government regulations and cereal marketing in
Senegal,
in Delgado, Reardon, Matlon (eds.) Commodity priorities
and changing food patterns in West Africa, IFPRI, Washington, D.C.
,,
.
.
1
.I
.
:
I
130
7.
LIVESTOCK TRANSACTIONS
INTRODUCTION
This chapter examines livestock sales, purchases, and consumption
from household herds.
It addresses five questions:
a
Which animal groups are most important in sales, consumption,
and purchases?
0
Why do people sell?
0
Is there any seasonality in livestock sales?
0
What does animal husbandry contribute to the total level of
income?
0
How is income from livestock distributed?
The data used for the analyses in this chapter come from monthly
interviews of household members owning livestock.
The interviews
enumerated all sales, purchases, gifts, births, and deaths (including
slaughtering for home consumption).
They also enumerated all cash
expenditures for livestock maintenance (vaccinations,
herder fees,
supplementary feed).
This analysis does not take into consideration income (gains or
losses) due to changes in stocks during the year.'
Gifts are not
analyzed either.
Maintenance costs do not cover inputs from family
resources, such as unpaid family labor and home-produced peanut hay or
cereal straw.
The analysis concerns only rural households because livestock
transactions in urban households were extremely minor.
Results are
presented for harvest years 1988/89 and 1989/90. Data collection began
well after the beginning of the 1988 harvest year (January or February
of 1989, for most zones).
In a few cases, the interviewees were able to
do a full recall to the beginning of the harvest year; most of the time
' Ndiaye (1993) reports results of a more in-depth analysis of the
IFPRI/ISRA livestock data for three zones (Niakhar and Colobane in the
central basin and Passy in the southwest) in his M.S. paper Micro-level
Small Ruminant Transaction Analysis and Livestock Policy in Seneqal. He
estimated beginning and ending stocks for the period October 1989
through September 1990 and used these estimates to develop income
statements for small ruminant enterprises.
Appendix 7 contains the
summary, conclusions,
and policy recommendations reported in the final
chapter of his paper and several pertinent tables.
131
this was not possible.
As a result, data presented for harvest year
1988 underestimate total transactions.
The period missing is a period
of both sales and purchases, hence,
it is difficult to draw any
conclusions about the direction of any bias in net transactions income
that has been introduced.
Net transactions income is calculated as follows:
sales - purchases t home consumption - maintenance.
Table 7.1 presents a summary of net livestock income across zones
and shows the share of sales receipts by animal category. More detailed
tables are in Appendix 7.2.
Table 7.1 -- Summary of net income from animal transactions
NPB
CPBN
CPBC
SUP6
SEPB
cso
Net income
FCFA/AE
7015
749
6473
1457
6119
3778
Share of sales receipts
from
Cattle (%)
4 4
41
28
50
5 6
a 4
Small ruminants (%)
5 5
5 4
58
47
4 3
1 5
Poultry (X)
1
5
14
3
1
1
Maintenance Expenditure
(FCFA/AE)
a01
908
490
160
18
1340
Pal-t: 3:' r9&9/9o
NPB
CPBN
CPBC
SUPB
SEPB
cso
Net Income
FCFA/AE
N.AV.
988
3032
1248
7878
4904
Share of sales receipts
from
Cattle (X)
N.AV.
54
28
4 6
53
91
Small ruminants (%)
N.AV.
4 2
61
52
4 6
a
Poultry (X)
N.AV.
4
11
2
1
1
Maintenance Expenditure
(FCFA/AE)
N.AV.
866
1120
1074
726
720
:
.I
.
.
I,
‘_
.
132
IMPORTANCE OF ANIMAL GROUPS
Sales
In the periods and zones surveyed, the value of sheep and goats
sold roughly equalled that for cattle and traction animals.
The
exceptions are in the central Peanut Basin (Colobane), where sales of
small ruminants were double those of larger animals in harvest year 1989
and in Senegal Oriental, where sales of cattle in both harvest years
were more than 80 percent of total sales.
The contribution of poultry to total value of livestock sales was
generally very low (O-Z percent) except in the central basin (Colobane)
where poultry accounted for lo-15 percent of sales depending on the
year.
Consumption
The contribution of home consumption to net transactions income was
less than 15 percent in most zones and years.
The most striking
exception is the southwestern Peanut Basin where home consumption
accounted for 62 percent of livestock income in harvest year 1989. The
central Peanut Basin (Colobane) also had relatively high home
consumption income (31 percent of total livestock income in 1988 and 33
percent in 1989),
as did the southeastern Peanut Basin in 1988 (24
percent).
Sheep and goats provided the bulk of animal protein consumed from
owned stocks, supplying more than 90 percent of the livestock income
derived from home consumption in all zones but two. This is undoubtedly
the "Tabaski" effect from households slaughtering their own sheep to
fulfil1 religious obligations.
A somewhat lower share of home
consumption came from small ruminants in the northern Peanut Basin in
1988 (79 percent) and Senegal Oriental in both harvest years (84 and 42
percent).
Consumption of cattle contributed to net transactions income in
only two zones.
In the northern Peanut Basin, cattle accounted for 20
percent of home consumption value and 4 percent of net transactions
income.
In harvest year 1989, 55 percent of home consumption and 6
percent of net transactions came from cattle in Senegal Oriental.
Where household stocks were consumed, poultry contributed from 1 to
16 percent of income derived from home consumption.
Although home
consumption of poultry was a small share of total livestock income,
home-consumed poultry accounted for a relatively large share of net
poultry income in the southwestern basin (53 percent), southeastern
basin (81 percent),
and Senegal Oriental (46 percent) in harvest year
1989.
No home consumption of owned poultry stocks was reported in the
Niakhar zone of the central basin in either period and none in the
southwest Peanut Basin for harvest year 1988/89.
133
The major conclusion to be drawn here is that animal husbandry
activities do not make a very large direct contribution to food security
because the amount of protein provided by home consumption (valued at
less than 2,000 FCFA per adult equivalent per year) is quite small. In
very rough terms, a single chicken costs about 750 FCFA and a kilo of
beef from 500 to 700 FCFA, depending on the zone. This translates to a
maximum of 4 kilos of home-produced meat per adult equivalent per year.
Purchases
The relative importance of different types of livestock as a
percent of total livestock expenditures varies with region. In 1988/89,
nearly all of the livestock purchased in the northern and southwestern
Peanut Basin and in Senegal Oriental consisted of sheep and goats.
During the same year, expenditures on cattle and small ruminants were
about equal in the southeast.
In the Colobane zone of the central
basin,
expenditures on small ruminants were almost double those on
cattle, while the reverse was true in the Niakhar zone of the same
region.
Livestock purchases in all zones increased 2.5 to 3-fold between
harvest years.
Part of this increase is probably due to the incomplete
enumeration of transactions in the postharvest period of harvest year
1988.
The relatively good crop harvest in 1989, however, did provide
many households with a big increase in cropping income (see Chapter 5),
thereby encouraging a higher level of livestock purchases than in
harvest year 1988.
In the central basin, the composition of purchases in the second
year was similar to the first.
Niakhar continued its preference for
cattle and traction animals while Colobane gave priority to small
ruminants.
In the southwest and Senegal Oriental, the
1988/89
preference for small ruminants was shifted to cattle in 1989/90.
Poultry purchases to increase stocks were minor (O-6 percent of
purchases) except in the Colobane zone in the Central Basin, where they
were 15 percent of purchases in 1988.'
REASONS FOR SELLING
In all regions, livestock was sold largely to obtain food. The
portion of annual sales that went for food is generally in the 30-50
percent range.
The exceptions are Colobane in 1989/90 and Dioly in
1988/89 - - 19 and 24 percent respectively.
The next most important
reasons households sold their livestock, both as a portion of the budget
and frequency cited, was to improve their herd by reinvesting in
different animals.
Only in two regions during the survey period
* Note that purchases for immediate consumption are not considered
in calculating net livestock income.
,.
._
t
.
.
.
.,
;
*
!
‘.
,’
:
134
(central and southwest basin) was livestock sold to obtain agricultural
inputs.
It should be noted, however, that only the principal reason for
selling an animal was enumerated.
Frequently the money obtained from
the animal sale was used to purchase food as well as other goods. In
transactions data for purchases of agricultural inputs (particularly
seeds), "animal sale" was frequently cited as the source of financing
for the purchase.
There were a number of other reasons for selling livestock, such as
paying debts, paying for ceremonies, buying gifts, financing trips, and
assorted other activities.
Except in Senegal Oriental where 19 percent
of monthly sales in 1988/89 went for ceremonies, and the northern Peanut
Basin where 17 percent of sales went for gifts in 1988/89, the other
reasons comprised less than 15 percent of annual sales and there is no
pattern or consistency among the regions. The reasons households cited
for selling livestock during peak periods are discussed in the section
on seasonality.
The major conclusion to be drawn from this section is that animal
husbandry plays an important role in ensuring food security by providing
a relatively liquid form of savings that can be quickly transformed into
cash when food supplies run low.
Animal husbandry's direct contribution
to meat consumption is relatively low, with small ruminants slaughtered
at Tabaski accounting for a large share of this consumption.
DISTRIBUTION OF NET INCOME FROM LIVESTOCK TRANSACTIONS
Livestock transactions are one of the most difficult activities to
treat in a general context because the level of holdings as well as the
importance of transaction income can be extremely variable both within
and across zones.
Appendix Figures A7.1-A7.11 are histograms showing the distribution
of household income from livestock transactions for each zone and each
year of the study.
The following conclusions can be drawn from these
histograms:
Incomes are most skewed in the zones with the highest average and
median
livestock incomes.
The northern Peanut Basin and the
southeastern Peanut Basin have average livestock incomes in the 6,000-
8,000 FCFA range and median incomes in the 2,000-4,000 FCFA range. The
gap between lowest and highest incomes in both zones is extremely large
(-3,000 to 72,000 in the north and -500 to 300,000 in the southeast for
the 1988/89 harvest year).
While average incomes in other zones are consistently higher than
median incomes, the gaps tend to be smaller than those in the north and
southeast, suggesting that average values are less influenced by a few
households with very high livestock incomes.
.I’
:
135
The range from lowest to highest incomes is also much smaller in
these zones.
With the exception of Senegal Oriental in 1989/90, the gap
between high and low incomes ranges from 12,000 to 37,000 FCFA per year.
One household in Senegal Oriental had a particularly high income in
1989, pushing the gap for this zone up to 66,000 FCFA.
Ndiaye (1993) estimated Gini coefficients of livestock holdings
using IFPRI/ISRA data for three of the six rural zones.
His results
(see Appendix 7.3) confirm that the distribution per capita is quite
skewed with coefficients greater than .65 for cattle and .42 for small
ruminants.3
The implication of this extremely variable level of income from
livestock activities is that programs to help households increase
livestock production will have important equity effects. Care must be
taken to ensure that any policy interventions to improve livestock
production do not accrue primarily to those households already at the
top of the pyramid for livestock income.
Another policy implication can be drawn from the fact that one of
the zones with the biggest herds of livestock (Niakhar in the central
Peanut Basin) has the smallest livestock income.
There are clearly
cultural explanations for this behavior (see comments in Chapter 4);
however, analysts should be asking if there is not some way to encourage
greater economic exploitation of livestock holdings in this zone, which
exhibits the lowest levels of per capita income in the entire survey.
CONTRIBUTION TO TOTAL LEVEL OF RURAL INCOME
Chapter 4 discussed the sectoral composition of total household
income in detail.
The analyses presented (Table 4.5) show that net
transactions income from livestock contributes less than 25 percent of
total household income across all zones and years.
Zones where livestock was most important were the northern Peanut
Basin (17 percent of total income) and Colobane in the Central Basin
with 23 percent in 1988/89 (but only 5 percent in 1989/90).
In both
cases, the large share occurred in 1988 after the very bad harvest.
Livestock was more than 15 percent of income in the southeastern
Peanut Basin and Senegal Oriental during both years. These zones are
not traditionally identified with livestock production. It appears that
pasture and land constraints further north are pushing livestock
production toward the south.
In these zones, livestock is the second
most important single source of income after crop production.
3
The Gini coefficient is an indication of how skewed the
distribution of livestock is across households; a coefficient close to
zero implies a relatively equal distribution while a coefficient
approaching 1 means the distribution is unequal.
136
Zones with relatively small shares of livestock income (5 percent
or less) are the central Peanut Basin (Niakhar in both years and
Colobane in 1989) and the southwestern Peanut Basin (Passy).
Low
livestock incomes in the central basin appear to be due to investments
(purchase of new animals for the herd) that exceed sales severalfold.
This is a pattern that is typical of the Serer ethnic group, which
predominates
in Niakhar and is important in Colobane.
In the
southwestern Peanut Basin, the overall level of transactions is quite
low, with the value of purchases 2-3 times sales.
Tables in Appendix 7.2 present information about changes in stocks
during the two study periods.
Although no value is placed on changes in
stocks resulting from births, deaths, and gifts, the data do help one to
draw some tentative conclusions about what role animal husbandry
activities play in changing a household's total assets. For example, in
the Niakhar zone, which has a very low level of livestock income, stocks
increased substantially during 1988/89 (38 new goats).
I n 1989/90,
stocks of goats declined due to deaths (down 7 animals net) while sheep
increased by 6.
This suggests that livestock activities are probably
more profitable to households in this zone than the net transactions
analysis presented above suggests.
A quick look at the tables for other
zones suggests, that in general, livestock herds increased during the
two-year survey period.
Additional analyses have been done by Ndiaye (1993) to combine
transactions income with capital gains and losses to obtain a full
picture of net income from raising small ruminants in harvest year
1989/90.
A summary table of his results for the central and
southwestern basins is presented in Appendix 7.4.
Ndiaye's analysis shows that poorer households (those with average
income per AE less than the median income for the zone) lost money on
small ruminants in the Niakhar zone of the central basin and in the
southwest.
The primary cause was a substantial drop in the value of
ending inventories due to high death rates.
Poorer households in the
Colobane zone realized a net income of 12,500 FCFA (6,500 per small
ruminant head).
The wealthier half of the sample in each zone (income per AE
greater than the median) realized a net profit. Colobane had extremely
good returns (21,600 FCFA per small ruminant head), while the profits in
the southwest and Niakhar were more modest (3,800 and 500 FCFA per
head).
These results suggest that poorer households may not have adequate
resources for animal feed and medical care, thereby limiting the extent
to which they can supplement cropping income with livestock income.
137
SEASONALITY OF SALES
Appendix 7.2 contains tables illustrating the seasonality of sales
for each zone and year.
The 1988/89 harvest year was very bad in the northern Peanut Basin
and 41 percent of livestock sales for the year took place in October and
December.
These sales were primarily for food, ceremonies, and gifts,
and consisted mostly (82 percent) of cattle, with sheep and goats making
up the rest.
In the Niakhar zone of the central basin, which also suffered in
1988/89, 69 percent of the year's livestock sales occurred in June. A
quarter of these sales were for food, another quarter for reinvesting in
livestock, and the remaining share (42 percent) for unspecified reasons.
Sales were mainly cattle (42 percent -- this was the only month cattle
were sold all year) and hogs (28 percent).
In the 1989/90 y e a r , sales peaked in May and July, with the
livestock sales consisting of cattle (42 and 46 percent) and goats (40
and 32 percent respectively).
Sales in May were to finance trips, to
buy food, and for ceremonies and other uses, while all of the sales in
June were for livestock reinvestment.
In the Colobane zone in the central basin, peak sales in the
1988/89 harvest year occurred during June and July, with the bulk of
sales being sheep sold for food, agricultural inputs, and other reasons.
The following year also saw peak sales in June and July composed of
sheep (44 percent), cattle (28 percent), and horses (26 percent). These
sales were mostly to finance another activity or to pay off debts.
In the southwestern Peanut Basin during the 1988/89 season, the
heaviest livestock sales were distributed fairly evenly from February
through May.
Horses made up 58 percent of sales in February, but
farmers shifted to cattle after February (46, 72, and 73 percent
respectively for March, April, and May).
The horses sold went primarily
for reinvestment in livestock and some for food. The later sales were
for food, to pay off debts, taxes,
and for livestock reinvestment. In
1989/90, sales peaked in January, March, and May, with horses making up
70 percent of sales in January, and cattle in March and May (90 percent
and 58 percent respectively).
Debts absorbed most of the income from
sales in January while livestock was sold almost exclusively for food in
March and for food and agricultural inputs in May.
In the southeastern Peanut Basin, sales peaked in March of 1988,
with cattle supplying 64 percent of the sales, more than half of them
for livestock reinvestment. In the next harvest year, sales peaked in
March and May and consisted mostly of cattle and horses.
In March
nearly all the sales were for food while in May about a quarter were for
agricultural inputs.
,
’
)’
;
.
.
.
:
1
.
.
138
In Senegal Oriental, peak sales in the 1988/89 harvest year took
place in January and February and consisted largely of cattle (75 and 97
percent respectively).
Some of the January revenue went for food but
also to finance trips and for other uses.
In February almost half of it
went for ceremonies.
In the 1989/90 season, peak sales occurred in
October and December and June, and again were largely cattle (92 and 81
percent).
Some of the sales in January were for food, but a number of
the animals were sold because they were sick.
In June more than 60
percent of sales were for food.
Only in the northern Peanut Basin during 1988/89 and in Senegal
Oriental during 1989/90 did households sell animals and purchase food
early when prices were relatively low.
Most sales for food occurred in
June/July when cereal prices were high. The question is, does the value
of an animal increase enough between harvest and June/July to warrant
this type of behavior? The experience in Senegal Oriental suggests that
it does not because prices of goats and sheep were very low in June and
July in 1989/90 when many households were in distress and selling
animals.
Households in Colobane in the central basin and those in the
southwestern basin are the only ones that seem to sell animals in large
numbers to pay for agricultural inputs or to pay off agricultural
credit.
In the southwest this is related to credit from a confectionery
peanut program.
Farmers with poor harvests will liquidate other
resources
to reimburse
suppliers for credit to maintain their
eligibility for the program in the following year.
In Colobane, peanut
production is very important and farmers sell animals to purchase seed.
The links between livestock and household food security have
already been mentioned in Chapter 4.
Information presented here on the
seasonality and reasons for livestock sales only serves to underscore
the important contribution that animal husbandry makes to smoothing
income across years and seasons.
This income smoothing helps households
cover cereal production shortfalls with purchases and, in some cases,
provides liquidity for purchasing agricultural inputs, particularly
peanut seed.
SUMMARY OF MAIN POINTS
Animal husbandry income is used by most households to smooth total
household income and compensate for crop shortfalls.
As a result,
livestock income tends to fluctuate across years (between 5 and 20
percent of total household income,
in general) and is negatively
correlated with cropping income.
Interannual fluctuations are lower,
however, in the southeastern Peanut Basin and Senegal Oriental where
livestock income was consistently at least 15 percent of total household
income in both years.
139
Income from livestock sales is used, above all, to purchase food.
The next most common reason for sales is to reinvest in different types
of animals.
Distribution of livestock assests across households is quite skewed
with Gini coefficients in the .4 to .5 range for small ruminants and in
the .7 to -9 range for cattle (Ndiaye, 1993).
Distribution of livestock income across households is also quite
skewed with many households having zero or negative incomes and a few
households having very high incomes.
Size of herds is not always positively correlated with the level of
net income from livestock transactions nor with the share of livestock
income in total income.
The Colobane zone of the central basin has one
of the smallest herd sizes but one of the higher levels of livestock
income per AE.
The Niakhar zone in the central basin has one of the
largest herd sizes but very low levels of transactions income.
The
former zone appears to raise animals as a means of supplementing income
while households in the latter zone appear to use livestock as a form of
wealth accumlation
and savings.
Policies to improve livestock
productivity will need to develop programs that respond to these
different types of strategies.
Using IFPRI/ISRA data, Ndiaye (1993) shows that poorer households
tend to have much higher loses than the rich due to animal deaths. As
a result, the poor realize negative net livestock income when changes in
the value of animal stocks are considered. If the livestock sector is
to serve as a complement to the cropping sector, enabling both poor and
rich households to better cope with cropping shortfalls, greater
attention will have to be paid to improving the health of livestock
raised in poorer households.
:
.
.
:
.
140
REFERENCES
Ndiaye, 0. 1993.
Micro-level Small Ruminant Transaction Analysis and
Livestock Po1ic.y in Seneqal.
Masters Paper,
Michigan State
University, E. Lansing.
,
.:
.
:
:
s
:
‘.
141
8.
FOOD CONSUMPTION PATTERNS
INTRODUCTION
The primary objective in examining food consumption behavior in
this study is to understand how different types of households will
respond to changes in consumer and producer prices for basic food crops.
Such knowledge will make it possible to predict the probable outcome of
proposed policy initiatives and to evaluate the impact of such policies
on households that differ by income level, location, and sectoral
composition of income, for example.
This chapter presents brief summaries of food consumption for each
of the agroclimatic regions by zone and compares factors that influence
the level, composition,
and source of food consumed by the sample
households.
More detailed analyses for each zone and village are given
in the appendixes for this chapter, which also contain the tables and
charts upon which the analyses are based.
While hypotheses are offered about the role of prices, income, and
market village infrastructure in shaping food consumption behavior; the
primary objective of this chapter is to describe current patterns. More
rigorous,
quantitative analysis of how prices, incomes, and other
relevant variables influence food consumption will be conducted in the
future using econometric techniques.
Quality of Harvest
One expects food consumption levels and composition to vary in
rural Senegal with the quality and composition of the harvest.
The
quality of the harvest depends to a large extent on the climatic
conditions,
amount of crop disease, and level of pests.
These factors
are highly variable in Senegal; one village may have terrible rains
while a neighboring village no more than 5 kilometers away might have
good rainfall.
The next few paragraphs provide a general description of
the quality of the harvest for the two years of the study.
Due to
substantial
intrazone variation
in the quality of the harvest,
situations that depart from these generalizations will be mentioned when
necessary.
The 1988/89 crop production year ended with a very poor harvest in
the northern Peanut Basin and in a large part of the central Peanut
Basin.
Poorly distributed rains were responsible to some extent, but
142
the main culprit was a locust attack that occurred just before the
harvest.
Cereals suffered the most from this attack.
The southern parts of the Peanut Basin and Senegal Oriental were
not particularly affected by locusts in 1988 and the harvest was
considered average in most parts of these zones.
The 1989/90 cropping season was one of the best in the past ten
years for most of Senegal.
Although pockets of irregular rainfall
persisted in some zones,
the 1989/90 gross value of agricultural
production per adult equivalent (AE) across all survey areas was 1.5 to
2.4 times greater than that for 1988/89 (see Chapter 5 for a full
discussion of crop production).
Nature of Data and Analyses Performed
Data on food consumption in each of the six rural zones were
collected over the two-year study period (HY88 and HY89); data on the
two urban zones were collected for a one-year period (July 1990~June
1991) .'
The surveys were designed to provide information on:
Caloric adequacy per adult equivalent (AE);'
Composition by product;
Sources (home production, purchase, gift);
Seasonal variation in levels, composition, and sources;
Influence of market villages on consumption patterns;
Differences in consumption between the poorest and the
richest income quartiles.
The data were collected fortnightly. The interviewer asked the
respondent to recall, for the previous day, household consumption of
cereals, pulses, tubers, animal products, oil, and bread -- at home and
outside the household.
1
Several preliminary versions of this chapter have been
distributed
in both French and English.
Among them are "Note
d/information sur la consommation des cereales et des legumineuses en
milieu rural senegalais,"
issued in July 1991, and project document 7,
Food Consumption in the Peanut Basin and Seneqal Oriental: October 1988
- September 1989 issued in February 1992.
Results presented here
supersede those presented in all previous reports where weighting for
disproportionate sampling of market villages was not taken into account.
* In calculating adult equivalents present at each meal we use the
following rules adapted from FA0 1975 standards:
adult male 15
2 years = 1
adult female 2 15 years = .73
boy 5-14 years =
.83
girl 5-14 years =
.70
boy < 5 years =
.44
girl < 5 years =
.44
.;
;
143
The number of household members and guests present for each meal
was also recorded, as were dishes sent to others or received as gifts.
The woman who cooked the day before was the principal respondent; those
having purchased food away from home were interviewed separately about
their consumption; data on food consumed away from home are not analyzed
in this report.3
The "Office de recherches sur l'alimentation et la nutrition
africaine" (ORANA) has established 3000 kilocalories (kcals) per day as
the intake norm for a moderately active Senegalese male.4
Populations
consuming less than 80 percent of this recommended level (fewer than
2400 kcal) are considered to be "at risk." The IFPRI/ISRA analysis
evaluates the extent to which the sample households have consumed at
least 80 percent of the recommended norm and, therefore, are no longer
"at risk".
Throughout the discussion, we are interested in the extent
to which the 2400-kcal level of "minimum requirements" is met by sample
households.
Calorie calculations for cereals took into account processing
losses
and nonconsumption of bran for millet, sorghum, and maize.
Using secondary data as well as a mini survey of processing losses for
a subset of sample households in Senegal Oriental, the following caloric
availability was established for coarse grains:
Calories Available per 100 grams of Whole Grain
Product
Before Processinq
After Processinq
Millet
353
281
Sorghum
354
281
Maize
357
285
Calories available for other products are taken from ORANA and are as
follows:
Product
Calories per 100 qrams
Peanuts (shelled)
550
Cowpeas (shelled)
342
Milk
79
Bread (wheat)
252
Vegetable Oil (peanut)
884
Cassava
146
Rice (processed)
363
3
See Fall, Kelly, and Reardon (1989) for details of survey
methods.
4 ORANA, "Cartes des calories et des vitamins pour 150 aliments
africains," no date.
' See Appendix 8.2 - for a fuller discussion of this issue.
.
,-,
,
:
.:
‘
.
”
‘.
,
144
The results reported are also weighted to adjust for non-
proportional sampling of households in market villages (see Chapter 2).6
A few weaknesses in the data need to be pointed out. The results for
the 1988/89 harvest year are based on data for just the period December
1988 through September 1989.
Given that the harvest season (October and
November) is missing, these results understate the overall consumption
level and the share of home production, while overstating the share of
rice.
The biases are probably minor in the northern and central Peanut
Basin because the harvest was relatively poor in these zones.
The analysis examined calories obtained from cereals and pulses
plus those obtained from vegetable oil, milk, bread, and tubers. The
term "total calories" means total calories obtained from the products
analyzed.
In discussions of composition and source, "percent of total
calories" means percent of total calories for the food qroups analvzed
to date.
The major food groups not addressed in this report are sugar,
vegetables, meat, and fish.
Given that the primary objective of the research was to provide
data for an examination of agricultural price policies with respect to
cereals and pulses, a great deal of time and effort was devoted to
ensuring high-quality data for these products.
The "other food"
products examined are more difficult to measure because of the small
quantities consumed and irregular patterns of consumption. Estimates of
average calories consumed per AE for these products exhibit very high
coefficients of variation and must be used with caution.
Given the
dearth of information on consumption of these products, it is important
to present what is available, fully acknowledging that it is a very
rough estimate of calories obtained from these products.
In other chapters (particularly Chapter 4), we divided sample
households into income terciles to examine the relationship between
income and different types of production or consumption behavior. In
this chapter, we compare the consumption patterns of the richest
quartile with that of the poorest quartile to examine the income effect.
This was done because the use of terciles failed to provide us with
useful information on how consumption varies by income group. The range
of caloric intake levels in the sample is not extremely large.
This
made it difficult to find significant differences when the sample was
divided into only 3 income groups.
Even with the use of quartiles,
differences in cereal and pulse consumption between the wealthiest and
poorest quartiles are not statistically significant. We believe that
the greatest differences in consumption between rich and poor relate to
the amount of expenditure on products not included in the caloric
analysis presented in this chapter (fish, meat, vegetables, and sugar).
This hypothesis
needs to be tested, however, with econometric
techniques.
6 This weighting produces results that are occasionally different
from preliminary,
unweighted results reported in Project Document 7.,
Food Consumption in the Peanut Basin and Senegal Oriental: October 1988
- September 1989 by Kelly et al., 1992.
145
SUMMARIES OF FOOD CONSUMPTION BY ZONE
This chapter is a summary of a more detailed zone by zone
discussion of results that is contained in the Appendix along with a set
of detailed tables presenting analysis of results for each zone and
village.
Tables 8.1 and 8.2 present summary statistics on the levels and
composition of calorie intake by zone and year.
The summary statistics
illustrate the extent to which access to market infrastructure and level
of income influence consumption levels, composition and sources.
Northern Peanut Basin: Zone of Saqatta
The typical AE in the northern Peanut Basin consumes about 2220
calories daily from products analyzed to date. Ninety-three percent of
minimum requirements are covered by these products, suggesting that - on
average - the typical AE will not fall in the "at risk" category when
unanalyzed products that are regularly consumed in the zone (sugar,
meat, fish, and vegetables) are added.
The typical northern Peanut Basin AE relied on purchases for 82
percent of cereal/pulse and 90 percent of cereal calories in HY88.
Purchasers in the zone favor imported rice, which accounts for 66
percent of cereal calories consumed.
Although the Northern Peanut Basin
is not very far from rice producing zones of the Senegal River Valley,
there was no evidence that any rice from the river valley was being
consumed by sample households.
Sixty-eight percent of households in the sample averaged less than
the 2400 minimum requirements during HY88. The poorest income quartile
of the sample consumes an average of 1925 cereal/pulse calories per AE.
This is 200 calories less than average consumption by the richest
quartile.
It is unlikely that households in the poorer group reached
the 2400 minimum requirements by purchasing foods to supplement cereal
and pulse consumption.
Although seasonality in consumption is not extreme (ranging from
93 to 105 percent of the annual average, depending on season),
consumption drops during the March-July period and rises as soon as the
cowpea harvest becomes available in August.
Recent Government of Senegal policies to increase cowpea production
in this zone appear to have had a strong positive effect on ensuring
food security during the August-October period when the new cowpea
harvest is available but millet and peanuts have not yet arrived at
maturity.
Surprisingly, households and villages with better harvests (higher
value of production per AE) tend to have lower levels of caloric intake.
The households with better harvests also tend to have less diversified
incomes with greater than 30 percent of total income coming from
cropping while households with better caloric intake levels realize only
146
lo-15 percent of total income from cropping. These results suggest that
food security in this zone is better ensured by concentrating on non-
cropping activities.
The market village has more diversified consumption patterns, with
less reliance on cereals and pulses than the nonmarket villages. This
is as expected given that "other" foods are more readily available in
market villages and that opportunities to earn cash to purchase "other"
foods also tends to be greater.
Looking at only cereals and pulses,
average calories per AE differ across villages by as much as 555
calories per day.
When "other foods" are added, the maximum difference
drops to 264 calories.
More work is needed in analyzing the links between income
diversification and food security in Senegal. Our impression is that
food security in this zone could be enhanced by rural development
policies that help farmers to diversify income sources while using crop
production as a means of supplementing non-cropping income (see Chapter
4 for more discussion of this issue).
Central Peanut Basin: Zone of Niakhar
During HY88, the typical household in the central Peanut Basin zone
of Niakhar consumed only 2162 calories per day per AE from the product
groups analyzed to date.
It is unlikely that products not yet analyzed
(sugar, meat, fish, and vegetables) would have pushed this level to the
minimum requirement of 2400 calories.
HY89 was a better year, with
average consumption in the zone reaching 2415 calories per day per AE.
In HY88, 68 percent of households failed to achieve an average
level of consumption (from products analyzed) of 2400 kcal/AE.
The
share below the minimum requirements dropped to 32 percent following the
better 1989 harvest.
Households in the lowest income quartile consumed 154 fewer
cereal/pulse calories per AE than the highest quartile in HY88 and 279
fewer in HY89.
In both years it is unlikely that households in the
lowest quartile consumed the minimum 2400 calories per day, even after
consumption of all other foods is taken into account.
The top income
quartile passed the 2400 level in HY89 with cereal/pulse products alone.
It is not clear that the wealthy quartile passed the 2400 line in HY88
because cereals and pulses covered only 2065 calories, leaving about 350
calories to be obtained from other products in this zone,
which exhibits
fairly low consumption of noncereal pulse products.
In sum, the zone appears to have some problem ensuring food
security at the minimum required level of 2400 calories.
Following poor
harvests, both rich and poor have problems; following good harvests, the
poor continue to have problems while the wealthy pass the "at risk"
level and move toward normal consumption.
The typical central Peanut Basin household in Niakhar is much more
self-reliant in terms of cereal and pulse consumption than other parts
.”
*
‘.
:
:
..;
.,
;
147
of the Peanut Basin.
Following the poor 1988 harvest, 21 percent of
total cereals and pulses were purchased while just 8 percent of m
produced cereals and pulses were purchased; this dropped to 5 percent in
HY89.
Although this appears commendable given the government's
objectives of encouraging self-sufficiency in cereal production, one
must not lose sight of the fact that consumption levels were not
adequate in HY88.
The degree of self-reliance would diminish if all
households in the zone had adequate resources to bring their consumption
up to the recommended minimum level through purchases.
Farmers in this zone tend to give more attention to growing their
own millet than to expanding their peanut production, despite the land
shortage that. plagues the zone.
Farmers in other areas of the Peanut
Basin (see the discussion of the Colobane zone) have been turning
increasingly to peanuts and relying on the market for cereals.
Imported rice is not as large a share of the overall diet (15
percent of cereals in HY88 and 14 percent in HY89) as it is further
north.
Although prices and incomes play some role in determining the
mix of rice and coarse grains, the Serere people living in this zone are
known to have stronger preferences for millet than other ethnic groups
in the Peanut Basin.
Rice is a relatively small part of total cereal/pulse consumption,
yet it is a very important part of purchases (65 percent of all
cereal/pulse purchases in HY88 and 73 percent in HY89). Interestingly,
the share of rice in purchases increased after the better harvest. This
was most likely due to (1) higher incomes that permitted purchases of
rice and (2) a desire to introduce some variety into the essentially
millet-based diet in the zone.
In looking at seasonality, we find that households in the zone tend
to consume their own production first, increasing purchases of both rice
and millet when stocks of millet run out.
This results in the bulk of
cereal products being purchased when the price of millet is at its
highest point relative to the stable (government-controlled) price of
rice.
This leads one to ask whether rice consumption would be lower if
cereal purchases were made when millet prices were still low.
The phenomenon of an early and abundant 1989 cowpea harvest that
helped families supplement their calories during the end of the rainy
season occurred in this zone, but on a smaller scale than in the
northern Peanut Basin where more cowpeas are cultivated.
Unlike the
northern Peanut Basin, where increased cowpea consumption was
concentrated in the lowest income quartile, both rich and poor exhibit
approximately the same share of cowpeas in total cereal/pulse
consumption.
Although the share of imported rice in total cereal/pulse
consumption is not strongly affected by income following a good harvest,
the share of imported rice in the diets of market village households is
about double that of households in other rural villages both years'.
_’
.
‘.
:. . .
'i
.-
:
:
I
.
'.
148
Households in market villages also purchase a larger share of their
cereals (about 35 percent versus 16 percent for the nonmarket villages).
Central Peanut Basin: Zone of Colobane
The typical household in this zone consumed on average 2624
kilocalories per adult equivalent in harvest year 1988/89 and 2462 kcal
in HY1989/90.
This average consumption met the minimal calorie
requirements for both harvest years.
Despite the acceptable average, 50
percent of households in HY88 and 32 percent in HY89 exhibited average
intake lower than the 2400 kcal/AE "at risk" level.
Consumption was
based primarily on cereals (74-77 percent),
particularly millet, with rice as a supplement. Although cereals remain
import, Colobane has the lowest share of cereals in total calories of
all rural zones but those in the northern Peanut Basin.
This zone also
exhibits the second highest level of imported rice consumption (about 30
percent of cereals).
Rice consumption in the market village is two
times higher than in the nonmarket villages.
The zone also has an
extremely high level of vegetable oil consumption, which resulted in
more calories coming from the "other food" category than from pulses in
1988/89.7
This zone has the second highest dependence on purchases (following
the northern Peanut Basin), with more than 40 percent of cereal/pulse
calories and 48 percent of cereal calories being purchased.
This
behavior is especially marked in the market villages and in both the
highest and lowest income quartiles.
In poor households, there was no
substantial increase in the quantity of home-produced millet consumed
following the better 1989 harvest.
Wealthy households exhibited much
greater consumption of home-produced millet in 1989-90, suggesting that
they were more able to profit from better rains than the poor.
Seasonality is not a problem, given that average annual consumption
levels are adequate, and the seasonal variation never goes below 93
percent of the annual average.
Demand for rice does vary by season and
harvest year.
The pattern differs from that in the Niakhar zone of the
Central Basin.
Southwest Peanut Basin: Zone of Pass-v
On the whole, the typical household in the southwest basin consumed
an average of 2451 kcal/AE/day, which supplied the minimal calorie
requirement for the two harvest years.
In both years, 40 percent of
households in this zone failed to average 2400 kcal/AE. This was the
lowest rate of inadequate consumption in HY88 and one of the lowest in
HY89.
' Transforming peanuts into oil is an important income-generating
activity in this zone.
149
The share of calories coming from cereals is highest in this zone
(83 percent) while the contribution of "other foods" is quite small (4-5
percent).
Overall consumption was based on millet (65-66 percent of the
cereals and pulses consumed), rice (more than 20 percent), and peanuts
as a supplement (a little more than 10 percent). Production remains the
primary source of calories but purchases are common (39 percent of
cereal calories in HY88 and 27 percent in HY89).
Rice is the most frequently purchased product -- 68 and 83 percent
of cereal/pulse purchases in 1988 and 1989, respectively. The share of
rice purchases increases following the better harvest, probably to add
variety to the diet.
As in the central Peanut Basin zones, seasonal variation in caloric
intake is not high (92-106 percent of annual average).
Surprisingly,
cereal purchases and imported rice account for about the same share of
cereal calories in one of the non-market villages as they do in the
market village.
When both non-market villages are averaged together,
however, the market village rice consumption and purchases are slightly
higher.
In 1988 there was little difference in consumption levels of the
rich and poor.
In 1989, however, the rich consumed about 400 calories
more.
The poor in this zone ate a smaller share of rice (35 percent of
cereal calories) than the rich (43 percent) in 1988. In 1989, rice was
a smaller share of the cereal diet for both rich and poor, and rich
continued to consume a larger share than the poor (36 versus 29
percent).
Southeast Peanut Basin: Zone of Nsanda-Dioly
The typical household in the southeastern basin consumed on average
2232 kcal/AE/day in 1988/89 and 2304 in 1989/90.'
This represents on
average, 95 percent of minimal requirements over the two harvest years.
Although average
levels approached minimum needs, 70 percent of
households in HY88 and 60 percent in HY89 fell below the 2400 kcal/AE
minimum needs.
This suggests the existence of a serious food problem
that needs to be clarified.
In the present analysis, nothing suggests
that including other food groups would make up the calorie deficit.
There is a big difference between overall consumption levels in the
market village (92-97 percent of minimum needs covered) and the
nonmarket villages (78-90 percent).
On average, the nonmarket villages
had a higher production sufficiency ratio than the market village in
' The draft report showed much lower consumption levels for this
zone (2024 and 2006 kcal/AE for HY88 and HY89, respectively). A review
of the data files and analysis resulted in a number of corrections.
Time did not permit us to rerun all the analysis tables in the Appendix
nor to revise all the graphs.
The text of the final report, however,
has been modified by removing comments concerning the extremely low
average consumption levels for this zone.
;,
I
.
‘.
150
both years (see Chapter 5).
About one quarter of households in the
nonmarket villages, however, had an extremely low (0.24) production
sufficiency ratio in 1989/90 and one equal to that in the market village
(0.44) in 1988.
This suggests that the overall low consumption levels
in the zone are probably being strongly influenced by consumption levels
for this set of about six households.
The results also suggest that
nonmarket villages in the zone have much more trouble purchasing
cereals.
The extent to which this is due to inadequate supply (the
nonmarket villages are quite isolated) versus lack of purchasing power
needs to be examined more thoroughly.
Grains supply the essential food needs, in particular millet (67-69
percent of grains consumed), sorghum (9-12 percent), peanuts (11-14
percent), and finally imported rice (6-8 percent). Less than 5 percent
of calories came from other foods analyzed (mostly vegetable oil in 1988
and milk in 1989).
About three-quarters of the calories consumed came
from production and a quarter from purchases. Market villages purchase
a much larger share of cereals than nonmarket villages.
Seasonality is evident in rice consumption, which is high at peanut
marketing time when money is available and low during the rest of the
year.
Peanut consumption is also seasonal -- high at harvest and
declining regularly until the next harvest.
Millet and sorghum consumption patterns differed during the two
years.
The extent to which this was a price phenomenon is examined in
Chapter 9.
Finally, consumption of cereals and pulses by the poor was only 71
percent of that of the rich.
In no other part of the basin was one able
to observe such a large difference.
Following the 1989 harvest, the
rich were able to reduce purchases from 25 to 14 percent while the poor
increased purchases from 24 to 28 percent,
Clearly, there are serious
constraints preventing the poor from realizing the full benefits of a
better rainy season.
Seneqal Oriental: Zone of Missirah
The typical household in the Senegal Oriental consumed on average
2113 kcal/AE/day in 1988/89 and 2047 in 1989/90.
Given the very low
level of "other food" consumption enumerated in the zone, it is unlikely
that analysis of the remaining products (sugar, meat, fish, vegetables)
will bring average consumption to a satisfactory level.
On average,
only 85-88 percent of the minimal level of calories was met during the
two harvest years; this was among the lowest in the zones surveyed. The
share of households not meeting minimum requirements was also high --63
percent in HY88 and 59 percent in HY89.
These low consumption levels are surprising because the higher
rainfall in this region should provide better yields and therefore
better food security.
Although the zone had a higher cereal sufficiency
ratio in 1988/89 than all zones but the southwestern Peanut Basin,
cereal production still only provided 55 percent of minimum caloric
151
needs (45 percent if sales are netted out).
It appears that the
nonagricultural
income
stream,
with a large
share
coming from
intermittent migration remittances, failed to cover the gap, leaving
many households with inadequate consumption levels.
A peculiarity of this zone is that cereal consumption is very
diversified.
No grain dominates the others in the overwhelming fashion
that millet does in the most part of the Peanut Basin or imported rice
does in the northern basin or in the urban zones. Cereals and pulses
supply essentially all the calories consumed, especially maize (39-43
percent of cereals and pulses consumed), sorghum (15-20 percent), millet
(16-20 percent) peanuts (16-18 percent), and finally imported rice (4-6
percent).
The very large share of calories derived from peanuts (about
equal to those of millet and sorghum) is surprising in Senegal Oriental.
Does this suggest that the Peanut Basin is moving toward the southeast?
About three-quarters of the calories consumed come from production and
a quarter from purchases.
Gifts are more important in this zone than
elsewhere (3-5 percent) but still supply a very small share of total
calories.
The market village households had lower average caloric intake in
HY88 for cereals, pulses, and "other" products.
The difference
remained, but was less pronounced in HY89. Share of cereals purchased
and share of rice were consistently higher in the market village. The
difference in share of consumption coming from rice in market versus
non-market villages was much more pronounced in this zone where the
ratio was 1 to 4 in HY88 and 1 to 10 in HY89 (the ratio for other zones
was about 1 to 2).
Differences between the richest and poorest quartiles were not very
large, and not consistent across years.
Total caloric intake was about
the same.
Share of purchases and share of rice was larger for the poor
in HY88 but the pattern reversed itself in HY89.
Inter-annual variation in the composition of consumption is more
important in this zone than elsewhere.
Purchases of maize and sorghum
were more important relative to rice in 1989/90. This was probably due
to an excellent maize harvest which reduced the maize price
substantially during February-June 1990 (see Chapter 9).
Intra-annual seasonality is also more pronounced in the zone than
elsewhere.
The variation around the mean level of caloric intake ranged
from 91 to 111 percent.
This is not very different than the range in
other zones.
It has much more serious consequences in this zone,
however, due to the low average annual consumption level. Consumption
levels were highest during the harvest and lowest during the rainy
season.
Maize plays a very important role as a hungry season crop
because it can be consumed "green" toward the end of the rainy season
before other cereals become available.
Caloric intake may well be
higher during the "green" maize season than reflected in our data
because it was extremely difficult to enumerate the consumption of this
product, which is eaten in large quantities as a snack food.
152
Urban Zone:
Town of Kaolack
Approximately 50 percent of Kaolack households surveyed failed to
attain an average intake level of at least 2400 kcal/AE. The average
daily consumption per adult equivalent in this zone is about 2100 kg
from the products analyzed to date.
This represents a level covering 87
percent of the 2400 kcal/AE minimal caloric requirements. Given that
products not yet analyzed are likely to be more important in urban than
in rural zones, it is logical to think that average intake levels will
approach 2400 kcal/AE when all products are considered.
The typical Kaolack household purchases nearly all the products
that it consumes.
Imported broken rice constitutes the principal
element of the food regime, followed by millet, vegetable oil, and
peanuts.
Consumption of cereals and pulses appears to be most important
during the period April to September relative to the other months of the
year (October to March).
Urban Zone:
Town of Tambacoumba
During 1990, a typical household in the urban zone of Tambacoumba
consumed 2328 kcal/AE/day.
This attains nearly total coverage of the
minimal level of calories needed with only the products already
analyzed.
Despite the generally acceptable average level, 53 percent of
households in the sample were
below the 2400 kcal/AE minimum
requirements - about the same share as found in Kaolack.
Cereal consumption is more diversified in Tamba than in Kaolack.
Imported rice remains the principal source of calories (45 percent of
all calories analyzed and 70 percent of cereal calories).
Maize and
sorghum are the next cereals of importance, with millet in last place.
This preference for sorghum and maize is linked to proximity to the
zones of production.
The importance of peanuts is surprising (14
percent of total calories) in this zone, which lies outside the Peanut
Basin.
The majority of calories consumed by households in Tamba are
purchased (more than 90 percent of total calories). In contrast to the
urban zone of Kaolack, consumption of home-produced grains does exist,
even if the share remains low.
In October-December the portion of home-
produced cereals and pulses consumed by the households climbed to nearly
17 percent of total cereal/pulse calories.
It is during the harvest (October-December) when the highest
consumption of cereals and pulses was recorded in this zone and from
April through September that lowest levels were recorded.
Purchases
were highest from July through September and lowest during October,
November,
and December.
Maize is most important from October to
December and sorghum in January, March, and June.
.
.
.
.,
:
.
:
‘.
153
SUMMARY OF FACTORS INFLUENCING THE LEVEL, COMPOSITION, AND SOURCE OF
FOOD CONSUMED BY SAMPLE HOUSEHOLDS
Appendix 8.1 contains detailed analyses of the level, product
composition,
and source of the food basket for each of six rural and two
urban study zones, looking at how such factors as income, seasonality,
commercial infrastructure, and harvest quality influence consumption
patterns within each zone.
The zone-level findings were summarized
briefly in the preceding pages.
In this section, the analyses are used
to make comparisons across zones and to draw generalizations about how
these various factors might influence
aggregate consumption for all
zones covered by the ISRA/IFPRI study.
The discussion is divided into four major sections based on
effects:
(1)
agroclimatic
and
study
zones,
(2)
commercial
infrastructure, (3) income, and (4) harvest quality and seasonality. In
addition to drawing broad generalizations across zones, the discussion
is used to develop hypotheses about the types of variables that should
be included in a demand model and the anticipated size and signs of
model coefficients.
Differences in Consumption Due to Asroclimatic and Study Zone Effects
level of Ca'loric Intake
A number of study zones fall in the "at risk" category, with fewer
than 2400 calories per AE per day.
This becomes apparent in looking at
Figure 8.1 and Table 8.1, which show levels of caloric intake across
years and zones.
Surprisingly, the rural zones considered to have the
greatest agricultural potential due to higher rainfall, lower population
density,
and better soils are the ones that have consistently lower
levels of caloric intake.
Average consumption per AE in the
southeastern Peanut Basin and Senegal Oriental is 100 to 300 calories
below minimum needs both years of the survey. Given the extremely low
level of expenditures on other foods in these zones, it is unlikely that
the average will change significantly when foods (sugar, meat, fish, and
vegetables) not yet analyzed are added.
The Niakhar zone in the central Peanut Basin experienced relatively
low consumption following the poor harvest of 1988; average levels were
acceptable, however, in HY89.
Overall consumption is low in both the urban zones but more so in
Kaolack (2099 calories/AE) than in Tamba (2328 calories/AE).
The
caloric value of products in Tamba not yet analyzed should exceed the 75
calories necessary to bring average consumption levels up to the 2400
minimum.
Even though urban diets have a much larger "other food"
component than rural zones (suggesting that products not yet analyzed
account for an important share of total calories) and snacking is
154
figure 8. --
Consumption by product group, zone and year. SA =
Sagatta; NI, Niakhar; CO, Colobane; PA, Passy; DI,
Dioly; MI, Missirah; KA, Kaolack, and TA, Tambacounda
SA NI CO PA DI MI
SA NI CO PA DI MI
KA TA
1988/89
1989/90
1990/91
0 CEREALS m LEGUMES
OTHER
Note: The correspondence between the names of zones used in Figures 8.1 - 8.12 and the names of
zones used in the text is as follows:
SA = Sagatta = northern Peanut Basin
NI = Niakhar = central basin (Niakhar)
CO = Colobane = central basin (Colobane)
PA = Passy = southwestern Peanut Basin
DI = Dioly = southeastern Peanut Basin
MI = Missirah = central Senegal Oriental
KA = Kaolack = urban zone
TA = Tambacounda = urban zone
Corrections to data files for Dioly (southeastern Peanut Basin) were made after these
graphs were prepared. The caloric intake appears to be about 2300 Kcal/AE for this zone rather
than the 2000 Kcal/AE shown on the graphs.
See discussion in text concerning these changes.
‘,
:
:
.
155
Table 8.1 --
Summary of Finding on Caloric Intake Levels and Shares.
Z 0 N E S
NPB
CPB.N
CPB.C
SYPB
SEPB
cso
LEVELS OF CALORIE INTAKE:
TOTAL SAMPLE
CEREAL/PULSE KCAL/AE
1820
2008
2199
2350
1955
2033
SHARE CEREALS PURCHASED (X)
9 0
2 1
4 8
3 9
2 4
17
SHARE OF RICE IN CEREALS (%)
7 0
15
33
28
9
9
"OTHER" KCAL/AE
488
154
425
131
6 9
a0
MARKET VILLAGE
CEREAL/PULSE KCAL/AE
1545
1922
1818
1761
2062
1856
SHARE CEREALS PURCHASED (X)
100
38
aa
51
4 4
37
SHARE OF RICE IN CEREALS (X)
8 0
2 6
57
3 4
19
2 2
"OTHER" KCAL/AE
587
230
634
251
143
4 6
NON-MARKET VILLAGE
CEREAL/PULSE KCAL/AE
1943
2046
2300
2410
1954
2083
SHARE CEREALS PURCHASED (%)
8 4
18
39
3 6
2 2
13
SHARE OF RICE IN CEREALS (%)
4 7
13
28
2 6
a
5
"OTHER" KCAL/AE
N/A
121
341
110
60
9 6
POOREST QUARTILE
CEREAL/PULSE KCALIAE
1925
1911
2126
2411
1586
1964
SHARE CEREALS PURCHASED (X)
66
20
55
3 3
2 7
2 1
SHARE OF RICE IN CEREAS (%)
50
10
3 5
22
9
1 2
RICHEST QUARTILE
CEREAL/PULSE KCAL/AE
2124
2065
2162
2326
2237
1979
SHARE CEREALS PURCHASED (%)
9 4
2 5
57
5 0
2 8
11
57
22
43
3 6
1 9
8
i:-s~~~~~F-::::~~,“:E
::I N: CEREACL~
~‘4 :::::
.
.
,.., :: .,, :::.:::::.j.j::.::j::::.j:.:~:
. . . . .:::I’
7:::::.., .,,: .:,,. .,.,. ....::.,x..:~
.c.;::+. .: . . . :.: :I..,. ..:.,,::......: .: :;::::,.,
,, ,.:::.;,
:,‘PA’RT
...3- .:: j:i,:::.;,
.
. . . . .._. .
LEVELS OF CALORIE INTAKE:
TOTAL SAMPLE
CEREAL/PULSE KCAL/AE
2274
2224
2320
1981
1952
SHARE CEREALS PURCHASE0 (X)
17
4 7
27
2 2
1 9
SHARE OF RICE IN CEREALS (X)
14
31
2 5
6
5
"OTHER" KCALIAE
141
238
119
2 5
9 5
HARKET VILLAGE
CEREAL/PULSE KCAL/AE
2035
1868
1829
2260
1925
SHARE CEREALS PURCHASE0 (%)
3 3
7 6
4 1
3 0
2 9
SHARE OF RICE IN CEREALS (X)
28
52
3 4
16
2 0
"OTHER" KCAL/AE
211
405
179
7 8
3 2
NON-MARKET VILLAGE
CEREAL/PULSE KCAL/AE
2324
2310
2409
1952
1989
SHARE CEREALS PURCHASED (%)
15
4 2
2 5
1 9
1 5
SHARE OF RICE IN CEREALS (X)
11
27
23
5
2
"OTHER" KCAL/AE
1 2 1
167
07
20
120
POOREST QUARTILE
CEREAL/PULSE KCAL/AE
2135
2176
2305
1597
2069
SHARE CEREALS PURCHASED (%)
2 8
72
27
17
12
SHARE OF RICE IN CEREALS (%)
1 6
29
21
5
4
RICHEST QUARTILE
CEREAL/PULSE KCAL/AE
2414
2337
2733
2207
’ 1911
SHARE CEREALS PURCHASED (%)
2 1
5 5
3 9
1 9
20
SHARE OF RICE IN CEREALS (%)
1 7
3 6
2 8
5
6
156
Table 8.1 -- continued.
LEVELS OF CALORIC INTAKE:
TOTAL SAMPLE
CEREAL/PtlLSE/KCAL/AE
1530
1823
SHARE CEREALS PURCHASED (X)
9 5
82
SHARE OF RICE IN CEREALS (%)
62
7 0
"OTHER" KCAL/AE
569
505
source :
IFPRI/ISRA Survey.
common and extremely difficult to capture in urban survey data, it is
likely that products still needing analysis would approach but not
surpass the 300 calories needed to achieve an acceptable average level
for Kaolack.
Looking at average consumption levels does not address the question
of what share of households within each zone fall below or above the
2400 calorie minimum. Data presented in Table 8.1 show that in HY88, 50
to 70 percent of sample households, depending on the zone, had average
levels of caloric intake below 2400 kcal/AE from cereals, pulses, milk,
bread, vegetable oil, and tubers.
In HY89, 30 to 40 percent were below
the "at risk" level in the central and southwestern Peanut Basin while
60 percent were in the "at risk" category in the southeastern basin and
in Senegal Oriental, the urban zones had about 50 percent of households
with inadequate consumption.
In summary, two of the zones thought to have high agricultural
potential (Senegal Oriental and the southeastern basin) tend to have
lower average consumption levels and/or a higher share of households in
the "at risk" category than other zones.
The analysis in Chapter 5 revealed that in 1988/89, the cereal
sufficiency ratios for these zones were higher than for the lower
potential zones, suggesting that production is not the problem in the
first year,
In 1989/90,
the cereal sufficiency ratio for Senegal
Oriental was lower than all other zones but the northern Peanut Basin
while the ratio was highest in the southeast.
Households in the southeastern Peanut Basin had the highest average
income per AE in both years (Chapter 4), suggesting that households
should have been able to ensure food security through purchases.
The
causes of this apparent contradiction are being investigated.
Two
possible explanations are an income distribution that is highly skewed
(this might give a misleadingly high average income for the zone) or
poorly distributed over time (causing periods of low consumption).
.
I
:
.
1
.
<’
157
Another possible explanation could be poor supply due to the remote
nature of many villages in these zones.
The contribution to total caloric intake made by each of the major
product groups is summarized in Figure 8.2 and Table 8.2.
In Senegal
one would clearly expect cereals to contribute the largest share of
calories, followed by pulses, and then other foods..
Figure 8.2 and
Table 8.2 show that the assumption about cereals dominating is correct
across all rural and urban zones, but that the picture for pulses and
'other food' is mixed.
Cereals accounted for a greater share of total calories analyzed
(more than 80 percent) in the southwest and southeast Peanut Basin in
both survey years.
Cereals were also more than 80 percent of the diet
in the central Peanut Basin (Niakhar) following the poor peanut harvest
of 1988/89, which reduced households' ability to supplement cereals with
pulses and other foods.
The urban zones and the northern Peanut Basin
consume the smallest shares of cereals (63-65 percent of total
calories).
Cereals provide from 1364 to 2058 calories per AE, depending
on the zone and year -- there are no zones where cereals alone cover the
minimum needs of 2400 calories.
Pulses are usually the second most important source of calories.
"Other foods" are more important than pulses only in the urban zones and
in two of the rural zones (northern Peanut Basin and central Peanut
Basin - Colobane) following the poor 1988/89 harvest. There is no zone
where total cereal/pulse calories alone attain the minimum 2400 calorie
requirement.
The southwestern Peanut Basin zone of Passy is the only
one that comes close with slightly over 2300 calories per AE coming from
cereals and pulses in both years.
P r o d u c t C o m p o s i t i o n
In this section the product composition of each of the major groups
is discussed.
Cereals.
Millet dominates the cereals group in the four study
zones representing the central and southern Peanut Basin, providing
between 1000 and 1500 calories per AE per day. Imported rice provides
the remaining cereal calories in the central and southwestern Peanut
Basin while sorghum, rice, and maize (in order of importance) complete
the cereal diet in the southeast Peanut Basin.
Maize provides slightly more than half the cereal calories in
Senegal Oriental, with sorghum and millet contributing another 45
percent and imported rice about 5 percent.
The urban zones and the northern Peanut Basin rely most heavily on
imported rice, which accounts for 62 to 70 percent of cereal calories.
Millet and very small amounts of maize complete the cereal picture for
the northern Peanut Basin and Kaolack, while Tamba residents complete
their cereals with maize (12 percent), sorghum (12 percent), and millet
(6 percent).
.
.,;
:
.
158
Figure 8.2--
Cereal and pulse consumption by product, zone, and
year.
SA = Sagatta; NI, Niakhar; CO, Colobane; PA
Passy; DI, Dioly; MI, Missirah; KA, Kaolack, and Ti,
Tambacounda.
> 2000 -
OQ\\ 1500-
2\\
; lOOO-
0
-I
;:
5 0 0 -
SA Nl CO PA DI Ml
‘SA NI ‘CO PA DI MI
KA TA
19as/a9
1989/90
1990/w
0 MILLET
@@$j SORGHUM m CORN
@j&j IMP RICE
NIEEE
m PEANUTS
..:i
:
.
(
:
159
Table 8.2 --
Share of minimum calorie requirements covered by type of product,
zone, and year
Peanut Basin Rural Zones
Urban Zones
Central
Central
Senegal
Share
North
(Niakhar) (Colobane) Southwest
Southeast
Oriental
Kaolack
Tambacounda
Share of 2400 cal/AE
covered by cereals
& pulses(%)
HY88/89'
76
84
92
98
91
85
n.a.
n.a.
HY89/90
n.a.
95
93
97
95
81
n.a.
n.a.
7-90/6-91
64
65
Share of 2400 cal/AE
covered by all foods
analyzed(%)
HY88/89'
96
90
109
103
93
88
n.a.
n.a.
HY&?9/90
n.a.
101
103
102
96
85
n.a.
n.a.
7-90/6-91
87
86
Source: ISRA/IFPRI project survey.
Notes: Minimum caloric requirement is 2400 calories per AE per day.
This is 80% of ORANA recommended
standards of 3000 kcallday for a moderately active adult male.
Those below this level are considered
"at risk." Foods analyzed are cereals, pulses, milk, tubers, bread, and vegetable oil. n.a. means not
available.
' Data for harvest year 1988/89 cover December 1988 through September 1989.
The above points are all well illustrated in Table 8.3 and Figure
8.2.
This review suggests three distinct cereal consumption patterns:
(1)
Millet-based consumption complemented with imported rice
(central and southern Peanut Basin)
(2) Rice based consumption complemented with coarse grains (urban
zones and northern Peanut Basin)
(3) Mixed coarse grain consumption (Senegal Oriental)
Although differences in consumption patterns begin to appear among
the various zones, it is not yet clear to what extent these differences
are due to agroclimatic and socioeconomic factors that distinguish the
zones, or to price, income, and infrastructure variables, which may
differ across zones.
Pulses.
Peanuts provide 80-99 percent of all pulse calories and
cowpeas the remaining l-20 percent in all zones but the northern Peanut
Basin where
’
:
.
.’
160
Table 8,3-- Product composition of cereals consumed by zone and year
Peanut Basin Rural Zones
Urban Zones
Survey
Central
Central
Senegal
Product Period
North (Hi a k h a r ) (Co1 obane) Southwest.
Southeast
Oriental
Kaol ack
Tambacounda
Rice
wia8/89’
70
15
33
28
16
9
HY89/90
-
14
31
25
10
5
7-90/6-91
-
62
70
Millet and
sorghum
tiYaa/a9’ 2
5
a4
66
72
a3
38
HY89/90
-
86
68
75
88
48
7-90/6-91
-
36
18
Maize
nY8aia9’
5
Cl
cl
<l
1
53
HY89/90
-
Cl
cl
<l
2
47
7-90/6-91
-
1
12
Source: ISRA/IFPRI project survey.
Note: Numbers are percent of total cereal calories per adult equivalent provided by the cereal products
listed.
' Data for harvest year 1988/1989 cover December 1988 through September 1989.
the pulse diet is split equally between cowpeas and peanuts. Table 8.4
provides the exact shares of each product by zone and year.
Peanuts provide additional calories in the form of oil in zones
where home processing of peanuts is common (parts of the central Peanut
Basin).
These calories are accounted for in the "other food" section
because data are not precise enough to allow peanut oil to be separated
from other vegetable 0i1s.~
Cowpeas are a crop that received substantial donor and government
support in the mid-f980s.
Cowpeas are traditionally consumed in Senegal
as a condiment, being added in small quantities to the sauce.
The
ISRA/IFPRI data suggest that in the north and central parts of the
Peanut Basin, where the support programs were introduced, cowpeas
account for a much larger share of pulse consumption (20-50 percent)
than in the southwest Peanut Basin and Senegal Oriental (1-4 percent of
' In recent years vegetable oil processed industrially and sold in
Senegal has been a mix of various vegetable oils and peanut oil,.
Consumers tend to be unaware of what type of oil they have purchased,
therefore we are unable to distinguish among the different types.
161
pulse consumption).
This raises the question of whether consumers in
these low-consumption zones could be encouraged to increase the share of
cowpeas in their diets if the surplus production from the producing
zones was made available.
Table 8.4-- Product composition of pulses consumed by zone and year
Peanut Basin Rural Zones
Urban Zones
Central
Central
Senegal
Product
North (Niakhar) (Colobane] Southwest
Southeast
Oriental
Kaolack
Tambacounda
Peanuts
tma/a9’
50
al
83
98
95
93
HY89/90
78
79
99
07
96
7-90/6-91
a7
99
Cowpeas
HY88/89'
50
19
17
2
5
7
HY89/90
22
21
1
13
4
7-90/6-91
13
1
Source: ISRA/IFPRI project survey.
Note:
Numbers are percent of total pulse calories per adult equivalent provided by peanuts and cowpeas,
respectively.
' Data for harvest year 1988/89 cover December 1988 through September 1989 only.
"Other Foods."
"Other foods" examined to date are vegetable oil,
bread, milk, and tubers.
Figure 8.3 clearly shows that vegetable oil
makes the biggest contribution to the "other food" category in all zones
but Senegal Oriental, where milk predominates.
Bread is the second
biggest contributor in the urban zones, the Northern Peanut Basin, and
the Colobane zone of the Central Peanut Basin.
In all other zones,
bread is inconsequential and milk is the second most important product.
If a larger share of "other foods" in the diet is associated with
a more urban type of consumption, one concludes from Figure 8.3 that the
"rural" zone of the northern Peanut Basin and the Colobane zone of the
central Peanut Basin are clearly moving in the direction of an urban
consumption basket.
The oil consumption in these two zones provides
more calories per AE than the sum of all
"other food" calories for each
of the remaining rural zones.
Tubers account for a very small share of "other food" calories in
all zones.
This is generally true for the rural areas.
Tuber
consumption in urban areas is probably somewhat underestimated here due
to difficulties in obtaining accurate unit measures for tubers consumed.
162
Figure 8.3--
Composition of "other food" by product, zone, and year.
SA = Saaatta; NI, Niakhar; CO, Colobane; PA, Passy; DI,
Dioly; f?I, Missirah;
KA, Kaolack, and TA, Tambacounda.
6007
1
500 -
2'400-
\\
T
\\300-
07
!A
g 200 -
i
loo-
0
SA NI CO PA DI Ml
KA TA
1988/89
1989/90
1990/91
m VEG OIL
m WH BREAD m MILK
0 TUBERS
,..
.
.i
:
163
T y p i c a l P a t t e r n s
Modifying the three major types of cereal consumption patterns
identified above by adding considerations concerning the caloric
importance of "other foods" and relative shares of cowpeas in the pulse
diet produces the six groups shown in Table 8.5.
Cowpeas are a very small and generally insignificant part of the
diet.
Using the level of cowpea consumption to classify consumption
patterns would not normally be justified.
Given the interest in
promoting cowpea production in the northern regions of Senegal and
problems encountered in marketing surplus production, we have paid more
attention to the crop than would normally be the case.
Eliminating
cowpeas from the classification system leaves four consumption types:
(1) millet base/low "other," (2) millet base/high "other," (3) rice
base/high "other," and (4) mixed coarse grains/low "other."
It becomes apparent from this system of classification that there
is not a sharp distinction between rural and urban patterns.
Rice,
typically associated with urban consumption, predominates in one rural
zone and accounts for 25-33 percent of cereal calories in two other
rural zones (central Peanut Basin: Colobane and southwestern Peanut
Basin).
"Other foods," also associated with urban consumption, are
important sources of calories in the rural zone of the northern Peanut
Basin and the Colobane zone of the central basin.
Cowpeas, sorghum,
and maize consumption do,
however, follow
expected patterns of being much more heavily consumed in their zones of
production than elsewhere.
Given the relatively low production
sufficiency ratios reported in Chapter 5 -- even in the high potential
zones -- it is not surprising to find cereals being consumed in their
zones of production.
There was no zone where sample households produced
enough cereal to provide an average consumption of 2400 kcal/AE during
the two study years.
The three best cases were Colobane in the central
basin and the two zones in the southern basin where approximately 80
percent of the 2400 kcal/AE could have been covered by home produced
cereals in 1989/90.
Because these zones obtain about 75 percent of
their calories from cereals, there is a margin of only 5-10 percent of
production which could have been "exported" from these zones to the less
productive ones.
The only serious maize-producing area could cover
about 40 percent of their 2400 kcal/AE needs in HY88 and 60 percent in
HY89 with home produced cereals.
S o u r c e s o f P r o d u c t s C o n s u m e d
Table 8.6 and Figure 8.4 show that cereal/pulse purchases account
for the smallest amount of calories in Senegal Oriental (about 350 per
AE per day or 17 percent of cereal/pulse consumption).
In the
southeastern Peanut Basin and the Niakhar zone of the central Peanut
Basin, purchases are also low (both approximately 450 calories per AE
per day -- about 20 percent of cereal/pulse consumption).
.
.
.
.
164
Table 8.5-- Classification of consumption patterns
C o n s u m p t i o n
P a t t e r n
Cereal
“Other”
Pulse
Zone
Millet based with
Low share
Low to medium
Southeast Peanut Basin
rice supplement
share of cowpeas
Southwest Peanut Basin
Millet based with
Low share
High share of
Central Peanut Basin: Niakhar
rice supplement
cowpeas
Millet based with
High share
High share of
Central Peanut Basin: Colobane
rice supplement
cowpeas
Rice base with
High share
High share of
Northern Peanut Basin:
mixed coarse
cowoeas
Kaolack
grain supplement'
Rice base with
High share
Low share of
Tambacounda
mixed coarse
cowpeas
grain supplement
Mixed coarse
Low share
Low share of
Senegal Oriental
grain base
cowpeas
Source: ISRA/IFPRI project survey.
' Coarse grain supplement is predominantly millet in these zones with a small amount of maize.
Table 8.6 --Share of cereals and pulses purchased by zone and year
Peanut Basin Rural Zones
Urban Zones
S u r v e y
Central
Central
Senegal
Product Period
North (Ni a k h a r ) (Co1 o b a n e ) S o u t h w e s t
Southeast
Oriental
Kaol ack Tambacounda
Cereals
HY88/89' 9 0
21
48
2 4
1 7
HY89/90
-
17
4 7
2 2
1 9
7-90/6-91
-
95
9 1
Pulses
HY88/89' 4 9
1 8
1 8
1 6
1 0
1 8
HY89/90
-
4
15
10
19
1 0
7-90/6-91
-
95
82
Source: ISRA/IFPRI project survey.
Note: Numbers are percent of total calories purchased in each product group.
In rural zones, most
unpurchased cereal and pulses are home produced.
In Kaolack 1 percent of cereals are home produced
- all other nonpurchased products come from gifts.
In Tamba 1 percent of cereals and 13 percent of
pulses are home produced, the rest come from gifts.
I
' Data for harvest year 1988/89 cover December 1988 through September 1989 only.
165
Figure 8.4--
Sources of cereals and pulses by zone and year. SA =
Sagatta; NI, Niakhar; CO, Colobane; PA, Passy; DI,
Dioly; MI, Missirah; KA, Kaolack, and TA, Tambacounda.
3000
E
SA NI CO PA DI Ml
SA NI CO PA Dl Iv41
KA TA
1988/89
1989/90
1990/91
0 PURCHASES m PRODUCTION
GIFTS
166
Households in urban zones practice very little agriculture; hence,
more than 90 percent of their cereals and pulses are purchased.
Rural
zones that are described above as having more urbanized consumption
habits (northern Peanut Basin and Colobane zone of the central basin)
exhibit the largest share of cereal/pulse purchases among the rural
zones (80 and 46 percent, respectively).
The southwestern Peanut Basin
falls between the two extremes
with about 30 percent of cereal/pulse
calories being purchased.
Given that average consumption in most zones just meets minimum
requirements or falls somewhat short, this relatively high dependence on
purchased cereals and pulses implies that rural households must have
substantial cash incomes -- from sales of crop produce or other
activities -- to ensure minimum acceptable levels of consumption.
This is true following the poor 1988/89 harvest as well as the good
1989/90 harvest.
Rural zones where purchases provide the fewest
calories are also the zones that have calorie deficits.
A low share of
purchases appears to be correlated with a low level of consumption
rather than with adequate levels of production that make it unnecessary
to resort to the market.
There is also evidence that when households enter the market to
purchase cereals, there is a greater tendency to purchase rice than
coarse grains.
Assuming that households want some diversity in their
diets, it is not surprising that one would purchase cereals that differ
from those produced, hence the large share of rice in total purchases
shown in Table 8.7.
Table 8.7-- Share of rice in cereal purchases by zone and year
Peanut Basin Rural Zones
Urban Zones
S u r v e y
Central
Central
Senegal
Period
N o r t h (Niakhar) ( C o l o b a n e ) S o u t h w e s t
Southeast
Oriental
Kacl ack
Tambacounda
HY88/89'
7 4
7 3
7 0
71
35
4 1
HY89/90
7 7
66
88
2 8
1 9
7-90/6-91
6 5
7 7
Source: ISRA/IFPRI project survey.
Note: Numbers are percent of purchased cereal calories per adult equivalent provided by imported
rice.
' Data for harvest year 1988/89 cover December 1988 through September 1989 only.
Differences in Consumption Due to the Level of Commercial Infrastructure
167
One-third of the sample in each rural zone is located in a village
that is the site of a weekly market.
Many of these market villages have
other infrastructure such as permanent shops, schools, health services,
government offices (arrondissement level), peanut and cotton purchasing
agents,
and so on.
It is our hypothesis that the presence of this
infrastructure encourages a more urban food consumption pattern than
that prevailing in villages lacking such infrastructure.
The urban
consumption pattern is encouraged for three reasons: (1) supplies of
rice and "other foods" are more readily available in market villages,
(2) prices of these foods tend to be lower in market villages than in
nonmarket villages, and (3) the higher level of economic activity in the
market village makes it easier for one to earn the cash needed for
purchasing these "urban" food products.
In this section the differences in level, composition, and source
of food products consumed by households in market and nonmarket villages
is examined.
level and Composition of Caloric Intake from Cereals and Pulses
Figure 8.5 presents two sets of bar graphs showing the level and
composition of consumption by zone and year.
In general, there is a lower level of caloric intake from cereals
and pulses in the market villages.
The difference between market and
nonmarket cereal/pulse consumption ranges from about 100 calories to as
much as 500 per AE per day.
The only exception here is the southeastern
Peanut Basin where the market village exhibits about the same level of
consumption as non-market villages.
There is consistently a larger share of rice in the cereal/pulse
diet of households in market villages.
In zones where rice consumption
is generally low, market villages consume as much as 8 times more rice
per AE than nonmarket villages.
In zones where rice is more common, the
market villages consume 1.5 to 2 times more than their neighbors.
Market villages with the lowest rice consumption levels still cover 15-
20 percent of cereal/pulse calories with this imported product. At the
higher end, as much as 68 percent (northern Peanut Basin) and 50 percent
(central Peanut Basin: Colobane) of cereal/pulse calories come from
rice.
Market villages generally obtain fewer calories from pulses, but
there is no apparent difference with respect to the relative shares of
cowpeas and peanuts.
" This observation is based on partial analysis of data files that
were revised subsequent to the draft report.
Further analysis of the
revised files is required to be sure about what is happening in
individual villages.
:
_“,
I’.
.,
:
i
#
.,
I
:
*
:
.’
168
Figure 8.5--
Product composition of cereal and pulses consumption
compared for (A) nonmarket and (B) market villages.
A
NON - MARKET VILLAGES
1
C
-F ‘ROOUCT COMf’OSlTiON OF C O N S U M P T I O N
I-
/-
NOT
AVAL
IT
SA .
Y
.
m
I.4
II
Y
'*'m'fu'a'u
HY 1988/89
HY 1989/90
B
:;-j/
~300, PROOUCT COMPOSrrKxl
O F C O N S U M P T I O N
:
I
:
.
‘_
I
.
169
S o u r c e o f P r o d u c t s C o n s u m e d
Households in nonmarket villages consume N-30 percent more
calories from home production than do their market village counterparts.
As Figure 8.6 shows, this leads to a greater share as well as a greater
absolute quantity of cereal and pulse calories being purchased by market
village households.
As mentioned previously, households purchasing
more cereals and pulses appear to favor rice over coarse grain
purchases.
This is illustrated in Figure 8.7, which shows that shares
and absolute quantities of rice calories purchased are greater in market
villages.
Policy Relevance of Urbanized Consumption Patterns in Market Villages
The 1988 census made it clear that the rate of urbanization in
Senegal is quite high (3.9 percent growth rate for Dakar compared with
2.8 percent for the entire country).
Although it is unlikely that
market villages are growing as fast as Dakar, a quick comparison of 1976
and 1988 census data for rural villages suggests that market villages
are growing moge rapidly than other rural villages.
In most of the
rural study zones, market villages account for 15-20 percent of the
rural populatidn.
As it becomes increasingly difficult to make a living
with crop production and animal husbandry activities alone,
rural
populations are likely to migrate toward market villages as well as the
urban centers.
This type of migration will increase demand for rice
substantially through both demand and supply factors because market
villagers.eat more rice and produce less coarse grain cereal (see
Chapter 5) than their counterparts in nonmarket villages.
Income Effects on Consumption Patterns
In order to examine differences in consumption that might be
associated with different levels of income, we divided the sample for
each zone into quartiles and compared the cereal/pulse consumption
basket of the lowest quartile with that of the wealthiest.
In the
following paragraphs,
the differences in cereal/pulse consumption of
rich and Door households are examined in terms of levels of caloric
intake,
product composition,
and sources of products consumed for the
rural sample."
L e v e l o f C a l o r i c Intake
Most rural zones exhibited surprisingly little income effect on
total cereal/pulse calories durinq HY88. The difference between the two
groups is only about 100 calories per AE. Given the high coefficients
of variation on these data, such a small difference is not likely to be
significant.
The only exception to this tendency is the southeast
Peanut Basin where there is a difference between rich and poor of about
600 calories per AE!
This is one of the zones where a large share of
" Time did not permit us to do a comparable analysis for urban
zones.
‘(%
;
,
j
Figure 8.7--
Composition of cereals and pulses compared for (A)
nonmarket and (6) market villages.
A
-
NON - MARKi VILLAGES
D
H-f 1988/89
H Y 1989/90
RI MARKET VlLLAGES
I
I
CXMPOSITION OF
PURCHASES
.“.
;:
,,
i
*’
:
.’
.
.
”
172
households (60-70 percent) failed to cover minimum needs of 2400
kcal/AE.
Following the much better 1989/90 harvest, there is a larger
difference between rich and poor, with the rich consuming from 150
(Colobane in the central basin) to 700 (southeastern Peanut Basin) more
calories per AE than the poor in all zones but Senegal Oriental.
Figure 8.8 illustrates these points by comparing levels and
composition of cereals and pulses consumed by the lowest and highest
income quartiles in each rural zone.
The top half of the figure
illustrates consumption patterns following the poor 1988/89 harvest and
the lower half represents consumption following the good 1989/90
harvest.
Source of Ca7ories by Income Group
The fact that there was very little improvement in caloric intake
for the poor following the good rainfall of 1989/90 suggests that (1)
production constraints prevent them from reaping the benefits of better
climatic conditions and (2) lack of alternatives to crop income keep
them from supplementing cereal/pulse production with cash purchases of
food.
The rich,
on the other hand,
appear
to have produced
substantially more food in HY89 than they did in HY88. The increase in
production translated into an increase of from 300 to 700 calories per
AE per day more coming from home production in HY89.
Figure 8.9 illustrates this point by comparing the source
composition of cereal and pulse consumption between the two harvest
years.
Senegal Oriental is the only zone where the wealthy were worse
off in HY89 and the poor had a better harvest."
Figure 8.9 also shows that both the rich and the poor relied
heavily on purchases of cereals and pulses in the northern Peanut Basin
(56 percent for the poor and 90 percent for the rich in HY88) and in the
Colobane zone of the Central Peanut Basin (53 percent in HY88 and 45
percent in HY89 for the rich;
50 and 62 percent, respectively, for the
poor).
In other words, both the rich and the poor in these zones share
a tendency to purchase a large part of their food.
Even in the zones relying the least on purchases (Niakhar in the
central Peanut Basin,
the southeastern Peanut Basin and Senegal
Oriental),
a minimum of 15-20 percent of cereal/pulse calories is
purchased.
This leads us to the question of whether these zones are all
deficient in cereals or if they are selling coarse grains and purchasing
rice to introduce variety into the diet.
The discussion of crop
production in Chapter 5 confirms that the average household
" This may be due to households who received large amounts of
migration remittance toward the end of the year being classified as
wealthy when they were really poor for most of the year.
.I
:
:
:
:
.
.
.
‘.
.
i
Figure 8.8--
Income effects on level and composition of cereal and
pulse consumption for two harvest years. NPB, Northern
173
Peanut Basin; CPBN, Central Peanut Basin - Niakhar-
CpBC, Central Peanut Basin, Colobane; SWPB, South&t
peanut Basin; SO, Senegal, Oriental.
A
JO00
2750
2500
2 2250
E 2000
a
w 1750
Q:
E 1500
2 1250
?I!
ig 1000
5
750
500
250
0
LOM
Lo M
Lo H
NPE
CPB.C &P;
so
mw=wJn~m mpwcm LYEiflm
B
2750
NPB
CPB.N
CPB.C SWPB
SEPB
s o
uzw~wo~RRxT
m-s
EfEamwPw
Figure 8.9--
Income effects on sources of cereals and pulses
consumed for two harvest years.
NPB, Northern Peanut
Basin; CPBN, Central Peanut Basin - Niakhar; CPBC,
174
Central Peanut Basin, Colobane; SWPB, Southwest Peanut
Basin; SO, Senegal, Oriental.
A
btu
IAH
Lola
LDHI
Lota
Lo H
NPB
CPB.N
CPB.C SWPB
SEPB
so
E3l-m-M~
1
B
Lo HI
Lo H
Lo HI
Lo Hl
Lo HI
Lo H
NPB
CP8.N
CPB.C SWPB
SEPB
SO
Ex-rsa- nlx-rs
3
. .
!..
,’
175
across all zones fails to produce 2400 kcal/AE of cereals.
Even when
one assumes that the caloric demand for coarse grains in rural Senegal
is only 1900 kcal/AE, home production only covers these needs in the two
zones of the southern Peanut Basin and only after the better than
average 1989/90 rains.
P r o d u c t C o m p o s i t i o n o f C o n s u m p t i o n
In HY88 the wealthier households consumed more rice than the poor
(Figure 8.8).
This is true for absolute quantities per AE as well as
share of rice in the diet.
Senegal Oriental was the only zone where the
poor ate more rice than the rich (see footnote 15).
Following the good 1989/90 harvest, the absolute amount of rice
consumed went down for both income groups in all zones with the
exception of the poor in the Niakhar zone of the central Peanut Basin
where rice consumption increased by 69 calories per AE.13
In HY89 wealthy households in all rural zones consumed more rice
per AE than the poor; the share of rice in the diet of the rich was,
however, substantially greater in only two zones (central Peanut Basin -
Colobane,
and southwestern Peanut Basin).
A large increase in
consumption of home grown millet was largely responsible for the reduced
share of rice eaten by wealthier households in HY89.
In HY88 the poor obtained 8 to 39 percent of their cereal/pulse
calories from rice, depending on the zone.
In HY89 the range was 4 to
24 percent.
Zones with the higher shares in rice are the most densely
populated; this means that a large part of the rural poor depend on rice
for about 20 percent of their cereal/pulse calories.
The fact that the rich eat more rice than the poor does not
necessarily mean that they eat less millet.
In HY88 the poor consumed
mire calories per AE from millet than the rich in every zone.
In HY89
this was true only in Senegal Oriental. In other zones, the poor ate
an equal or smaller absolute quantity per AE than the rich. Millet did,
however, account for a slightly larger share of the cereal/pulse diet in
poor households than in rich ones due to the lower overall level of
consumption by the poor in HY89.
In terms of pulse consumption, the biggest distinction between rich
and poor is found in the northern Peanut Basin where the poor consumed
substantially more cowpeas late in HY88 than the rich.
This came about
because the 1989 cowpea harvest was early (August) and abundant,
permitting households with inadequate food supplies to supplement millet
l3 Analysis of agricultural product transactions data in Chapter
6 suggests that increased rice consumption following a good harvest may
be more prevalent than a partial analysis of the lowest and highest
income quartiles suggests.
The effect of increased income due to a good
peanut harvest appears to encourage households to supplement their
millet consumption with purchased rice.
;
.’
:
176
and rice dishes with very inexpensive cowpeas.
On an annual basis,
cowpeas provided 14 percent of the poor's cereal pulse calories but only
5 percent of those for the rich.
There is no clear pattern linking income class and absolute
quantity or share of total calories obtained from peanuts -- sometimes
the poor eat more and sometimes the rich, depending on the year and the
zone.
Composition of Purchases
Figure 8.10 illustrates the product composition of purchases. Rice
is generally a more important component (in absolute quantity and share)
of purchases made by the rich than those made by the poor. An exception
to this rule is found in the northern Peanut Basin in HY88 where poor
households purchased less rice per AE but the share of their purchases
devoted to rice was quite high (70 percent vs. 57 percent for the rich).
Millet is generally a more important component (in absolute
quantity and share) in the purchases of the poor. Exceptions to this
were the northern Peanut Basin, where the rich purchased three times
more millet than the poor in HY88, and the southwestern Peanut Basin,
where the share of millet in purchases made by the rich was 6 percentage
points higher than the share for the poor.
There are no striking differences in purchasing patterns for pulses
between the rich and the poor that appear in both harvest years.
Absolute quantities and shares of purchased calories coming from pulses
differ little by income group.
The only exceptions occurred in HY89:
cowpeas accounted for 10 percent of purchased calories for the rich and
only 5 percent for the poor in the southeastern Peanut Basin; and
peanuts accounted for 23 percent of purchases by the poor, but only 6
percent of purchases by the wealthy.
Seasonal Chanqes in Consumption Patterns14
It is difficult to compare seasonality across zones in the
ISRA/IFPRI sample because the onset of the different seasons does not
fall in the same calendar month in all zones.
To simplify the
presentation,
four basic seasons are used in all zones:
l4 This section has not been revised to incorporate the higher
average levels of consumption calculated for the southeastern Peanut
Basin since the draft report was issued. It is unlikely that the higher
levels will substantially change the pattern of seasonality.
.
.
.’
‘.
:
Figure 8.10--
Income effects on composition of cereals and pulses
177
purchased for two harvest years. NPB, Northern Peanut
Basin; CPBN, Central Peanut Basin - Niakhar; CPBC,
Central Peanut Basin, Colobane; SWPB, Southwest Peanut
Basin; SO, Senegal, Oriental.
1750
PRODUCT COKPOSll7ON
OF PMXXASES
2 1500
E
a 1250
2
E 1000
E
c
750
f5
ij
500
250
NPB
CPB.N
CPB.C SWPB
SEPB
so
palucr tEEi= Emu
o=
IxxImm-
2000
1750
HARVEST YEAR 1989/90
>
PROCKT CoMposmoH OF PLIXtiASEs
Q 1500
n
(II
2 1250
W
Q
[r
1 0 0 0
t
z
750
z
d
500
0
0
Lo HI
Lo H
La Hl
Lo
M'
'Lo'HI'
'Lo'H
'
NPE
CPB.N
CPB.C SWPB
SEPB
so
mu msa=w ez4~
Elm
Ixxl-EEEI-
178
SEASON
CALENDAR MONTHS
Harvest period
October - December
Peak period for non-
agricultural
activities
January - March
Period of
clearing fields
and planting
April - June
Period of
weeding
July - September
The major problem encountered with this classification is that the
early rains in 1989 led to early harvests, particularly of cowpeas and
millet in the northern and central Peanut Basin.
The early harvest had
the unusual effect of increasing consumption of home produced crops
during the weeding season -- a period that one would expect to exhibit
serious calorie deficiencies following a harvest as poor as that for
1988/89.
Another shortcoming of this seasonal classification is that data in
the weeding season are an average over three months that can exhibit
very different calorie levels and composition.
An earlier report
presented data for August, September, and October for several zones.
Readers interested is the details of month-to-month changes in level of
caloric intake should consult this document (Kelly et al., February
1991).
The following paragraphs discuss seasonal changes in level,
composition,
and source of cereal and pulse consumption.
Leve7 of Ca7oric Intake
The level of caloric intake is examined from two perspectives: (1)
changes in absolute level and (2) percent fluctuation about the average
annual level.
Part A of Figure 8.11 shows average calories consumed per AE per
day during the harvest season for each zone and year covered by the
rural survey.
Each bar in the graph shows the average level per AE of
cereal and pulse calories consumed for each zone/year combination.
Parts B-C of the same figure illustrate consumption patterns for the
other three seasons.
Given that only cereal and pulse calories are
examined in this section, 2200 calories is used as a rough indicator of
minimum acceptable consumption levels (with this level of cereal/pulse
consumption,
most will attain the 2400 mark with "other foods"; 1900
represents the level below which we believe populations are clearly "at
risk" because it is unlikely that they would obtain 500 calories per AE
of "other food."
:
.
,.;
.
.
:
.
c
INTER-ZONE COUPARISON
HARVCST
PEnlOO CONSUMPTION
.
a
HY 19aa/a9
H Y 1989/90
i
J
.
.
B
r-
.
.
,cm
,I%0
INTER-ZONE COUPARISON
,%a I-----
MN - t<G PERIOD CONSUUPTDN
-
.
_
: .
180
At harvest time, 6 of the 11 bars are equal to or greater than 2200
calories;
this declines to 5 bars in the nonagricultural season and 3
bars in the planting and weeding season. Hence, at the upper end of the
scale, almost half the zone/year averages are below the 2200 level even
at harvest.
The number of zones falling below this cutoff point
increase throughout the year,
leaving only three cases above the 2200
level during both the planting and the weeding season.
Of the six cases maintaining this relatively high level of cereal
pulse consumption,
the southwest Peanut Basin appears three times
(planting seasons 1988 and 1989, and weeding season 1989). Niakhar in
the central Peanut Basin -- appears two times (planting and weeding
1989) and Colobane, also in the central basin, once (weeding 1988).
This result reflects the relatively strong preference for cereals vs
"other foods" in two of the zones (Niakhar in the central Peanut Basin -
and the southwestern Peanut Basin) as well as the ability of households
in these zones to manage stocks and finances to avoid periods of low
caloric intake.
Looking at the lower end of the scale, there is no zone/year
combination with average consumption below 1900 calories per AE at
harvest time,
one case during the HY88 nonagricultural season, four
cases during the planting season, and just two during the weeding
season.
Zones falling into this "at risk" group are the northern Peanut
Basin (nonagricultural, planting, and weeding HY88), Senegal Oriental
(planting HY88 and HY89 and weeding HY89), and the southeastern Peanut
Basin (planting HY88).
The fact that the northern Peanut Basin falls into this at risk
grouping is of less concern than the presence of the two other zones due
to the very high levels of "other foods" consumed in the north (about
512 calories per AE per day).
The low levels of cereal/pulse
consumption,
coupled with extremely low levels of "other food"
consumption suggest that one would find evidence of nutrition related
health problems in both the Southeastern Peanut Basin and Senegal
Oriental.
In looking at percent of variation around the mean level of
consumption for each zone/year combination, we find that seasonal
differences are usually not more than plus or minus 5 percent of the
annual average.
During the harvest period, consumption ranges from
0.98 to 1.15 of the annual average with 8 of the 11 observations falling
between 100 and 106 percent of the average.
During the nonagricultural season, the range is 0.93 to 1.07, with
6 observations equal to or below the annual average and 5 above. Moving
into the planting season, the range is 0.93 to 1.05 with only 2 cases
where consumption is equal to or greater than the annual average.
During the weeding season, the picture for HY88 differs from that
of HY89.
This difference is due to the early 1989/90 harvest, which
increased consumption during the HY88 weeding season, and the late
1990/91 harvest which was not available until late October 1990. IN
.
._
’
:
.
:
.
:
181
HY88 only two of the 6 zones fell below the annual average; in HY89 all
5 zones fell below, with the range being from 91 to 98 percent of the
average.
By way of summary,
we conclude that seasonal variation in
consumption levels exists but stays within a band of plus or minus 5
percent of the annual average in most cases. The exception to this rule
is Senegal Oriental, which has quite large fluctuations (plus 11 percent
in harvest period 1989/90 and minus 9 percent in weeding period
1989/90).
This wide range in seasonality is of concern because the
overall zone average is quite low.
It suggests that households in the
more drought-prone areas of the Peanut Basin have learned to better
manage food stocks and cash reserves,
thereby
ensuring less
interseasonal fluctuation in consumption levels.
In modeling demand for cereals and pulses one would normally expect
there to be a seasonal effect on the levels of demand (declining levels
of consumption as one moves from harvest to harvest) as well as a
harvest effect (better harvest, higher level of caloric intake across
all seasons).
The discussion above suggests that the seasonal effect
exists but is relatively small.
Earlier discussion concerning average
levels across years suggested that the harvest effect also exists; it is
also generally small (l-5 percent for all zones but Niakhar in the
central Peanut Basin, which had a 12 percent increase in average
consumption between HY88 and HY89).
We have, however,
found some tendencies that disturb these
apparently general patterns.
For example, even though the harvest
effect was relatively small in most of the zones, there was a very
significant harvest effect across
all zones for the wealthiest
households whose millet production rose significantly between HY88 and
HY89.
The general seasonal trend is also rendered more complex by an
increase in consumption levels between the planting and weeding season
in HY88 but a decline during the same period in HY89 due to interannual
differences in the timing of the rains and, subsequently, the harvest,
This discussion suggests that in modeling demand it may be
necessary
to introduce variables that will capture (1) harvest
quantity/quality, (2) harvest timing, (3) seasonal effects, and (4) some
type of interaction effect between income and harvest quantity/quality.
Composition of Consumption
The most striking aspect of seasonal changes in composition is the
role played by cowpeas during the weeding season of HY88. Without the
increase in calories obtained from cowpeas during this season, the
average cereal/pulse consumption in the north would have been about 1525
calories per AE.
Although less important in the central basin zones of
Niakhar (4 percent of total cereal/pulse calories) and Colobane (6
percent) than in the north (19 percent), it clearly helped households in
these zones get though a difficult period. These findings suggest that
expansion of cowpea production in the northern and central Peanut Basin
.,;
*
,:
.
.,;
.
.
:
;-
..
182
can be justified on nutritional grounds in addition to the more common
argument that cowpeas offer farmers in the zone a cash crop alternative
to peanuts.
Seasonal and interannual differences in rice consumption are also
noteworthy.
Part A of Figure 8.11 shows that rice consumption does not
disappear in the central and southwestern Peanut Basin during the
harvest season when new coarse grains are readily available. We would
expect relatively high rice consumption during the 1988 harvest period
in the northern and central Peanut Basin because their cereal harvests
were generally poor.
Rice remained important during the harvest season
of HY89 in all these zones.
In the Niakhar zone of the central Basin
and in the southwestern Basin, the HY89 level is about the same as that
in HY88;
rice consumption drops to 67 percent of the HY88 level in the
Colobane zone of the central Peanut Basin but remains high at 577
calories per AE per day.
This relatively high level of rice consumption during a period when
coarse grains are readily available and inexpensive is probably due to
the effect of large amounts of cash income from peanut and cotton sales
coming in during this same period.
The presence of the income effect
would have to be verified for each zone by looking at the exact timing
of the cash crop sales, but the official cash crop marketing period
usually opens in December and continues until March.
Tastes must also play a part because rice consumption virtually
disappears in the southeast Peanut Basin and Senegal Oriental during the
1989 harvest, suggesting that when coarse grains are available in these
zones, households show little interest in rice.
Tracking the ups and downs of rice consumption across the four
seasons, we find three patterns.
(1) Small, incremental increases of the share of rice in the diet from
harvest to harvest (Niakhar zone in the central Peanut Basin and
Senegal Oriental).
(2)
Large and relatively constant shares of rice throughout the entire
year (northern Peanut Basin, Colobane zone of central Basin,
southwestern Peanut Basin).
(3) A relatively small share of rice in total cereal consumption rising
from the harvest through the nonagricultural season and then
falling during the planting and weeding season (southeastern Peanut
Basin).
The first pattern is probably a reflection of three factors working
together:
(1) households in these two zones have strong taste
preferences for coarse grains,
(2) they do not have liquidity to
purchase coarse grain stocks during the harvest period when prices are
low (see Chapter 4 on income), and (3) as the seasons progress, the
price of rice relative to coarse grains is declining (rice price i's
constant but the coarse grain prices are rising).
1 8 3
The second pattern
is probably a reflection of tastes and
commercial
infrastructure
which
makes
rice
readily
available,
facilitates sales of agricultural produce, and provides opportunities
for nonagricultural income.
Although every zone has a market village,
these three zones have bigger, more important markets than the other
zones.
One of the three zones has a chronic cereal deficit. Cereal
production per AE for Colobane
in the central basin and in the
southwestern basin are among the highest in the study, yet purchased
rice remains important in these zones, suggesting that factors other
than production are driving the demand.
The quantity of rice consumed in the third case (southeastern
basin) is so small that it is hard to know if this is a real pattern.
One frequently hears farmers saying that only coarse grains provide the
energy required for the heavy field work during the planting and weeding
seasons.
This could possibly be the explanation for the higher rice
consumption during the periods of less intense agricultural work.
Seasonality o f P u r c h a s e s
Figure 8.12 (A-D) illustrates how the level of calories and the
product composition of purchases vary across zones in each of the four
seasons.
The level of calories obtained from purchased cereals and
pulses increases constantly from harvest to harvest in almost all zones.
Purchases start out relatively low during the harvest period. In most
zones levels purchased during the nonagricultural and planting season
are 1.5 to 2 times that purchased during the harvest period. During the
weeding season there tends to be another big jump, with amounts
purchased increasing another 1.5 to 2 times.
The composition of cereal purchases is mixed. Rice is an extremely
important share of purchases at harvest time because most households
have their home-produced millet to consume. During the nonagricultural
season the quantity of rice rises but the share remains steady or
declines as households in some zones begin purchasing millet.
This
behavior continues with both millet and rice purchases increasing in
tandem across the seasons until we reach the weeding period, when rice
purchases decline and millet increases.
Without examining the price data here, it is difficult to speculate
about the role of prices versus preferences in determining the mix of
cereals purchased.
Since the government-controlled price of rice does
not vary during the year, price response to rice is a function of the
relative price of rice and coarse grains.
When coarse grains are
inexpensive,
rice is most expensive; when the price of coarse grains
rises, rice becomes relatively less expensive.
:
:
_:
.
. 1
.‘.
Ava Y3d 3V Y3d S3lYOlV3
IV0 tr3d 37 Y3d S3tYOlY3
i
:
.
,#’
.,
I’
.
:.
.
’
.
.
185
The high processing losses associated with coarse grains means that
a simple comparison of per kilo prices for millet and imported rice is
inappropriate.
It is informative to compare the price of rice and
coarse grains per calorie rather than per kilo; most Senegalese cooks
probably do this unconsciously because they know that they need more
kilos of millet than rice to feed a given number of people. Using the
government price of 135 francs CFA (FCFA)/kilo of imported rice, the per
calorie price is 0.379 FCFA.
When the price of a coarse grain reaches
107 FCFA per kilo, this is equivalent to 0.379 FCFA per calorie. In
other words, a consumer interested in purchasing the most calories per
FCFA should be indifferent between rice at 135 FCFA per kilo and millet
at 107 FCFA per kilo.15
If price were the major factor in determining cereal purchases, we
would expect coarse grains rather than rice to be purchased during the
harvest period because the millet price was never above 100 FCFA at this
time.
We would also anticipate rice purchases to grow faster than
coarse grain purchases as the seasons progress and coarse grain prices
rise. This does not happen in most zones; millet purchases increase
significantly during the weeding period and rice purchases decline
somewhat.
Factors other than price are also at work here. Households
appear to have a preference for a particular mix of rice and coarse
grains.
As stocks of coarse grains decline, it is not a matter of
switching over entirely to rice but of purchasing a mix of the two types
of cereal.
Purchases of pulses account for a relatively small share (less than
10 percent) of total cereal/pulse purchases.
Peanuts are rarely
purchased during the harvest and nonagricultural season, but quantities
increase during the planting and weeding seasons. With the exception of
the Northern Peanut Basin, cowpeas are rarely purchased. This suggests
that there is room for developing markets for cowpeas outside the
principal zone of production.
Statistical tests of factors influencinq consumption patterns
Two-by-two tables and chi square tests (.05 level of significance)
were used to assess the statistical significance of factors that appear
to influence the level of caloric intake and the share of rice in cereal
consumption.
Differences in the level of caloric intake were found to be
statistically different across agroclimatic zones.
Households in the
higher potential zones of Senegal Oriental and the southeastern basin
I5 We are ignoring here the fact that the cost of the sauce that
goes with the millet or rice might differ and the decision to purchase
one cereal or the other involves an evaluation of the cost of the entire
meal, not just the cereal.
This is a more important consideration in
urban areas and market villages where "other foods" are a substantial
part of the diet, and less important in villages where sauces are made
from gathered products or home produced condiments.
186
had a greater probability of falling below the 2400 kcal/ae minimum
needs level while households in the Colobane zone of the central basin
and in the southwestern basin were more likely to exceed minimum needs.
Members of the Serer ethnic group are more likely to exceed the
2400 kcal/ae level while the Wolof, Manding, and Diakhanke are more
likely to experience inadequate consumption levels. Ethnic group and
agroclimatic zones are highly correlated, making it difficult to assess
the individual importance of each factor.
Households in market villages were shown to be statistically more
likely to purchase larger shares of their cereals, to consume larger
shares of rice, and to have a lower level of caloric intake from goods
analyzed.
Households that cultivated less than one hectare per AE (all zones
combined) were more likely than others to exhibit inadequate levels of
caloric intake.
We were unable to show any statistically significant link between
levels of caloric intake and dependency ratios, income categories,
degree of income diversification, cereal sufficiency ratios, and the
share of rice in household cereal consumption.
In other words, the
analyses conducted to date have not provided statistically significant
support for the hypotheses that these factors influence the level of
caloric intake.
More thorough forms of multivariate analysis will be
able to shed further light on these relationships.
Households living in a market village, earning more than 30 percent
of total household income in off-farm activities, and having lower
cereal production sufficiency ratios have a statistically significant
tendency to consume more imported rice than their neighbors.
Other
factors that one might expect to influence the share of rice in the diet
(dependency ratio, income category, and cultivated area per AE) were not
found to be statistically significant.
SUMMARY OF MAIN POINTS
By way of summary it is interesting to group the findings in three
categories:
conventional wisdoms confirmed, conventional wisdoms not
confirmed, and unanticipated results.
In the category of "conventional wisdoms confirmed" we include the
following points:
Millet is the most commonly consumed cereal in the Peanut Basin
(two-thirds of calories consumed) except in the north where
imported rice dominates.
In urban areas, rice provides 100 percent of cereal calories at
noon and 50 percent in the evening.
:
187
Rice consumption has made important inroads in rural areas,
particularly in the north, in the Colobane zone of the central
basin, and in the southwest (70, 32, and 26 percent of cereal
calories, respectively).
Maize consumption is important only in its zone of production,
Senegal Oriental (50 percent of rural and 12 percent of urban
consumption).
The northern Peanut Basin exhibits an essentially urban consumption
pattern due to its low levels of agricultural productivity and high
share of off-farm income in total income.
Purchases are an important share (at least 20 percent) of all
cereals consumed in rural areas; greater purchases are encouraged
by low levels of agricultural productivity and the monetization of
the rural economy (both agricultural and nonagricultural sectors)
which facilitates purchases of basic needs.
The consumption patterns in market villages are becoming
increasingly urbanized (more rice and more noncereal, nonpulse
products).
Overall, an important share (30 percent or more depending on year
and zone) of rural households do not meet the minimum levels of
caloric intake (2400 kcal/AE).
The most satisfactory levels of
consumption are found in the southwestern Peanut Basin and the
Colobane zone of the central basin.
In the category of "conventional wisdoms not confirmed" we have the
following finding:
The role of factors such as dependency ratio and level of income do
not seem to have a statistically significant effect on the levels
of caloric intake in the analyses conducted to date.
Given that
these are factors that normally exhibit an important effect on food
consumption patterns,
the final word on this issue will have to
await more thorough statistical analysis of the data.
The most surprising finding is that there appears to be a more
serious calorie deficit in the zones considered to have better
agricultural potential -- Senegal Oriental and the southeastern Peanut
Basin.
Limited access to off-farm incomes (particularly in the
southeast)
and
relatively less developed commercial
and
road
infrastructure (particularly in Senegal Oriental) are two factors that
might be keeping households in these zones from earning adequate
agricultural incomes and smoothing total household income throughout the
year.
The most important policy implication coming from the analysis of
consumption patterns is that purchased cereals play an important role in
assuring food security for both urban and rural households.
Imported
rice constitutes a substantial share of these purchases in urban areas
188
(70 percent) and throughout most of the rural zones covered in the
Peanut Basin (at least 20 percent). For this reason, it is extremely
important to examine the potential effect that a change in cereal prices
(either rice or millet) might have on the real income of Senegalese
households.
This issue is examined in greater depth in the next
chapter.
.’
:
‘.
189
REFERENCES
Fall, A.A., V. Kelly, and T. Reardon.
1989. Household-Level Survey Methods
Used in the IFPRI/ISRA Study of Consumption and Supply Impacts of
Agricultural Price Policies, IFPRI/ISRA Project Document 2, September,
IFPRI, Washington, D.C.
Kelly, V., T. Reardon, A.A. Fall, 6. Diagana, and L. McNeilly. 1992. Food
Consumption in the Peanut Basin and Senegal Oriental: October 1988 -
September 1989.
IFPRI/ISRA Project Document 7, September IFPRI,
Washington, D.C.
Hellegouarch, R., R. Giorgi, L. Monjour, and J. Toury. 1988. Enquete de
Consommation alimentaire dans une Zone pilote du Seneqal (1968).
Document du Travail, Organisme de Recherches sur 1'Alimentation et la
Nutrition Africaines (O.R.A.N.A.), Dakar.
Thirion, M-C. and J-L. Garnault. 1991.
"Le sous-espace ouest tourne vers
la mer," La lettre de Solaqral - Strateqies Alimentaires, no. 37,
April-May.
190
9.
PRICE RELATIONSHIPS AND THE IMPLICATIONS FOR CHANGES IN
RELATIVE PRICES
INTRODUCTION
The objectives of this chapter are two-fold:
(1)
to describe the methods used to develop price series for cereals
and pulses and make some observations about the price trends and
relationships during the survey period; and
(2)
to summarize results of price simulations conducted to examine the
potential household level impact of changes in relative prices of
principal crops.
PRICE TRENDS AND RELATIONSHIPS
A complete and accurate set of prices for the commodities of
interest is crucial to understanding the demand and supply behavior of
rural and urban households.
The IFPRI/ISRA survey was designed to
collect unit prices for all the agricultural products bought and sold by
sample households.
The method of collecting the data is described
briefly in Chapter 6 and in more detail in Fall, Kelly, and Reardon
(1989).
Two sets of consumer and producer prices were developed for each
study zone.
The first is a set of transaction-derived prices based
entirely on household-level survey data (Tables A9.1-A9.14, Appendix
9.1).
The second series uses the household data as a base but
supplements it with information from the ISRA/MSU market surveys
conducted from 1985 through 1988 and the CSA market surveys conducted
from the mid-1980s to the present (Tables A9.15-A9.20).
At present, the consumer price series is considered acceptable for
modeling demand and has been used in a number of preliminary runs. The
producer price series is still not complete. The very poor harvest in
1988 made it extremely difficult to obtain household-level observations
for producer prices
because there were few sales by the sample
households.
Because very few peanuts are sold after the end of the
official peanut marketing program,
it was also difficult to get a full
price series for the April through December period in both harvest
years.
Efforts to supplement the household data with market surveys are
still under way.
Appendix 9.2 discusses some potential problems with
the quality of the price data for particular products and zones.
,
191
Transaction-Derived Prices
A transaction-derived price (TDP) was calculated for each purchase
and
sale
recorded on
the
agricultural
product
transactions
questionnaire.
When unit prices were provided by the respondent, the
unit price was converted to a price per kilo using unit weights. If
unit price or unit weights were not available but total expenditure for
a transaction and total kilogram weight were available, this information
was used to calculate a price per kilo.
Monthly TDPs were then
calculated as the averages of all the prices recorded in each month
(weighted by the kilograms transacted).
Before the monthly averages were taken, one modification was made
to correctly match the prices with monthly time periods.
Initially,
each transaction price was allocated to the month of the interview, with
no regard for the recall period.
However, transactions with a recall
period that straddled more than one month were assumed to have been
distributed evenly over the recall period for the interview.
Consequently, these transactions were split into parts proportional to
the number of days falling in each month.
The final step was the
calculation of a TDP for each month.
For example, if 2 kilograms of
millet were purchased for 140 FCFA in the two-week period prior to June
7, then 1 kilogram would be allocated to June and one to May.
Both
months would then have a line of data for a millet transaction at a TDP
of 70 FCFA per kilogram.
The TDP series runs from October 1988 to October 1990, except for
the zone of Sagatta, where the series stops with October 1989. Separate
series were compiled for millet and/or sorghum, maize, unshelled
peanuts, shelled peanuts, and shelled cowpeas. Unfortunately, there are
few cases of a complete series for a good. In addition, several of the
prices are based on only a small number of transactions, raising the
question of their accuracy in reflecting market conditions. The next
section discusses how these missing and "thin" prices were estimated for
the full price series.
Full Price Series
In moving from a TDP series to the full price series, three
problems had to be addressed:
l
Measurement error from unit weights
l
Thin prices (few observations per month)
0
Missing prices
Measurement errors entered the analysis when the quantity per unit
was converted into kilograms
for all transactions that were not
.
.
192
originally made in kilograms.
A decision was made to use only
transactions enumerated in kilos or very common measures (for example,
tomato paste cans) for the price series.
Since most of the transactions
in the study zones were purchased in kilograms, this decision meant
eliminating fewer than 10 percent of observations.
Missing prices for the major crops (millet, peanuts, and so on)
were filled in using extrapolation techniques and secondary data, In
addition, "thin" prices based on only one or two transactions which did
not seem in line with seasonal trends were eliminated and then
estimated.
A number of procedures were used to estimate missing or
eliminated values; among the most common were:
Extrapolation by using the average price from adjoining
months;
Direct borrowing of a data point from the market survey data
if overlapping months of TDP and the market data were
comparable;
Borrowing with adjustment from market survey data if
overlapping months of TDP and market data followed a similar
trend but were not at identical levels;
Replacing the TDP with market survey data for months where
TDP data were very thin and contradicted other price
information.
Observations on Consumer Price Trends
This section discusses consumer price trends briefly to ascertain
whether they conform to predetermined expectations.
The discussion
refers exclusively to the full price series (Tables A9.15-A9.20).
Imported rice prices are government controlled and should be
consistently in the 130 to 140 FCFA/kilo range for the entire survey
period.
Prices in Senegal Oriental and the Southern Peanut Basin should
be somewhat greater than those closer to Dakar because of transportation
costs (the amount of the transport cost is also controlled).
Tables A9.1-A9.8 confirm these expectations. The price of rice
throughout the Peanut Basin is in the 130-135 FCFA/kilo range.
It is
lowest in the Colobane zone where households frequently purchase'large
sacks and receive slightly lower prices.
It is highest in Senegal
Oriental where it reaches a maximum of 140 FCFA/kilo.
Unlike rice, coarse grains are not subject to government price
controls.
Prices for coarse grains should be lowest at harvest
(October-November) and highest during the rainy season (July-September)'.
Prices should also be higher in the poorer harvest year (1988/89), and
,
.
.
.
;
Figure 9.1 -- Consumer price trends for millet/sorghum. (Source: ISRA/IFPRI
survey data, adjusted.
Note:
Prices in Missirah are for sorghum, while those
in other zones are for millet.)
121I-
IIC)-
IOC l-
80
70
60
50.
40.
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
Oct88
I
I
Jan89
I
Apr89
Jul89
I
Oct89
Jan90
Apr90
Jut90
Oct90
+ SAGATTA
- + - , NIAKHAR
~~*~~ COLOBANE
--El- MISSIRAH --X- PASSY
-& DIOLY
194
in the zones with greater production deficits (the Northern and Central
Peanut Basin in 1988/89 and Senegal Oriental in 1989/90).
Figure 9.1 shows millet and sorghum price trends for all rural
zones.'
In general the prices follow the pattern described in the
preceding paragraph. In Senegal Oriental, however, there are a number
of unexplained ups and downs:
these are more pronounced in 1988/89 than
in 1989/90.
In Passy and Colobane, prices are a bit more volatile than
in Niakhar and Sagatta.
The prices tend to rise in January and then
fall again until April/May when they begin to rise again and do so
continuously until the harvest.
The Southeastern Peanut Basin also
exhibits this tendency, but with much smaller movements. These price
movements reflect the fact that farmers start marketing peanuts in late
December and no longer need to sell cereals to obtain cash -- this cuts
down on cereal supplies available in the markets.
In comparing prices across zones, one notes that millet prices are
consistently higher in the Niakhar zone of the Central Peanut Basin
during both years.
This is true even though the Northern Peanut Basin
had a much lower production sufficiency ratio in 1988/89 than Niakhar
and is more distant from the cereal production zones.
Prices in the
Southeastern Peanut Basin are most often the lowest while prices in
Passy and Colobane tend to move around between the two extremes.
Maize prices are illustrated in Figure 9.2. Again, the highest
prices are in Niakhar zone of the Central Basin, with Colobane and
Sagatta prices being also quite high.
The Colobane price for 1989/90,
however,
is quite a bit lower than it was in 1988/89; we have no
explanation for this.
The lowest prices are in the production zone of
Senegal Oriental (Missirah), the Southwest Basin (Passy), and the
Southeast (Dioly).
The price series for Dioly remains incomplete for a
substantial number of months.
The seasonal trends are about as
anticipated,
with Missirah exhibiting almost as much volatility in
1988/89 maize prices as it did in millet/sorghum prices.
Maize prices are consistently greater than millet and sorghum
prices in all zones.
This is illustrated more clearly in Figures A9.1-
A9.6. The fact that more maize than millet is sold in already processed
form could be causing a slight upward bias to the maize prices (see the
discussion in Appendix 9.2), but this should not be a major problem
because much of the maize price series comes from market rather than
TDPs.
' Figures A9.1-A9.6 in Appendix 9.3 present price trends individually for
each zone.
Zone-specific trends are easier to see on these figures than on the
cross-zone comparisons presented in the text.
.’
:.
.
.
:’
‘.
’
.:
.
Figure 9.3 -- Consumer price trends for shelled peanuts. (Source: ISRA/IFPRI
survey data, adjusted.)
300-
2 8 Cl-
2 6 0l-
2 4 0
2 0 0
180
1 6 0
140,
IZO-
1 oo-
80- I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
Oct88
I
Jan89
I
I
Apr89
Jul89
octa9
Jan90
Apr90
Jul90
Oct90
-U- SAGATTA
-+- NIAKHAR
-3++. COLOBANE
-El-- MISSIRAH -X-- PASSY
-.A- DIOLY
Figure 9.4 -- Consumer price trends for cowpeas.
(Source: ISRA/IFPRI survey
data, adjusted.)
26CI-
24C)-
22cI-
200I-
180
160
. .
100
80
_:
60,
40,- -I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
-
~~
Oct88
Jan89
Apr89
Jul89
Oct89
Jan90
Apr90
Jul90
Oct90
-1c- SAGATTA
-+-- NIAKHAR
-316.~ COLOBANE
-B-- MISSIRAH ---X--. PASSY
-A- DIOLY
198
Figure 9.3 presents the rather complex picture of consumer prices
for shelled peanuts.
When they are not sold in official channels,
peanuts are often sold in shelled form. On average, a kilo of unshelled
peanuts yields 700 grams of shelled peanuts.
This implies that the
shelled price should be at least 1.43 times the unshelled price -- about
100 FCFA if the official price prevails for unshelled peanuts. In
developing the consumer price index for peanuts, quantities of unshelled
peanuts were converted to the shelled equivalent.
Prices tend to be lowest in November after the harvest has begun
and before the official marketing campaign has opened.
Prices
frequently peak during the planting season (May-July, depending on the
rains) and then again before the harvest (September-October).
Once again, Niakhar appears to have the highest prices.
The
Central Basin (Colobane) and the Southeastern Basin (Dioly) tend to have
the lowest prices.
As Chapter 5 revealed, these are the biggest peanut-
producing zones in the sample.
Figure 9.4 presents the consumer price series for cowpeas. The
series is very incomplete, but it does illustrate that prices in the
three northernmost zones where most production takes place are at
similar levels and follow similar patterns of seasonality.
Tables A9.9-A9.14 present the producer price series based entirely
on transaction-derived prices.
Average producer prices for peanuts are
70 FCFA/kilo in all zones because such a large share of peanuts is
marketed at the official price.
Producer prices of cereals are more
volatile and also more difficult to explain without doing some
econometric analyses.
In some cases there are producer prices that are
higher than consumer prices in the same zone. This is often due to the
thinness of the data.
It can also be due to lumping different types of
sales together.
For example,
there are numerous cases of sales to
friends and neighbors.
The prices for these sales may give consumer
prices that are equal to or even lower than normal producer prices.
More time is needed to thoroughly examine the implications of the trends
in the producer price series.
Additional Work Planned
It is difficult to develop complete price series from household
survey data.
It is very important that it be done correctly because
mistakes in the data or incorrect decisions made to fill in missing
values can lead to serious problems in demand and supply analysis. The
199
fact that market survey data are available in Senegal renders the task
somewhat more tractable than it would have been otherwise, even though
observations that are missing from household data are frequently also
missing from market survey data.
On the demand side, the next step is to begin examining
relationships between purchases and sales behavior presented in Chapter
6 and the price trends presented in this chapter.
Tables A9.1-14
provide a feeling for what such an analysis will yield because the
kilograms transacted each month are presented with the price data.
Demand in rural households, however, includes purchases as well as
consumption from home production.
To conduct a complete analysis of
demand, one would want to know how consumption of home-produced cereals
as well as purchased cereals responds to price changes.
On the supply side, a great deal of work remains because of the
many gaps in the producer price series.
Ultimately, we hope to examine
the relationship between aggregate production (discussed in Chapter
5),
producer prices, and marketed surplus of cereals and peanuts.
In addition to the demand and supply analysis, it would be useful
to look at price formation patterns and market integration within and
across agroclimatic regions.
This type of analysis is not planned as a
part of the current project but the data lend itself to such analyses.
One issue that could be addressed by this type of analysis is why
consumer prices are consistently higher in the Niakhar zone of the
Central Peanut Basin even though the roads and commercial infrastructure
in the area appear to be better than that in other zones.
HOUSE-HOLD LEVEL IMPACTS OF CHANGES IN RELATIVE PRICES
Background
Over the last decade two issues concerning trade regime policy
have been hotly contested in Senegal:
(i) devaluation of the franc CFA
and (ii) increases in the tariff on rice imports. Both of these issues
have been central to the debate on Structural Adjustment policies. To
date, devaluation has not been attempted in Senegal. During the 1980s
a number of changes were made in the consumer price of rice, with
substantial increases in 1982, 1983 and 1985.
These increases were
partially rescinded in 1988 and again in 1989 following protests in
urban areas.
.
.
.,
.’
.’
200
It can be argued that trade regime changes are necessary for
efficient long-term growth, both of the food and cash crop subsectors
and of the overall economy.
Timmer (1981, p. 209), however, exhorts us
by stressing that an important part of policy analysis is "facing
squarely the short-run welfare consequences of efficient long-run food
sector development strategies".
Timmer's observation is particularly important in the Senegalese
context where data are seldom available to evaluate the potential
household-level impact of trade regime changes.
Our inability to
predict potential impacts for different groups of people makes it
extremely difficult to develop temporary policy interventions that will
protect
'at risk' groups from bearing an unanticipated share of the
negative household-level consequences
associated with such policy
interventions.
For this reason, governments tend to find themselves
initiating policy changes that must later be rescinded in the face of
popular uprisings that threaten political stability.
The objective of the analyses is to simulate the potential impact
of selected trade regime changes on different types of Senegalese
households.
The objective is not to provide arguments against reforms
in macro-economic policy that are clearly needed for long run economic
development but rather to provide policy makers with information that
can be used to ensure that macro-economic policies succeed.
Using different hypothesis about devaluation and changes in rice
tariffs, we try to better understand which types of households will
gain, which types will lose and what will be the magnitude of the gains
and loses.
The method used is partial equilibrium and comparative static. We
limit ourselves to identifying the upper bounds for potential direct
effects of policy changes on peanut, rice, and coarse grain prices.
First, we present a maintained hypothesis (not tested) of how a given
trade regime change would affect the prices households face (e.g. how a
devaluation would affect the rice price for consumers). Second, using
information on patterns of household behavior described in previous
chapters (e.g. what is the share of rice in the budgets of various
groups of consumers), we surmise how the price change would change
nominal income on the production side (how much more cash the households
income is worth), or reduce real income on the consumption side (how
much more cash households have to spend to get the same amount of
product).
2
,
a’
:
.
:
201
The results summarized here are based on analyses conducted in
November 1992 and reported in a seminar paper "Potential Welfare Impacts
of Trade Regime Changes on Households in Senegal" (Reardon et al. 1992)
presented at Saly Portugal (Seminar on Regional Integration of
Agricultural Markets in West Africa). We have not been able to revise
the analyses to incorporate very useful suggestions made by conference
participants.
Some of the revisions subsequently made to the base data
files (particularly changes in consumption levels and composition for
the southeastern Peanut Basin) have also not been included.
It is
unlikely, however, that these revisions will substantially change the
overall conclusions of the original paper.
We turn now to a brief
review of hypothesis and assumptions underlying the analysis.
This is
followed by a summary of the results and a brief discussion of the
policy implications.
Hypothesis and Assumptions
The analytical question is what is the welfare impact on rural
households of a change in a price facing the households?
In this
analysis the price change is caused by a devaluation (change in tariff
for example), but the reason for the change could be anything.
We focus on the key tradeables and non-tradeable in the food
sector.
Devaluation would:
(a)
Raise the export border price of peanuts in CFA Franc equivalent
and -- with no government intervention -- raise the farmgate
price, given a world dollar price of peanuts;
(b)
Raise the imported price of rice and of comestible oil (at border,
and with no government intervention, to the consumer), given a
world dollar price of rice and of comestible oil;
(c)
Depending on the transmission effect of a rise in the price of
tradeable rice on non-tradeable coarse grains (call it millet
here), raise the price of millet
(4
Raise the price of imported inputs such as fertilizer and
fungicides.
The impacts of (a)-(d) on household welfare are conceptually as
follows. In each case we assume commerce and manufacturing absolute
margins stay the same, for simplicity.
(a)
An increase of the peanut price raises producer nominal incomes,
and decreases consumer real incomes (to the extent that peanuts or
peanut oil is consumed);
(b)
An increase in the rice price and imported comestible oil price
decreases real incomes of consumers;
‘.
.
.
I
.
.
‘.
,’
.
202
(c)
An increase in the rice price might or might not raise the millet
price;
assuming
it raises millet prices,
but weakly (weak
substitutes, see below), this will increase producer nominal
incomes and reduce consumer real incomes.
(d) An increase in imported crop input prices decreases producer
nominal incomes.
For any "average household" in a specific zone, the combination of
(a)-(d) will produce a net effect on nominal and on real income. These
four channels are simulated, given consumption, production, and income
patterns (drawn from survey data) and based on certain assumptions
concerning the 'transmission effect' from rice price to millet price,
and the 'passthrough policy' from FOB price change to producer price
change for peanuts. These latter two are discussed below.
The above direct, instantaneous effects in practice depend on the
answers to three key empirical questions, and one policy/empirical
question, as follows.
(i)
What is the share of each of the three goods above in consumption
and in production over groups and over zones? If one assumes that
home consumed goods can be given an imputed value of the market
price, then a change in price, regardless of how much of a good is
sold or bought by a household, will change the nominal income and
real income of the household in proportion to their entire
consumption or production of the good in question. For example, if
the price of millet rises, and a third of a household's income is
composed of millet sales plus home consumption of millet, then the
price rise increases the value of the entire production of millet,
regardless of marketed share; the household could trade all the
millet for more of another good than it could previously.
The evidence for these shares for consuming/producing households
in rural Senegal has been presented in previous chapters of this
report.
ii)
What is the share of imported inputs in agricultural production,
and how does this differ by zone?
iii) What is the "transmission effect" of a rise in the price of one
consumable or producible to another? This question applies to rice
on the consumption side and imported inputs (fertilizer and crop
protection chemicals) on the production side.
The most pertinent but also the most difficult question here is
whether and to what degree a rise in the price of rice (occasioned
203
by devaluation or protection) is transmitted to the price of
millet. In theory, if the goods are strong substitutes, the price
effect on millet will be strong (if substitutes, the price of
millet will rise sharply; if complements, it will fall sharply).'
We perceive the conventional wisdom in the Sahel as being
that increases in imported rice prices would have immediate strong
impacts on coarse grain prices. Very little empirical research has
been done for the Sahel on the determinants of grain prices, and
transmission effects of macro policy changes on sectoral prices.
Consumption studies in the Sahel, including in Senegal, generally
show the cross-price elasticity of rice and coarse grains to be
either very low, insignificant, or negative and significant.
Preliminary analyses of ISRA/IFPRI survey data for rural
areas suggest that some substitution of coarse grains for rice
does occur when millet becomes less expensive relative to rice.
There is ample evidence,
however, that non-price factors (tastes
and access to commercial infrastructure,
for example) also
influence the relative shares of rice and coarse grains in the
diet.
Evidence from ISRA/IFPRI data also suggests that the
tendency to substitute in urban areas is much lower as meals based
on coarse grains are rarely eaten at noon.
In our simulations of a rice price increase in rural Senegal
we opt for a middle ground assumption that rice and coarse grains
are weak substitutes.
This implies that part of the rice price
increase will be passed on to coarse grain prices as consumers
move away from rice and increase demand for coarse grains.
In the
simulations below, we assume that 20 percent of the rice price
increase is passed on to millet.
The second aspect of the 'transmission' question concerns
inputs:
To what extent is a price increase in (imported)
fertilizer,
fungicide, and other production inputs used by
different agroclimatic zones transmitted to the prices of other
products? One can also consider indirect effects that might work
through price increases in imported fuel (transport costs, for
example). Increases in prices of imported inputs should also be
expected to be transmitted to some degree (again, depending on
degree of substitutability) to other (non-traded) inputs such as
' By incorporating a passthrough effect on mi llet prices, we dev iate
from a strictly "static" analysis, but we believe
it better reflects the
way rural households will respond.
.,;.
.
.
:
.’
204
labor.
iv)
The combined policy and empirical issue concerns how devaluation
will affect domestic peanut prices.
We assume that Senegal is a "small country" in the world peanut
market.
This implies that Senegal can market as much as it wants at the
current world market price and therefore has no interest in selling at
a lower price.
In other words, Senegal maximizes peanut revenues by
selling at the established world price rather than by lowering its price
and increasing the quantity of its sales.
Given the decline in peanut
production in recent years, and indications that worldwide demand for
vegetable oil products is on the rise,
this is a reasonable short-term
assumption.
The alternative assumption is that Senegal cannot market all of
its production at the world market price and has an interest in cutting
prices to increase the volume of exports. For example, Senegal could
undercut other sellers with a lower price in a contract for peanut
deliveries to Nigeria.
We do not use this assumption directly in the
simulations.
Returning now to the small-country assumption, let us also assume
that a marketing board does the actual exporting of peanuts. Then, a
devaluation would increase the number of FCFA that the producer or
marketing board or both could earn given the world dollar peanut price.
The equity impact would then depend on the distribution of the
gain, which we call the "passthrouoh po1ic.v". Assuming a 50 percent
devaluation,
and the marketing board passed the whole increase on to
producers, the farmgate price in FCFA would rise 50 percent. If only
half were passed on (with the marketing board keeping the difference, a
more common scenario in developing countries that export agricultural
commodities), then the producer price increase would only be 25 percent.
We ignore here what the marketing board does with its share of the
increase.
As mentioned above, this analysis is static and partial. There
are a number of important issues concerning the dynamic, indirect, and
intersectoral effects of trade regime changes that are not addressed in
these simulations.
Simulatinq Devaluation
Two points made in previous chapters of this report are
particularly pertinent to the present discussion of the potential
impacts of devaluation:
2 0 5
(1) Average levels of caloric intake are, at the margin, barely
adequate in both rural and urban areas.
This suggests that any
policy that reduces real income will be likely to put more
households in the 'at risk' category unless non-food expenditures
are reduced by the full amount of real income lost and transferred
entirely to food expenditures.
(2)
Rural household income comes from a variety of sources. The share
of crop production income in total household income ranges from 22
to 80 percent.
The smaller the share of crop revenues in total
revenue, and the smaller the share of peanut production in total
crop revenues, the less likely a rural household will be to
benefit from the positive effects that a devaluation might have on
producer prices.
Table 9.1 shows the average share of household income derived from
crop production for each rural zone.'
Part A of the table shows the
shares of different crops in total gross agricultural revenue.
Peanuts
account for 60-70 percent of agricultural revenues in four of the five
zones in the Peanut Basin.
Cereals account for most of the remaining
income,
with very small shares of income coming from cowpeas and
cotton.3
Part B of Table 9.1 shows that the share of gross crop revenue in
total household revenue ranges from a low of 20% (this was during a very
bad harvest year
in the northern Peanut Basin) to 75% in the
southeastern Peanut Basin.
In all zones but the northern Peanut Basin,
crop revenues averaged at least 50% of total income during the two year
period (1988/89-1989/90).
During the poor harvest year 50-60% of income
came from crops in most zones.
In the good harvest year, 70-80% of
income came from crop production in most zones.
Table 9 C and D show gross peanut and coarse grain incomes as
shares of total household income.
Colobane in the central Peanut Basin
and in the southeastern Basin earned approximately 50 percent of their
income from peanuts during the two year survey period.
Niakhar in the
central basin and Senegal Oriental earn the largest share from cereal
' Note that these figures differ from results in chapter 4 which
have been revised to show net rather than gross cropping income. By
doing the simulation with gross cropping income, the positive impact of
an increase in producer price of peanuts is upwardly biased.
3 Data for the 1988 cotton harvest in Missirah is not included in
this analysis; in 1989 sample households in Missirah boycotted cotton
entirely.
.
:
.
‘.
206
production, about 40 percent during the survey period.
iable 9.1 -- Incane Patterns: (Averages 1988 & 1989)
N. P. 8.
Central P. B.
S. P. B.
Sen. Or.
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
Peanuts
0.69
0.44
0.68
0.60
0.66
0.29
Cereals
0.23
0.57
0.31
0.41
0.31
0.70
Cowpeas
0.09
0.00
0.01
0.00
0.00
0.02
Cotton
0.00
0.00
0.00
0.00
0.03
0.00
N. P. 6.
Central P. B.
S. P. 8.
Sen. Or.
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
0.20
0.68
0.70
0.50
0.75
0.58
N. P. 8.
Central P. 8.
S. P. 8.
Sen. Or.
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
0.14
0.30
0.48
0.30
0.50
0.17
N. P. 8.
Central P. B.
S. P. B.
Sen. Or.
Sagatta
Niakhar
Colobane
Passy
Dioly
Mssirah
0.05
0.39
0.22
0.21
0.23
0.41
Source:
ISRA/IFPRI survey data, 1988-1990.
Table 9.2 uses the information in Table 9.1 to simulate the
potential effect of devaluation on nominal income.
We present two
assumptions about how the devaluation will affect producer price of
peanuts and one concerning a passthrough effect that increases the
millet price.
Assuming a 50% increase in the domestic producer price of peanuts
with full passthrough to the producer, the producer's nominal income
will increase from 7 to 25% depending on the zone.
The lowest increase
is in the northern Peanut Basin, the highest in the central (Colobane)
and southeastern Peanut Basin.
207
TabLe 9.2 -- Policy Effects Via Production Output Side (average 1988 and 1989)
Part A: Specific peanut and cereal effects
5D%'hikcin.domestic producer price of peanuts due to devaluation;
fixed dollar price on world market; full passthrough
Increase in nominal income due to full passthrough on peanut price
(1)
N. P. Ei.
Central P. B.
S. P . B .
Sen. O r .
Sagatta
Niakhar
c01obane
Passy
Dioly
Hissirah
Peanuts
0.07
0.15
0.24
0.15
0.25
0.08
.,
,.
,. .,.
.,
A$smnjLDw 2:
5O%"hike'in"bomestic
producer price of peanuts due to devaluation;
fixed dollar price on world market; half passthrough
Increase in nominal income due half passthrough on peanut price
(2)
N. P. B.
Central P. B.
s . P . 8.
Sen. O r .
Sagatta
Niakhar
Colobane
Passy
Dioly
Missirah
Peanuts
0.03
0.07
0.12
0.08
0.12
0.04
Arsstmprl q i‘3 ;
Devaluation raises rice price by 50% causing coarse grain price to
increase by 10% thereby increasing nominal income from cereal production
Increase in nominal income due to cereal passthrough effect (3)
N. P. B.
Central P. El.
s. P . B.
Sen. O r .
Sagatta
Niakhar
c01obane
Passy
Dioly
Wissirah
Cereals
0.00
0.04
0.02
0.02
0.02
0.04
Full passthrough plus cereal effect (4)
N. P. 6.
Central P. B.
s. P . B.
Sen. Or.
Sagatta
Niakhar
Colobane
Passy
Dioty
Missirah
0.07
0.19
0.26
0.17
0.27
0.12
Half passthrough pius cereal effect (5)
N. P. B.
Central P. B.
s. P. B.
Sen. O r .
Sagatta
Niakhar
c01otxme
Passy
Dioty
Hissirah
0.04
0.11
0.14
0.10
0.15
0.08
..;. :
.
.
.
;
208
If only half of the benefits of devaluation are passed through to
the producer, the nominal income effect is halved and ranges from 3 to
12 percent.
In hypothesizing about effects on cereal prices, we assume that
the devaluation raises imported rice prices by 50% and that there is a
20% passthrough effect on millet prices which increases the millet price
by 10%.
This price change increases nominal income by only 0 to 4
percent, depending on the zone.
Part B of Table 9.2 shows the total production effect after
combining the peanut and millet effects. With full pass through on the
peanut side, nominal income increases from 7 to 27 percent. Zones with
peanuts accounting for more than 60 percent of crop income (northern
basin, Colobane in the central basin, and the southeastern basin), a@
crop income accounting for more than 70 percent of total income
(Colobane and the southeastern basin only) realize increases greater
than 25%.
With the half passthrough assumption on peanuts -- a more likely
scenario -- increases range from 4 to 15 percent,
The total production effect in Part B is also the net production
effect after accounting
for anticipated price effects on the
agricultural
input side.
This occurs because use of tradeable
agricultural
inputs (fertilizer,
insecticide,
herbicide) is at an
extremely low level in Senegalese rainfed agriculture.
Table 9.3
illustrates that the negative effect via purchases of tradeable inputs
is insignificant given rates of input use reflected in data for the
1989/90 agricultural season.
Table 9.4 illustrates consumption patterns during the 1988-89
period, showing share of expenditure devoted to rice, coarse grains,
peanuts,
vegetable oil, and all other foods,
We concentrate on the
first four product groups as they are the ones with prices likely to be
affected by devaluation.
Across all zones, cereals as a group account for about 50 percent
of total expenditure (Part C).
About a third of total expenditure is on
coarse grains in all zones but the northern Peanut Basin (18 percent)
and the Niakhar zone of the central Peanut Basin where the Serer
maintain strong preferences for millet, spending as much as 44 percent
on coarse grains.
209
Table 9.3 -- Producticm Input Expenditure Patterns and Policy Effect.
Centrat P. B.
S. P. 8.
Sen. Or.
Sagatta
Niakhar
Colobane
Passy
Dioly
Missirah
1.
Gross ag
rev 1989
5.15
7.92
15.49
10.20
22.16
11.96
season in
millions FCFA
2.
Level of expen.
on fert. in
insig
insig
insig
0.13
0.15
insig
millions FCFA
3.
Level of expen.
on other var.
inputs (ext.
0.13
insig
0.04
insig
0.15
insig
seed and labor) in
millions FCFA
4.
Total above inputs
millions FCFA
0.13
insig
0.04
0.13
0.30
insig
5.
Share of above
inputs in gross
ag rev
0.02
nsig
insig
0.01
0.01
insig
6.
Share of above
inputs in total
income
insig
nsig
insig
0.01
0.01
insig
Effect
insig
insig
insig
insig
insig
insig
of 50%
deval. on income
via inputs
Source:
ISRA/IFPRI Survey
Note:
FCFA values are suns over all househoLds in the sample for 1989/90 production season.
Data not available for production season 1988/89.
‘.
.
210
Table 9.4 -- Expenditure Patterns: Selected Products as Shares of Total Expenditure
Part A: Share of Rice in Total Expenditure
II. P . B.
Central P. 8.
S.P.B
Sen. Or-U
Sagatta
Yiakhar
Colobane
Passy
Dioly
Missirah
1988
0.27
0.09
0.19
0.18
0.08
0.06
1989
not available
0.09
0.14
0.14
0.04
0.03
1988-89
not available
0.09
0.16
0.16
0.05
0.04
Share of Coarse Grains in Total Expenditure
Part B:
N. P. B.
Central P. B.
S.P.B
Sen. 0r.U
Sagatta
Yiakhar
c01obarle
Passy
Dioly
Missirah
1988
0.18
0.44
0.26
0.38
0.41
0.43
1989
not available
0.44
0.20
0.29
0.29
0.36
1988-89
not available
0.44
0.23
0.33
0.34
0.39
Part C* Share of Cereals in Total Expenditure (A +
A
8)
N. P. B.
Central P. B.
S.P.B
Sm. 0r.U
Sagatta
Niakhar
c01obane
Passy
Dioly
Wissirah
1988
0.45
0.53
0.45
0.56
0.49
0.49
1989
not available
0.53
0.34
0.43
0.33
0.39
1988-89
not available
0.53
0.39
0.49
0.39
0.43
Part D.
- Share of Peanuts and Vegetable Oil in Total Expenditure
Y. P. B.
Central P. 8.
S.P.B
Sea 0r.U
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
1988
0.16
0.10
0.15
0.11
0.10
0.12
1989
not available
0.12
0.10
0.12
0.05
0.11
1988-89
not available
0.11
0.13
0.12
0.07
0.11
Part Em
A Share of All Other Foods in Total Expenditure
N. P. B.
Central P. B.
S.P.B
Sen. 0r.U
Sagatta
N i akhar
cotobane
Passy
Dioly
Hissirah
1988
0.24
0.25
0.23
0.18
0.20
0.18
1989not
not available
0.22
0.32
0.17
0.26
0.21
1988-89
not available
0.23
0.27
0.18
0.23
0.20
Part F.
A Share of All Food in Total Expenditure (C + D + E)+
N. P . B .
Central P. B.
S. P. B.
Sen. 0r.U
Sagatta
Niakhar
Colobane
Passy
Dioly
Missirah
1988
0.85
0.88
0.83
0.85
0.79
0.78
1989
not available
0.87
0.76
0.73
0.65
0.71
1988-89
not availabLe
0.87
0.79
0.78
0.70
0.74
Source:
ISRA/IFPRI survey data 1988-1990
.!
211
Rice expenditure reaches a high in the northern Peanut Basin (27
percent), ranges from 9 to 16 percent in the central and southwestern
Peanut Basin, and declines as one moves eastward to the Southeastern
Peanut Basin and Senegal Oriental (4 to 5 percent).
Peanuts and oil range from 7 to 16 percent of expenditures, with
most zones being in the 11 to 13 percent range.
All foods combined {including those not specifically mentioned
above) account for 70 to 87 percent of total expenditure.
The share
allocated to food is generally greater (7 to 15 percent) following the
poor harvest of 1988.
Table 9.5 presents the hypothesized direct effects of a 50%
increase in the price of imported rice and the indirect price effect
that is passed through to coarse grains.
As mentioned earlier, we
examine these impacts assuming that the relative shares of rice an
coarse grains in the diet do not change in response to the price
changes.
While this is not an entirely unreasonable assumption in urban
areas where coarse grains are very unlikely to be substituted for rice
at the noon meal, it is a less tenable assumption for rural areas where
a variety of coarse grain dishes are consumed at all times of the day.
We see that the rice price effect results in a decline in real
income ranging from 2 percent in areas of low rice consumption (Senegal
Oriental and the southeastern Peanut Basin) to 14 percent in the
northern Peanut Basin.
Income declines by 5 to 8 percent in other rural
zones.
Although we do not present a full analysis for the urban areas
of Kaolack and Tambacounda, we expect the effect in these towns to be
comparable to that of the northern Peanut Basin or greater as the shares
of calories obtained from rice are equal or greater.
Given our hypothesis about the 20% passthrough to coarse grain
prices, the drop in real income increases another 2 to 4 percent across
all zones.
The size of the passthrough effect in the short-run depends
on how price elastic demand for rice and coarse grains are.
If small
increases in rice price push consumers to reduce rice consumption and
increase coarse grain consumption substantially the price of the latter
good will rise rapidly.
The medium-run size of the passthrough effect
depends to a large extent on the supply responsiveness to increases in
millet prices.
Assuming that the devaluation increases the producer
price of peanuts more than it increases the producer price of coarse
grains,
there is unlikely to be a large coarse grain supply response
because peanuts will continue to be a more profitable crop.
.’
.
212
Table 9.5 -- Policy Effects via Decreases in Real Income on Cmsurption Sick
.,.
Paf-jz:~~:.I$ff&%s”id&, :~~~:::~i~Fcj:;product~:by'~~~r::~~::'::~~
.,... . . . .,.,.,
,. . . ,.
.
AS,+HPTIdi( 3:
Rice price increases by 50%
Decrease in real income due to change in rice price (I)
N. P. B.
Central P. 8.
S. P. 8.
Set-~. 0r.U
Sagatta
Niakhar
c01obane
Passy
Dioly
Hissirah
1988
0.14
0.05
0.09
0.09
0.04
0.03
1989
not available
0.04
0.07
0.07
0.02
0.02
1988-89 not available
0.05
0.08
0.08
0.03
0.02
ASW.TXDN..Zi
..,. . . . . . . . . ..A. :;....i. ..A.., . . . . .
Coarse grain price up by 10% c&z to passthrough frm rice (2)
Decrease in real income due passthrough effect on coarse grains
N. P. B.
Central P. B.
S. P. B.
Sea 0r.U
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
1988
0.02
0.04
0.03
0.04
0.04
0.04
1989
not available
0.04
0.02
0.03
0.03
0.04
1988-89 not available
0.04
0.02
0.03
0.03
0.04
@,.$k&$J~~~~j3 Price of peanuts and peanut oil increase by 50% (3)
Decrease in real income due to peanut and oil price effect
N. P. 8.
Central P. 8.
S. P . B .
Set-~. 0r.U
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
1988
0.08
0.05
0.08
0.06
0.05
0.06
1989
not
available
0.06
0.05
0.06
0.03
0.06
1988-89 not available
0.06
0.07
0.06
0.04
0.06
Part'?: ~fotal:‘d~rea&in reat’:iya+ ,,F”average durinq:S988;1989 period,(&)
. . . ::. .“. ,.
N. P. B.
Central P. 8.
s. P. 0.
Sen. 0r.U
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
Rice
0.14
0.05
0.08
0.08
0.03
0.02
MMS
0.02
0.04
0.02
0.03
0.03
0.04
PN/Oi1
0.08
0.06
0.07
0.06
0.04
0.06
Total
0.24
0.14
0.17
0.17
0.09
0.11
.
‘
.
.
213
Although peanuts serve as the major cash crop throughout most of the
rainfed areas of Senegal, the crop is also an important food crop.
Peanuts are consumed directly as a condiment in sauces or a snack food
and indirectly as processed oil.
Vegetable oil consumed is of three
types: home-processed, locally produced peanuts, industrially processed
peanuts,
industrially processed oils from imported oil seed products.
The price of all three types of vegetable oil will be influenced by a
devaluation.
As rice consumption has increased over time, there has been a
tendency for oil consumption to also increase (large quantities of oil
are used to prepare the most popular Senegalese rice dish 'tcheb u
jen').
The ISRA/IFPRI data suggest that there is a secular trend toward
increased rice and oil consumption as rural areas tend to be adopting
more urban consumption patterns.
In the current analysis, we hypothesize that the price of all peanut
and oil products
increase by 50 percent due to the devaluation
(assumption 3, Table 9.5).
This results in a 6 to 7 percent decrease in
real income in most rural zones.
Part B of Table 9.5 summarizes and adds up the three consumption-side
policy effects described above.
Using the average effect across the two
years
of survey data we find that real income declines by 9
(southeastern Peanut Basin) to 24 percent (northern Peanut Basin). The
effect is strongest in the north, west, and center of the Peanut Basin
and diminishes as one moves east to the southeastern basin and to
Senegal Oriental -- the two zones with lowest rice and oil consumption.
Net real income effects followinq a hypothesized devaluation
Table 9.6 presents the final picture of the overall impact on rural
households.
Using our best case scenario (full passthrough for peanuts
and 20% passthrough from rice to millet prices) the net effect on real
household income ranges from a 17% decline (northern low production and
high rice consuming zone) to an 18 % increase (southern high peanut
producing and low rice consuming zone).
Two of the relatively high
potential zones (the southwestern basin and Senegal Oriental) are
virtually unaffected by the change,
In the central Peanut Basin, the
Colobane zone, which produces more peanuts and consumes more rice than
the nearby zone of Niakhar,
fairs much better (9% increase in Colobane
verses a 4% increase in Niakhar).
‘.
~
:
:
:
‘.
214
214
Table 9.6:
Calculation of Net Effects on Real Household Inccm 1988-89
N. P. B.
C e n t r a l P . 6.
S . P . 8.
Sen. Or.
S a g a t t a
Niakhar
Colobane
Passy
D i o l y
Hissirah
Total
Negative
-0.24
-0.14
-0.17
-0.17
-0.09
-0.11
Consumption
Effect (1)
Total
Positive
Production
Effect
0 . 0 7
0.19
0.26
0.17
0.27
0.12
Full
Passthrough (2)
Net
Effect
Ful I
- 0 . 1 7
0.04
0.09
-0.00
0.18
0 . 0 1
Passthrough (3)
Total
Positive
Production
Effect
0.04
0.11
0.14
0.10
0.15
0.08
Half
Passthrough (4)
Net
Effect
H a l f
- 0 . 2 0
-0.03
-0.03
-0-08
0.05
-0.03
Passthrough (5)
,
:
_,;
:
215
The last line of Table 9.6, using the half passthrough option,
presents a more realistic picture of what might really happen following
a devaluation.
All zones but the southeastern Peanut Basin experience
a loss in real income.
The increase in income is 5%. The decline in
income ranges from an extreme of 20% in the north, to the more typical
case of 3% declines in the central basin and Senegal Oriental.
The
southwestern basin suffers slightly more with an 8% decline in
purchasing power.
This analysis suggests that the ranking of zones by
income prior to a devaluation will no longer hold after devaluation.
The northern Peanut Basin will clearly move from one of the wealthier
zones to one of the poorest.
The southwestern Peanut Basin may also
drop from being one of the top three income earners. The southeastern
Peanut Basin, the wealthiest zone prior to devaluation, will remain the
wealthiest after devaluation.
Although we do not do a separate simulation for the urban zones,
their consumption patterns are similar to those in the northern Peanut
Basin.
Given that urban households do not realize the gains associated
with peanut production, we would anticipate the net effect in urban
areas to be somewhat more negative than that for the northern Peanut
Basin -- at least a 25 percent drop in real income.
To repeat, the key factors in determining whether one gains or losses
following this hypothesized devaluation are (1) the importance of
peanuts on the production side (those with a greater share of peanut
revenues gain proportionately more) and (2) the importance of rice on
the consumption side (the more rice consumption the more one's real
income declines).
Given the descriptive information on consumption and income patterns
presented in the rest of this report, one can draw some tentative
conclusions about types of households and individuals most likely to
gain or lose from a devaluation.
Rural households in the poorest tercile are more likely to produce a
larger share of cereals than of peanuts because they cannot afford
peanut seed (poor access to credit) and they need to devote a large
share of land to cereal crops to assure a minimum level of food
security.
This means that poorer households will be less likely to
benefit from the positive production side effects associated with peanut
production.
Rural households in the poorest tercile are net cereal purchasers.
Of cereals purchased, imported rice accounts for 40-60 percent.
In the
northern, central, and southwest Peanut Basin rice accounts for anywhere
from 9 to 39 percent of total cereal/pulse calories, depending on the
,i/.
:
,
.!
216
survey year.
Unless these households are able to increase coarse grain
production or purchase coarse grains at predevaluation prices, they will
experience substantial loss in purchasing power after devaluation.
Given that poor households were unable to significantly increase crop
production during the favorable 1989 season and given that they are not
big peanut producers, the net effect is likely to be quite negative for
the poor tercile in these zones.
The fact that these are among the most
densely populated regions in Senegal magnifies the problem.
Women do not generally earn a great deal of income from peanut and
cereal production.
When women do have their own fields, however, they
usually produce a cash crop such as peanuts or cowpeas rather than
cereals (this pattern has been less pronounced since the government
stopped their seed distribution program in the mid-1980s). As women are
generally not responsible for purchasing family cereal supplies with
their own income, they stand to realize a net gain from their peanut
production.
This same scenario would be true for other household
members who grow peanuts and have little or no responsibility for cereal
supply ("sourga," unmarried adult males, dependent household heads).
On the other hand, the household head on whose fields the family
cereal supply is grown, will gain less than other household members to
the extent that he specializes in cereal production rather than peanuts.
Most household heads grow both cereal and cash crops.
One scenario
might be for the household head to increase peanut production in
response to the rise in nominal peanut prices and cut back on millet
production believing that the increased revenue will be adequate to
cover the cereal shortfall with purchases.
If enough producers do this,
coarse grain supplies are likely to decrease putting further pressure on
cereal prices.
The advocates of devaluation have the high ground in that it is clear
that there are very severe fiscal and balance of payment deficits and
currency overvaluation in the Sahel.
On the other hand, evidence presented in this chapter suggests that
only one rural zone survives a half-passthrough devaluation with a net
gain for the average household.
It is clear that urban households on
average risk a sharp drop (20-25 percent) in real income as most urban
households are in the service sector or local commerce and not likely to
realize any compensatory production side benefits.
We have pointed out that one of the benefits of examining potential
effects of a devaluation on households is that it permits one to design
policy interventions that lessen the negative impacts for households
’
5
.
.’
,
.
:
217
likely to experience large loses in real income and to stimulate the
positive impacts from the production side.
Three broad categories of interventions come to mind:
Investments in reversing structural constraints to agricultural
supply response, especially in the high potential zones.
Second is investment in ways to make coarse grain processing
cheaper and more accessible so that local grains can gradually replace
rice as the grain that meets the needs of the poor in growing cities.
Third is investments in infrastructure and transport, and
certainly protection from spending cuts in this domain, so as not to
allow devaluation and public spending cuts to increase price instability
and transaction costs, thus choking off supply response.
It is evident that in the very short-run, supply response will not be
adequate to provide consumers with large stocks of low cost coarse
grains that will be able to substitute for more expensive rice. It is
here that better organized intra-regional markets could help if some of
the excess production in neighboring countries (perhaps Mali or Niger)
could be made available in Senegal.
One might also consider some type
of targeted program to assure that poor households in the zones most
likely to realize strong negative impacts (urban areas and the northern
Peanut Basin) do not have to support the full impact of the devaluation.
Our data suggest that poor households in the north, for example, rely
most heavily on crop production and livestock for their incomes. We
also found that poor households ate a much larger share of cowpeas than
wealthy households following the poor 1988 harvest. These facts suggest
that programs to increase cowpea production and marketing systems in the
northern Peanut Basin could be a significant help to poor farmers.
Cowpea production could improve caloric intake, provide a reliable cash
crop, and also provide high quality feed supplements for livestock in
the form of cowpea hay.
IFPRI/ISRA data show that livestock production is frequently the
second most important source of income for poor households in rural
areas.
Policy interventions
to make livestock enterprises more
profitable for poor households could also protect them from the negative
effects of devaluation -- credit for fattening activities and better
market information systems are two interventions that come to mind.
Timely investments in infrastructure could also improve production and
marketing response.
At the beginning of the ISRA/IFPRI survey one of
our study zones was just 20 minutes from a paved road; by the end of the
218
survey the road had deteriorated so much that it took about an hour to
get to the study zone.
Water is another problem. Much time that could
be devoted to crop production is now allocated to transporting water due
to inadequate well systems.
The above is clearly not an exhaustive list of possible policy
interventions but merely suggestive of the types of measures that could
be undertaken to protect 'at risk' groups from the negative consequences
of a devaluation and encourage producers to realize the full benefits of
better nominal prices for some of their products.
,..
.
.
I .
..
;
I
219
REFERENCES
Fall, A.A., V. Kelly, and T. Reardon. 1989. Household Survey Methods
used in the IFPRI/ISRA study of consumption and supply impacts of
aqricultural price policies in Seneqal.
Document 2, IFPRI,
Washington, D.C.
Timmer, C.P. 1981.
"Developing a Food Strategy," in Proceedinqs of the
Conference on Food Security in a Hunqrv World, International Food
Policy Conference, San Francisco.
.’
‘.
:
.
I
’
220
SUMMARY OF PRINCIPAL RESULTS AND POLICY IMPLICATIONS
This report has served as a vehicle for presenting the most salient
characteristics of income (levels, composition, and sources), crop
production, agricultural product and livestock transactions (purchases,
sales, gifts, and prices), and food consumption for rural and urban
households in the Peanut Basin and Senegal Oriental during the period
October 1988 through July 1991.
The data were collected to permit
econometric modeling of household demand and supply for agricultural
products, particularly cereals and coarse grains. Although the modeling
work is still under way, there are a number of important findings that
are unlikely to change during the course of future analysis.
These
findings are summarized below.
A discussion of the insights of these
findings for agricultural price policies follows.
The next section
describes patterns of income diversification and their implications for
rural development policies.
A discussion of the policy implications can
be found at the end of the summary.
PRINCIPAL RESULTS
Food consumption
Adequacy:
Average daily caloric intake per adult equivalent
(AE)' for each zone was generally below recommended levels (3000
kcal) but close to minimum needs of 2400 kca1.l Exceptions were
' Adult equivalents are calculated by weighing different persons
in the household by a coefficient representing their relative demands
in consumption.
For example, an adult male will have a much larger
weight than a 5 year old boy.
FA0 adult equivalent standards were
used.
Many results are standardized and reported in adult equivalent
terms (income per AE, for example) to facilitate comparisons across
households of different sizes.
' Minimum daily needs of 2400 kcal/AE are equivalent to 80
percent of the 3000 kcal/AE norms recommended by the Organisme de
Recherches sur 1'Alimentation et la Nutrition Africaine (ORANA). When ,
the average intake per AE falls below 2400 kcal/day, the group is
considered to be "at risk".
IFPRI/ISRA estimates of caloric intake
include all cereals, pulses, tubers, milk, vegetable oil and bread.
encountered in the central Peanut Basin zone of Niakhar (2200 kcal),
in Senegal Oriental (2100 kcal), and in the town of Kaolack (2100
kcal ) .
Given that products not included in the analysis (see footnote 2)
were consumed in very small quantities in Niakhar and Senegal
Oriental, it is unlikely that average intake for these zones would
have reached 2400 kcal if all products were considered. Diets being
much more diversified in urban areas means that average consumption in
Kaolack may have approached, but probably not surpassed, the 2400 kcal
level if all products had been considered.
Although these zone averages serve as a vehicle for making
interzone and interannual comparisons, they tend to mask important
differences among households within the same zone.
About half of all rural households in the sample fell into the
"at risk" category consuming less than their minimum needs of 2400
kcal in both years (59 percent in 1988/89 and 47 percent in 1989/90).
Approximately 25 percent of the entire sample averaged less than 2000
kcal/AE.
The southeastern Peanut Basin had the most households falling
below 2400 kcal in both years (71 percent in 1988/89 and 60 percent in
1989/90), while the southwestern Peanut Basin had the fewest in
1988/89 (44 percent) and the Colobane zone of the central Peanut Basin
the fewest in 1989/90 (32 percent).
In the urban zones, half of the
households interviewed had average annual caloric levels lower than
2400 kcal/day/AE, illustrating that undernutrition is both a rural and
an urban phenomenon.
Composition:
The important role of imported rice has long been
acknowledged for Dakar but generally considered of little importance
outside the capital city.
The IFPRI/ISRA data now confirm that
imported rice provides more than half of all household consumption of
cereal calories in the secondary towns of Kaolack (62 percent) and
Tambacounda (70 percent), as well as in rural areas of the northern
Peanut Basin (70 percent).
In the other zones sampled in the central
and southwestern Peanut Basin, rice provides from 10 to 33 percent of
cereal calories, with higher consumption following poorer harvests.
Rice provides only 5-9 percent of cereal calories in Senegal Oriental
and the southeastern Peanut Basin.
Millet and sorghum account for 99 percent of coarse grains
consumed by sample households throughout the Peanut Basin.
Maize
consumption is important only in Senegal Oriental (the principal maize
production zone), where it provides about 50 percent of cereal
calories in rural areas and 12 percent in the town of Tambacounda.
Prices for imported rice compared to coarse grains: For coarse
grains to provide the same amount of consumable calories per CFA franc
as imported rice, the retail price of whole grain millet, sorghum, and
Products not included are sugar, fruits, vegetables, meat and fish.
222
maize would need to be below 92 FCFA/kilo.3 There are instances in
the IFPRI/ISRA data when the coarse grain price exceeded this
threshold
-- primarily in the northern Peanut Basin and the Niakhar
zone of the central basin from April through September 1989. Across
all rural zones and years, however, coarse grains generally provided
less expensive calories at prevailing grices than imported rice, even
when processing costs are factored in.
The large share of imported rice consumed in the northern basin
(70 percent of cereal calories) can be explained to some extent by
relative prices of rice and millet.
Price, however, does not appear
to explain the large share (about 30 percent) of rice consumed in the
central Peanut Basin (Colobane) and the southwestern Peanut Basin, nor
the small share of rice (15 percent) consumed in the central Peanut
Basin (Niakhar) which exhibited millet prices higher than those in the
northern basin where substantial amounts of rice were eaten.
This
behavior suggests that factors other than price (dietary preferences
and access to markets where rice is readily available, for example)
play an important role in decisions about cereal consumption made by
households in most rural zones.
In urban areas, the relative importance of price and nonprice
factors is less clear.
During the survey year average prices in both
zones exceed the threshold price (which underestimates somewhat the
urban processing costs).
Average annual millet prices in Kaolack were
115 FCFA/kilo.
Average annual sorghum and maize prices in Tambacounda
were about 96 FCFA/kilo.
Given that rice was generally less expensive
than coarse grains during the entire urban survey, it is not
surprising to find 60 to 70 percent of cereal calories coming from
rice.
This makes it difficult, however, to evaluate the extent to
which an increase in rice price might increase coarse grain
consumption.
Rice consumption per AE in market villages is consistently higher
(1.5 to 8 times greater, depending on the zone) than that in nonmarket
villages.
Consumption patterns in market villages serve as an
indication of the secular evolution in rural consumption patterns
given increasing levels of commercialization. The survey provides
little grounds for believing that the evolution of rural consumption
patterns over time will reverse the current trend toward greater rice
consumption.
3 The 92 FCFA/kilo price factors in the FCFA/kilo cost of
processing manually millet at a village millet mill. If processing
costs are not considered, the threshold price would be about 107
FCFA/kilo.
4 Imported rice consumed by sample households was almost
exclusively Thai brokens.
Only one instance was enumerated of higher
quality rice being eaten (an urban household in Tambacounda).
:
.
_‘,’
:
,I,
.
./
.~
223
Cereal purchases:
Rural households, in general, purchase 20 to
40 percent of all cereals consumed.
Although the share of purchases
in total cereal consumption is lower following the much better 1989
harvest, the drop in all but one zone is only 2-3 percent. This
suggests that household cereal purchasing behavior is heavily
influenced by factors other than household cereal harvests.
Higher
income levels in 1989 (see income section below) might have encouraged
the higher levels of consumption of purchased rice in some households.
Preferences for diversified diets or easy access to markets (which
increases opportunities for off-farm income and reduces transactions
costs) are other factors that may keep purchases high even when
production is adequate.
Policy implications of food consumption patterns: The fact that
a large share of households do not cover minimum caloric needs implies
that some type of policy adjustment is necessary. Because more
households are at nutritional risk in higher potential agricultural
zones, nutritional inadequacy may be linked to problems of income
smoothing rather than total income per se (see discussion on income
diversification below).
Credit or food for work programs should be
considered in such situations.
Imported rice is an important component of the diet in most rural
households.
Any policy that increases relative rice prices has the
potential to negatively affect real incomes of rice consuming
households.
Crop production and marketinq
Production technology:
Farm size varies significantly across
zones with the smallest farms (4-5 hectares) found in the zones with
the lowest cropping incomes per AE (northern basin, Niakhar in the
central basin, and Senegal Oriental).
The correlation between small
size and low cropping income per AE suggests that the smaller farm
sizes are not being compensated for by more fertile soils or higher
yields as is sometimes the case in higher potential zones that have
small farms.
Results from a pre-survey village reconnaissance suggest that
small 'size in the central basin zone of Niakhar is due to a land
constraint, while small size in Senegal Oriental is related to labor
and equipment constraints.
Median farm size in other zones ranges
from 7 hectares in the southwest to 11 in Colobane and 13 in the
southeastern Peanut Basin.
Use of "modern inputs" such as fertilizer, pesticides, and
herbicides is extremely limited,
being found primarily 'in households
with production contracts for cotton or confectionery peanuts.
Fewer
than 10 percent of sample households benefitted from these contracts
in any given year.
Most participating households contracted for a ,
hectare or less.
224
Use of fertilizer has declined dramatically since the price
increased in the mid-1980s, following the reduction and then the
1
elimination of fertilizer subsidies.
Credit constraints also
discourage farmers from using fertilizer and from investing in new
traction equipment.
Peanut seeds account for the largest single input cost;
approximately half the peanut seed comes from household stocks and
half is purchased from neighbors or at weekly markets. Despite the
relatively low levels of modern inputs used, input costs represent 14
to 21 percent of the gross value of all cropping output; this is due
in large part to the seed intensive nature of peanut production.5
"Informal" markets based on labor exchange among neighboring
households with inking payments or reciprocity provide most nonfamily
cropping labor, particularly at harvest time.
"Formal" markets for
day laborers do not appear to be well developed. This observation is
based on the fact that hired labor is seldom used and wages paid for
similar tasks are extremely variable.
Production levels:
Zone averages for cereal sufficiency ratios
(ranging from -14 to 1.03) show that households could have covered
from 2 to 12 months of coarse grain "needs" with home production,
depending on year and zone.6
All zones failed to cover minimum
coarse grain "needs" from their own production in 1988/89. Following
the better than average 1989/90 harvest, three zones (Colobane in the
central basin and both zones in the southern basin) had average
sufficiency levels of 12 months while the remaining zones covered 3 to
9 months.
Producer prices of cereals and peanuts: The government
controlled producer price for peanuts was found to be generally
effective during the official marketing period (December - March) when
most peanuts were sold.
The price during the survey period was highly
subsidized (at least 25 percent) given the prevailing world price for
5 Peanut cultivation is particularly demanding in seed. Farmers
use at least 100 kilos of unshelled peanuts to seed one hectare; a
typical yield is 800 kilos (reproduction rate of 8). For cereals only
4 kilos of millet seed are needed to harvest about 400 kilos
(reproduction rate of 100).
6 The cereal sufficiency ratio is calculated by dividing
calories available per day from home production by a household's
average daily coarse grain needs, estimated from IFPRI/ISRA
consumption data to be 1900 kcal/ae.
Use of 1900 kcal makes allowance
for the fact that all households obtain some calories from imported
rice, pulses and other foods.
225
Senegal's peanut oil exports and the net returns realized after
processing.7
Average annual producer prices for coarse grains were in the 60
to 70 FCFA/kilo range for most zones during the two year period, with
prices at the lower end of the range following the better 1989/90
harvest.
A direct comparison of producer prices is not, however,
informative of the economic incentives carried by those prices unless
it is put in the context of returns to land and/or labor.
Returns to family labor:*
Median returns to an 8-hour workday
in household crop production ranged from 150 to 1500 FCFA/day,
depending on crop and zone, during the relatively good 1989/90 season.
The higher end of these returns compares favorably to the SMIG
(official minimum wage in the formal sector) of 1500,FCFA/day.
Analysis of returns to family labor shows that there is little
incentive for most farmers to increase cereal production given current
technology and price relationships. Cereal production was
significantly more remunerative than peanut production (almost double)
in only one zone -- Niakhar in the central Peanut Basin.' In all
other zones median returns to labor in peanut cultivation were higher
(12 to 500 percent) than those for cereals. In three .zones (the
northern Peanut Basin, Niakhar in the central Peanut Basin, and the
7 The 25 percent figure is the estimated subsidy given the
current overvalued CFA franc.
Work by Badiane and Kinteh (June 1993)
suggests that the subsidy serves in large part to compensate for
farmers' loss due to overvaluation.
* Returns to family labor are calculated for each field as net
returns divided by the number of hours of household labor used on the
field.
Labor inputs are not weighted by AE (AE is an appropriate way
of weighting consumption data but is not indicated for production
analysis).
Net returns are calculated by subtracting costs (imputed
or actual) of all variable inputs from the gross value of crop
production (valued at the average producer price during the harvest
year for each crop/zone combination).
', Numerous explanations for the higher returns to millet in this
zone can be hypothesized:
climatic and pest patterns during the study
period that were more favorable to millet than to peanuts, poor
quality peanut seed, or farming techniques that favor millet over
peanuts.
Further analysis of the IFPRI/ISRA crop production data and
additianal field work are necessary to fully understand why this zone
appears to be different.
I
‘.
:
.
3 .
226
southwestern basin) the difference in mean returns to labor for the
two crops was statistically significant at the .05 level."
Crop marketing:
Gross receipts from peanut sales account for 85
percent or more of receipts from crop sales in each year and zone
except Senegal Oriental where cotton was the most important crop in
1988/89 (86 percent of gross receipts). Zones with the largest share
of sales receipts from cereals were Colobane in the central Peanut
Basin (8 and 12 percent of sales in 1988/89 and 1989/90 respectively),
Dioly in the southeastern Peanut Basin (7 percent both years), and
Passy in the southwestern basin (2 and 6 percent)." These sales were
virtually 100 percent millet.
Although some households in Senegal Oriental had a cereal surplus
in 1988/89, they sold less cereal than deficit households in the
Peanut Basin.
This appears to be due to the zone's isolation from
cereal deficit zones and differences in consumption habits (maize and
sorghum are seldom eaten in the Peanut Basin) that limit extra-zone
demand for Senegal Oriental's maize and sorghum.
Despite the generally low cereal sufficiency ratios, more than 90
percent of all households sold some cereal each year. Given that
approximately 50 percent of sample households fall below the "at risk"
level of caloric intake, it is clear that food insecure households do
sell cereals.
Overall quantities sold per household, however, are usually
small.
More than half of the sample households sold less than 200
kilos per year.
This means that half of sample households are not
even marketing enough cereal to satisfy typical annual demand by one
adult equivalent."
To better understand which types of households are marketing
large quantities of coarse grains analysis of households selling more
lo In the Colobane zone of the central basin, the difference was
significant only at the .15 level.
" According to Commissariat de Securite Alimentaire (CSA)
records, the zone of Colobane had a net cereal deficit during the
entire two-year survey period.
ISRA and CSA personnel are currently
collaborating to determine the source of these differences.
Given the
large share of purchased rice in this zone's cereal consumption, it is
possible that farmers are selling millet and purchasing rice.
'* Coarse grain demand per AE is estimated at 1900 kcal per day
(from consumption data).
Allowing for loss of calories in
transforming whole grain to its consumable equivalent, one kilo of
coarse grain provides about 2800 calories:
(200 kilos * 2800
kca1)/365 days = 1534 kcal per day.
227
than 200 kilos of cereal was conducted. These households are more
likely than others to:
be located in the Peanut Basin
be located in villages without weekly markets
belong to the Wolof or Fulani ethnic group
belong to the highest income tercile
earn more than 40 percent of total income in cropping
activities
have a cereal sufficiency ratio greater than .75
cultivate more than 6 hectares per household and more than 1
hectare per adult equivalent
consume less than 10 percent of their total calories from
imported rice.
The most unanticipated result here is that households in market
villages are not selling more quantities of cereals than those in
villages without markets.
This can be explained by a combination of two factors:
1)
Households living in market villages have more diversified
opportunities and rely on non-agricultural income rather than
cereal sales to supplement cash crop income.
This observation is
supported by the fact that farm size and the share of income from
crops both tend to be lower in market than in non-market
villages.
2)
The relatively dense market infrastructure in the Peanut Basin
provides most non-market village households with easy access to a
weekly market where they can market crops.
This ease of access
is confirmed by the fact that most households had some members
making purchases at nearby weekly markets every week.
As market
infrastructure is much thinner in Senegal Oriental it may be
constraining cereal sales in that zone.
Policy implications of crop production and marketing behavior:
Farmers' production and marketing patterns during the 1988-1990 survey
period suggest that GOS cereal production goals (80 percent self
sufficiency by the year 2000) are unlikely to be met if current price
relationships and marketing policies continue.
The small-scale,
largely non-commercial nature of smallholder participation in coarse
grain supply (unlike the case for peanuts) suggests that attempts to
encourage greater coarse grain consumption by increasing rice prices
may require prohibitively large relative price changes before major
changes are observed in coarse grain supply.
Low levels of cereal
production and marketing suggest that a guaranteed outlet and price
for peanuts and cotton may be diminishing the relative incentive for
cereals, which, unlike the two export crops, have no guaranteed price.
and purchaser.
Thin commercial infrastructure in Senegal Oriental
also appears to be constraining cereal production and marketing in
.’
..
.’
228
that zone.
The relative importance of infrastructure versus other
factors (demand for maize or production technologies, for example)
requires further investigation,
however.
Household Income: Levels, distribution, and sources
The level of income is traditionally considered one of the most
important determinants of a household's consumption behavior.
It can
also influence production behavior, particularly a household's ability
to make capital investments or to purchase costly inputs when credit
markets do not function well.
The share of income earned in nonfarm activities can also
influence consumption and production behavior, yet rural household
studies rarely address the issue of multiple income sources and what
this implies for household response to policy changes. Without the
intensive survey methods used in the IFPRI/ISRA study, there is a
considerable risk of deriving an erroneous impression of the relative
importance of different economic activities in household income
strategies.
For example, assuming equal levels of income, a Senegalese
household earning 30 percent of income from cropping and purchasing 80
percent of cereal needs will have a different reaction to a rice price
increase than a household that earns 80 percent of its income from
cropping and only purchases 10 percent of its cereals.
These two
households could also be expected to respond differently to programs
designed to improve agricultural productivity by decreasing soil
erosion if such programs competed with nonfarm activities for
household labor.
Levels and sources of incomes in rural Senegal differ across
zones due to differences in the natural resource endowment (climate,
soils), the infrastructure (transportation, communication), and the
institutional environment (savings, credit, markets).
Understanding
how both levels and sources differ across zones is an essential first
step to understanding how households are likely to react to various
policy initiatives.
In the IFPRI/ISRA study, "income" means net household income for
one harvest year.
The complete flow of income for all household
members and from all sources is taken into account; this is only
doable in the case of bi-weekly interviewing as was the case for the
IFPRI/ISRA study.
Cropping income is calculated by valuing total
harvest at the annual producer price for the zone. The costs of
variable inputs for all activities are netted out as is depreciation
when relevant.
The household income is standardized and reported as
income per AE to facilitate interhousehold comparisons.
Levels:
Average household income in rural zones ranges from
27000 FCFA ($95) per AE in the central Peanut Basin zone of Niakhar i'n
1988/89 to 72000 FCFA ($250) per AE in the southeastern Peanut Basin
229
during harvest year 1989/90.
Average income for urban households is
higher -- about 105,000 FCFA per AE ($365) in both Kaolack and
Tambacounda in 1990/91.13
Following the poor 1988/89 harvest, income did not cover minimum
needs14 for 50 percent of households in the Niakhar zone of the
central basin.
In the Colobane zone of the central basin, the
southwestern basin and Senegal Oriental, 20 to 30 percent of
households had incomes less than minimum needs.
Following generally better 1989/90 harvests, Niakhar (central
Peanut Basin) and Senegal Oriental still had a substantial share of
households with combined income from all sources less than "minimum
needs" levels (35 and 21 percent, respectively).
Interannual differences in levels of rural income are large.
Following the better 1989/90 harvest, average household incomes (from
all sources) increased by 20 to 70 percent throughout the Peanut
Basin.
The central Peanut Basin zones had both the lowest (Niakhar)
and the highest (Colobane) increases.
Most of the increase came from
better cropping outcomes which increased by 50 to 100 percent between
the two years, but some also came from higher off-farm incomes.
Only Senegal Oriental failed to experience an important increase
in income in 1989/90.
This was due in part to poor rainfall but also
because farmers in the zone boycotted cotton production the second
year.
Distribution across households: The distribution of income
across households within each zone is more skewed than that found in
other Sahelian countries for which comparable data are available.
Gini coefficients measure the extent to which income is concentrated
in a small share of households; the larger the coefficient, the
greater the concentration of income.
Gini coefficients calculated
for each rural zone and year are most often in the .35 to .45 range.
Estimates for Nigeria, Niger and Burkina Faso are -28, .30 and .33
(Matlon, 1979; Reardon,
1989; Hopkins and Reardon 1993). The urban
distribution is more skewed than the rural with Gini coefficients in
the .5 range.
Sources of income:
In urban areas, the commerce and services
sectors dominate.
Commerce accounts for 46 percent of net household
13
The per capita equivalents of the Niakhar and southeastern
Peanut Basin figures cited above are 20,000 and 52,000 FCFA,
respectively ($70 and $180); the urban per capita figure is 82,000
FCFA ($280).
During the survey period US$l.OO was worth approximately
290 FCFA.
l4 Minimum needs is defined as the cost (using zone-specific
1
prices) of 1900 kcal/AE/day of coarse grains plus 20 percent to cover
other essential food and nonfood expenditures.
.
~
,’
.
.’
.
:
230
income in Kaolack and services for 22 percent.
In Tambacounda, the
relative importance of the two sectors is reversed -- the service
sector (which includes civil servants) provides 75 percent while
commerce contributes only 12 percent. The contribution by sector is
fairly constant across income terciles.
In rural areas, cropping activities contributed from 20 to 70
percent of rural household income in 1988/89 and 40 to 80 percent in
3989/90.
The extreme case of only 20 percent cropping income was
found following the locust invasion of the northern Peanut Basin in
1988/89.
Consistently large market shares of cropping income (greater
than 70 percent) were found only in the southeastern Peanut Basin.
Most zones were in the 40 to 50 percent range in 1988/89 and 70 to 80
percent range in 1989/90.
Interannual differences in levels of cropping income are quite
large.
In the central Peanut Basin, levels of cropping income doubled
between 1988/89 and 1989/90.
In the southern Peanut Basin zones, the
level increased by 50 percent.
In Senegal Oriental the level of
cropping income dropped between the two years by about 20 percent (due
largely to the cotton boycott and poor rains in 1989/90).
Net income from livestock transactions was an important share of
total income (13 to 23 percent) in all zones but Niakhar in the
central basin and the southwestern basin (5 percent) following poor
1988/89 harvests.15
The following year, livestock income accounted
for less than 5 percent of income except in the southeastern basin and
Senegal Oriental where it represented about 20 percent. The tendency
for livestock income to vary inversely with cropping income and,
therefore, smooth total income is clearly evident.
Off farm income has a variety of sources; commerce and service
activities account for the largest share in most zones while
handicrafts (northern Peanut Basin) and migratory remittances
(northern basin and Senegal Oriental) tend to be more zone specific.
Findings concerning differences in income diversification patterns
across zones and the policy implications of these differences are
presented following the discussion of price policies.
Insights of consumption, production and income results for
aqricultural price policies
Results from simulation analysis show economy-wide policy reforms
(nominal devaluation, for example) have the potential, in the short-
run, to substantially reduce current real incomes of households in
l5 The extreme reluctance of the Serere households in the
Niakhar zone of the central Peanut Basin to sell their livestock is
evident in the IFPRI/ISRA data.
This zone has the largest livestock
herds, the lowest income levels, and the lowest caloric intake in
1988/89 yet they sold very little livestock to supplement income.
:
.’
;
231
urban areas and those in rural areas (northern Peanut Basin, for
example) that obtain a large share of cereal calories from imported
rice.
The same type of policy reform also has the potential to
increase real income for households that consume little rice and
produce large quantities of peanuts (southeastern Peanut Basin, for
example).
These changes may be quite different once associated changes
occur in factor markets such as labor and capital, most particularly
if accompanied by a liberalization of the rice market that would tend
to reduce any devaluation-induced increase in the price of rice. Even
if the associated changes were allowed to occur, however, substantial
disruption of consumption patterns can be anticipated, in the short-
to-medium run.
Both production and consumption patterns may be
difficult to change in less than one harvest year given the realities
of cropping cycles.
The simulation conducted was based on the anticipated price
changes that would occur if border prices for rice and peanuts rose by
50 percent.
Because concrete numbers on how recommended structural
adjustment measures are likely to affect relative prices are not
available, the simulation analysis uses educated assumptions about how
much a devaluation would be likely to raise both the domestic retail
price of imported rice and the domestic producer price of peanuts.
Assuming that the full impact of the real income effect on both
the consumption and production sides is passed through to consumers,
one would anticipate the following consequences:
Households in the northern basin will be worse off (17
percent drop in real income);
Households in the southwestern basin will have no change in
real income;
Households in the remaining four rural zones will be better
off (1-18 percent).
Using a more probable scenario that permits only half of the
price benefit on the production side to be passed through to
producers, one finds that:
Households in the southeastern Peanut Basin are the only
ones to realize a net increase in real income (5 percent);
Households in the northern basin experience greatest loss
(20 percent) and those in southwest the second greatest loss
(8 percent);
Households in the remaining three zones experience 3 percent
drop in real income.
;
232
Urban households in Kaolack and Tambacounda will clearly be
affected negatively as they are big consumers of imported rice and
they do not produce exportable crops.
These results show the average effect for all households in a
zone.
Some preliminary assessment of the differential impact on rich
and poor households has been conducted using the different consumption
and production patterns observed for these groups.
Results show that
imported rice provides an important share (19 to 39 percent) of total
calories in the diets of the poorest quartile of households in three
zones (the northern Peanut Basin,
Colobane in the central basin, and
the southwestern basin).
This implies that the negative income
effects linked to more expensive rice consumption are likely to affect
poor as well as rich households in rural areas -- rice consumption is
not limited to the wealthy who are better able to survive a drop in
real income.
The purpose of this analysis is not to provide arguments against
reforms in macro-economic policy that are clearly needed for long-run
development.
The analysis provides policy makers with information
about which groups of people are most likely to be negatively affected
by the policy changes in the short run, alerting the GOS to the
advisability of developing countervailing measures to protect the most
at-risk groups from the negative short run effects.
Implementation of
such countervailing measures provides greater potential for the policy
changes to survive and elicit the desired long run macroeconomic
effects.16
Patterns of household income diversification
Knowing what factors are driving households to diversify income
and the long-run economic implications of existing patterns of
diversification is an important prerequisite to designing rural
development policies that will foster growth in rural incomes and
increased agricultural production.
Furthermore, understanding income
diversification strategies of rural households permits one to better
judge household-level response to proposed policy changes. To better
grasp these elements, a typology of income diversification patterns
and a model of factors driving households to diversify were
developed.17
l6 In the past, policies that have increased the price of
imported rice have been met with civil disturbances and subsequent
lowering of the rice price.
The most recent case was in 1988 when the
imported rice price of 160 FCFA/kilo was reduced to 130 FCFA/kilo.
" There is an extensive literature on income diversification in
Asia, but evidence from Nigeria (Matlon), Burkina Faso (Reardon,
Delgado, and Matlon), and Niger (Hopkins and Reardon) suggests that
there are some important differences between Asian and Sahelian
diversification patterns.
The topology and model presented here are
:
:
:
.
‘.
233
Topology of diversification patterns: The topology provides a
framework that facilitates the tasks of (1) diagnosing the underlying
causes of the diversification in each zone, (2) understanding what
factors are motivating production and consumption behavior in each
zone, and (3) anticipating the type of household response that
different policy interventions are likely to elicit. Three major
styles of diversification have been identified:
(1) an outward-oriented
"coping" pattern characterized by:
declining levels of cropping income over time due, in
large part, to a poor, deteriorating natural resource
base;
a dearth of activities with production and consumption
linkages to agriculture
an important share of household income coming from off-
farm activities only in emergency situations following
severe crop shortfalls
consistently low total household income due to poor
cropping outcomes and limited off-farm opportunities;
(2)
an outward-oriented, lonq-run, adaptive pattern characterized by:
declining levels of cropping income over time (as in
the "coping" case) or low levels of cropping income
because technological transformation has not yet taken
place and/or price risk is high due to poorly developed
markets and physical isolation from centers of demand
nonfarm income often earned outside of the zone of
residence;
little reinvestment, if any, of nonfarm income into
cropping activities;
total incomes that are low but generally adequate
(higher than those for the "coping" case)
in the long-run, households may move out of crop
production entirely;
specific to Senegal and based on survey findings; the
conceptualization of the different factors to consider in the topology
and model drew heavily on the analyses of the other West African
research cited above.
.
.
234
(3) an inward-oriented, lonq-run, adaptive pattern, characterized by:
good cropping performance stimulating growth of
production and consumption linkages;
off-farm activities that complement rather than
substitute for cropping income;
income from off-farm activities that is reinvested in
crop production;
the presence of well developed transportation and
market infrastructure
a policy environment that fosters development of factor
markets (particularly credit and labor).
The "copinq" pattern is found in the Niakhar zone of the central
Peanut Basin.
Households in this zone had the lowest average incomes
in the study during both survey years.
Following the extremely poor
1988/89 harvest, off-farm income was 53 percent of total income.
Following the better 1989/90 harvest, cropping income doubled and off-
farm income dropped to only 20 percent of total income. Even with the
doubling of cropping income, however, total income remained low.
An example of the outwardlv-oriented, lonq-run, adaptive pattern
is found in the northern Peanut Basin, where cropping potential is
extremely low due to poor rainfall and degraded soils.
Diversification is based to a large extent on migration and provision
of products and services to Dakar.
Another zone that appears to be moving toward an outwardlv-
oriented strateqy is Senegal Oriental.
In this case, however, the
strategy does not appear to be well adapted to the zone.
Senegal
Oriental has promising cropping and livestock potential given high
average rainfall, less degraded soils than the Peanut Basin, and
abundant supplies of crop and pasture land due to low population
density.
Nevertheless, households earn a relatively low cropping
income ker AE and a high share of "outwardly-oriented" off-farm
income.
There are many indications that market failures and lack of
infrastructure are constraining agricultural productivity and
unnecessarily pushing Senegal Oriental's households into migration (22
percent of household income in 1989/90).
Senegal Oriental clearly has agricultural potential. The policy
issue here is how to design policies that will encourage households to
" The zone had the lowest average cropping income in 1989/90
and the third lowest in 1988/89.
,’
.
.
:
,‘.
.’
235
invest in the intensification of agriculture rather than in outward-
oriented off-farm activities, particularly migration.
An inwardly-oriented adapted strategy is currently found in the
two zones in the southern Peanut Basin and Colobane in the central
basin.
These zones have relatively low levels of migration income.
Much of the off farm income in these three zones is linked to crop
production activities (peanut marketing and processing, cereal
marketing, restaurants).
Animal husbandry is also an important source
of revenue in Colobane and the southeast but not in the southwest.
The challenge in these zones is to provide the incentives
necessary to improve agricultural productivity.
Without growth in
agriculture the impetus for growth in the local non-farm sector will
be lost.
If this happens, the prevailing credit constraints will be
magnified because reinvestment of off-farm income in crop production
will decline.
Determinants of off-farm income shares: Understanding why
diversification off-farm is taking place is important to policy
analysts because households will respond differently if the factors
pushing them to diversify are different. Information about factors
driving diversification can be used to develop nonprice policies that
will speed up and increase market response to price signals.
It is relatively easy to understand the role of climate and
natural resource constraints in pushing households into "outwardly-
oriented" off-farm activities in zones such as the northern basin. It
is more difficult to assess the role of inadequate infrastructure,
credit constraints and other market failures.
Regression analysis was
used to gain a better understanding of the relative importance of
these factors.
Elasticities estimated for these variables suggest that:
*
A 10 percent increase in the cereal sufficiency ratio
decreases the share of off-farm income by 7 percent;
*
A 10 percent increase in land assets (proxied by cultivated
area) decreases the off farm share by 5 percent;
*
A 10 percent increase in livestock assets decreases the off-
farm share by 3 percent;
*
A 10 percent increase in agricultural and food credit
decreases the off-farm share by 1 percent.
The elasticity for the cereal sufficiency ratio is large and
negative, as expected.
Households better able to meet food security
needs from home production have less incentive to seek off-farm
,
236
income, all else being equal." This suggests that programs or
policies which reduce food insecurity (food aid, food banks, or
credit, for example) could encourage households in zones with good
agricultural potential to concentrate more on agricultural production.
The same types of programs in zones with an outwardly-oriented
adaptive pattern of diversification or a coping pattern would be
unlikely to elicit greater agricultural productivity.
The elasticity for land assets is also large and negative as
anticipated.
All else equal, more land permits households to earn
more cropping income by producing the desired mix of cereals and cash
crops; higher cropping income reduces the pressure to move off-farm to
supplement income.
This result is not easy to translate into policy
because it involves land tenure issues that were not covered by the
study.
The implication is, however, that larger farms should be
encouraged if the objective is to reduce unnecessary diversification
that reduces productive efficiency.
Larger herd size also decreases the share of off-farm income
because livestock provide savings that are easily converted to cash
when households need liquidity for food or input purchases, reducing
the need to move off-farm.
Policies to improve animal health and
water supplies (a major constraint to animal husbandry in several
zones) when the carrying capacity permits larger herd sizes would be
most likely to diminish off-farm diversification and promote more
integrated cropping/livestock systems in zones with inwardly-oriented
diversification patterns.
Such policies would be unlikely to improve
crop productivity in zones with coping or outwardly-oriented, adapted
strategies.
Although the size of the elasticity for credit is smaller than
that for other factors,
it is an extremely important variable from a
policy perspective.
The credit elasticity implies that better access
to credit permits households to increase the share of cropping income
in total income (that is, it diminishes the need to seek off-farm
income).
This suggests that improving access to credit might permit
households
to use labor which is now engaged in off-farm activities to
improve farm productivity -- planting trees or building windbreaks,
for example.
In sum, the significant coefficients on three of the above
variables -- cereal sufficiency ratios, livestock holdings, and access
to credit -- provide ample support for the hypothesis that growth in
agricultural productivity is being constrained by liquidity problems
that push farmers into greater income diversification than would be
the case were institutions for credit and savings better developed.
" Cereal sufficiency is defined relative to beginning stocks of
cereals, avoiding problems of reverse causality.
237
Although these regression results provide some insights into the
nature of rural income diversification in the Peanut Basin, the model
did not directly address the links between risk and the share of off
farm income in total income.
Additional work on the risk aspects of
diversification is clearly needed.
POLICY IMPLICATIONS OF THE STUDY
The most important policy contribution of the IFPRI/ISRA study
are the insights it provides on the radical differences in household
consumption and production patterns found across rainfed areas of
Senegal that have long been considered relatively homogeneous for the
purposes of agricultural policy design.
Differences in production and
consumption patterns explain,
to a large extent, why households do not
respond to price policy incentives in anticipated ways.
The major
production and consumption patterns found and the implications of
these findings for policy design are summarized below for each zone.
Implications for hiqh potential areas
In the IFPRI/ISRA study, Senegal Oriental and the southern Peanut
Basin zones are located in high potential areas. These zones have
relatively good rainfall and less degraded soils than the rest of the
Peanut Basin.
The study clearly shows that Senegal Oriental is embarking on a
path of "outwardly-oriented" income diversification. failing to
realize its agricultural potential,
The zone is characterized by very
low returns to agricultural labor, very low participation in cereal
markets, low caloric intake, low total income, high outmigration, and
low reinvestment of migration or other non-cropping incomes in
cropping activities.
Study results suggest that implementation of price policies
(increasing producer prices or decreasing input prices) without
attention to increasing market infrastructure, decreasing
transportation costs, and improving crop production technologies is
unlikely to increase agricultural production in the zone.
Low returns to agricultural labor suggest that labor enhancing
technologies need to be further developed and encouraged with
appropriate extension and marketing programs.
Animal traction, for
example, is much less advanced here than in the Peanut Basin.
The
high cost of modern inputs is also a problem, reducing net returns to
cropping labor and discouraging input use, even when credit is
available.
Although production technologies are constraining, the major
cause of the unrealized potential appears to be physical isolation
that limits inter-regional trade and keeps input costs high. Policies
to reduce transportation and transactions costs associated with inter-
.’
:
:
:
I
.
. ‘.
238
regional trade could reduce the wholesale price of Senegal Oriental's
maize in other zones of Senegal and make it more competitive.
A more
competitive price might not stimulate increased human consumption in
the areas of Senegal that consume millet and rice, but there are other
sources of potential demand (poultry producers, for example) that one
can expect to expand rapidly in the near future.
Relatively thin market infrastructure within the zone also
mitigates against intra-regional cereal marketing, making it difficult
for cereal deficit households to purchase local production from other
farmers in the zone and unnecessarily encouraging consumption of
cereals imported into the zone.
Removal of constraints to the
creation of local markets could contribute to better intra-regional
cereal trade that, in turn, might reduce seasonal fluctuations in
caloric intake.
The southeastern Peanut Basin is the most striking example of a
zone pursuing an "inwardly-oriented adaptive" strategy of income
diversification.
It is a high potential zone that comes closer to
realizing its agricultural potential than Senegal Oriental, but
remains constrained by a number of factors.
Incomes in this zone are
the highest in the IFPRI/ISRA survey, yet caloric intake is frequently
inadequate and returns to family agricultural labor remain lower than
those in the lower potential zones of the central Peanut Basin. There
is a need for improved technologies that enhance labor productivity
and increase yields.
Land clearing and weeding, for example, are
extremely labor intensive in this zone.
Fertility enhancing inputs
(fertilizer, manure, compost) are rarely used.
Results show that animal husbandry is a major source of regular
income in the southeastern Peanut Basin rather than simply a savings
account that is depleted when harvests are poor or ceremonies take
place.
Research on improved agricultural technologies should take
livestock assets into account (better use of manure, for example).
Policies that increase returns to animal husbandry (marketing
information systems and permanent sources of water, for example) would
also increase incomes and overall productivity in this zone.
The major overriding constraint in the southeastern basin,
however, appears to be a lack of liquidity. Households in market
villages are able to overcome liquidity constraints, to some extent,
through off-farm activities.
As a result, households in market
villages have higher total income per capita, generally higher
cropping and livestock incomes, and better levels of caloric intake.
Households in nonmarket villages, conversely, have much lower levels
of participation in the off-farm sector, lower total incomes, and
lower caloric intake.
Two types of policies can help reduce the liquidity constraint in
this zone -- savings and credit programs and better infrastructure
(roads and markets).
.
.
,.
239
Policies that encourage savings and credit are the most direct
way of addressing the liquidity problem.
Such programs should be
designed to increase investment in agriculture. The relatively high
cropping incomes currently realized suggest that the potential for
credit repayment exists.
The low returns to labor and low levels of
modern inputs used suggest that credit could encourage greater
investment in existing technologies (animal traction, fertilizer,
improved seeds) and therefore improve cropping productivity.
Increases in cropping productivity stimulated by credit and
savings programs would also increase off-farm incomes through
consumption and production linkages.
For these linkages to be fully
exploited policies are needed to foster better roads and markets
within the zone.
Although the southeastern basin is better endowed
with markets and roads than Senegal Oriental, this type of
infrastructure is still less developed than in the other areas of the
Peanut Basin.
This is the only Peanut Basin zone where households in
nonmarket villages need to make relatively long trips over difficult
roads to get to market villages.
This diminishes the incentives to
engage in off-farm activities that thrive in market village settings
due to the consumption and production linkages fostered by the zone's
relatively dynamic agricultural sector.
The southwestern Peanut Basin is also following an "inwardly-
oriented adaptive" diversification strategy. Households in this high
potential zone are characterized by relatively high levels of total
income, fifty percent of which is earned in off-farm activities.
Many
of the off-farm activities are directly linked to agricultural
production (cereal and peanut marketing, in particular).
Returns to family labor used in cereal production are among the
highest in the study, but even higher returns (by 50 percent) are
realized in peanut production.
Average levels of caloric intake are among the highest in the
study.
The share of imported rice in cereal consumption is also
relatively high but typical of areas with important shares of off-farm
income.
Unlike the southeastern basin, the southwest benefits from
relatively dense market and road infrastructure, although the latter
deteriorated significantly during the survey period.
Given the extremely large difference in returns to labor for
peanut and cereal production, producer cereal price increases would
have to be substantial to elicit even small increases in cereal
production.
It is even conceivable that an increase in the consumer
price of imported rice might have a greater impact on cereal
production than a change in the producer price of coarse grains given
the relatively high rice consumption in this zone.
\\
240
Part, but not all, of the difference in returns to labor for
different crops is due to the production of confectionery peanuts and
the accompanying fertilizer credit. The extent to which the crop
(confectionery peanuts) or the fertilizer enhances returns to labor
needs to be more thoroughly examined and the implications for input
and commodity policies reviewed.
The unusually large share of off-farm income in this zone of high
returns to cropping labor reflects both the liquidity constraint found
in the southeastern basin and a land constraint that is beginning to
make itself felt.
Credit programs are also indicated to assure that
capital is available to invest in land enhancing technologies and that
labor which could be used to improve land productivity during the dry
season (composting, tree planting, and windbreaks, for example) is not
needlessly allocated to noncropping activities.
Despite the high reliance on off-farm income in this zone, the
IFPRI/ISRA data suggest that most of the off-farm income is dependent
on the level of productivity in the agricultural sector.
For many
households, a generalized drop in crop production in the zone leads to
a drop in off-farm income.
This correlation highlights the importance
of investing in agricultural research to improve crop and land
productivity thereby encouraging general economic growth in the zone.
Implications for lower agricultural potential areas
The northern and central Peanut Basin zones are considered the
low potential zones in the IFPRI/ISRA study due to less reliable
rainfall and seriously degraded soils.
Among the three study zones in these low potential areas, one has
developed a successful pattern of diversification away from crop
production, another has been able to maintain a relatively good (but
precarious) level of income based on a mix of cropping, livestock, and
off-farm activities, and the third is barely coping.
These zones each
require a different set of policy initiatives to assure food security
and encourage economic growth.
The northern Peanut Basin has adapted well to recurrent droughts
and low cropping productivity by following an "outwardly-oriented,
long-run adaptive" pattern of diversification. Income per capita and
caloric intake levels fall about mid-way in the distribution across
zones.
Share of noncropping income is, however, the highest (about 70
percent) as is consumption of imported rice (about 70 percent of
cereals consumed).
Crop production constraints are severe in this
zone and households are clearly moving into other sectors of the
economy.
The livestock sector is particularly important, both as a
source of income from animal husbandry and as a source of income from
activities linked to livestock production (sale of byproducts,
commerce and transport of animals to markets in Dakar).
~
.,
.
.
:
:.
‘.
241
Because households in this zone have adapted to the adverse
cropping environment, food security is currently less of a problem for
them than it is for some households in higher productivity zones.
Incomes are generally adequate to purchase food and imported rice is
readily available in weekly markets and village shops,.
This situation
could change dramatically, however,
if the consumer price of imported
rice were to increase.
Economic growth in the zone is highly dependent on transportation
and trade links with Dakar where a large share of production from off-
farm activities is sold.
Policies that reduce real incomes in Dakar
(a sharp increase in the price of imported rice, for example) could
reduce demand for livestock products produced in this zone.
Policy initiatives that can encourage further growth in the
sectors now providing a large share of northern basin income are
improved water sources for livestock,
improved livestock marketing
information, and credit for both livestock producers and traders.
The Colobane zone in the central Peanut Basin falls into the
"inwardly-oriented, adaptive" pattern of income diversification at
present.
It is surprisingly well adapted given that cropping
potential is lower here than in the southern basin and more risky.
There is, however, a need to improve incomes in the zone. Average
incomes are in the low to middle range.
The share of off-farm income
is relatively high.
Off-farm income is derived from a mix of
primarily local activities supplemented with some temporary migration.
Caloric intake is the highest in the survey both years despite
cereal production shortfalls.
The zone exhibits the second highest
level of rural rice consumption.
The zone benefits from relatively good road and market
infrastructure that encourages households to sell coarse grains and to
purchase rice to a greater extent than most other zones. Road
quality, however, deteriorated substantially during the survey period.
The highest returns to family labor used in cropping activities
are found in this zone.
Returns in peanut production are higher than
those for cereals (30 percent). Large farm sizes, extensive use of
traction equipment, and light, easily cleared and weeded soils appear
to contribute to the higher returns to family labor. This zone has
the second largest average farm size,
suggesting that quantity (if not
quality) of land is not yet a constraining factor.
The sustainability of these relatively high returns to labor is
an important issue.
No chemical fertilizer was used by any sample
household in this zone during the entire study period. Few fields
received manure.
Yields per hectare are low and soil fertility will
continue to decline if measures are not taken to stop erosion and
,
restore nutrients to the soil.
.I
.
.
*’
.
.
;
242
Efforts to improve cropping technologies in this zone should
concentrate on fertility enhancing and soil preserving objectives.
Livestock herds in the zone are relatively small limiting the
possibilities for expanded use of manure.
Liquidity problems in this zone are resolved to a large extent by
selling livestock.
Following the poor peanut harvest in 1988 a large
share of receipts. from livestock sales were used to replenish supplies
of peanut seed and secure food.
Credit programs could diminish the
need to liquidate livestock, but the issue of capacity to reimburse
cropping credit in this high risk zone is a real one.
This is the
type of situation where food for work programs could be employed
successfully in conjunction with credit and fertility enhancing
efforts to raise overall levels of income in the zone and move toward
more sustainable agricultural production.
Given the high levels of rice consumption, households in the zone
will be particularly vulnerable to increases in consumer rice prices.
The higher returns to labor in peanut production suggest, however,
that a very substantial change in relative prices will have to occur
before the zone increases the share of cereals in total production.
Niakhar in the western part of the central Peanut Basin is the
one survey zone that clearly falls into the "outwardly-oriented,
coping" pattern of diversification.
It is the zone with the most
discouraging economic prognosis.
The zone is characterized by the
lowest total incomes during both survey years. Cropping income is
constrained by low soil fertility and a shortage of land. Returns to
family labor in peanut production are the lowest in the survey, while
returns to cereal production are among the highest.
Diversification
into off-farm activities to supplement cropping incomes is poorly
developed.
The zone exhibited the lowest average levels of caloric intake
following the poor 1988 harvest,
illustrating that off-farm income did
not adequately substitute for crop shortfalls.
Rice consumption in
this zone is relatively low given the sizeable cereal production
shortfall.
Households show a strong preference for millet, even when
cereals need to be purchased.
More land intensive and fertility conserving production
technologies need to be developed for the zone. Explanations need to
be sought for the very low returns to peanuts. Given the high
population density and relatively low involvement in off-farm
activities, this zone could be expected to have the labor necessary
for labor intensive soil enhancement activities (composting, tree
plant, or ridging, for example).
Greater attention also needs to be paid to developing
opportunities for off-farm employment in this zone where few
production and consumption linkages are evident due to the low
agricultural productivity.
Survey results do not provide a clear
:
243
picture of why households in this zone have not developed the same
type of outwardly-oriented adaptive income strategy that was found in
the northern Peanut Basin.
Households in the zone own the largest livestock herds in the
entire sample, yet animals are sold primarily for food purchases and
ceremonies.
Income from sales is seldom invested in cropping or other
economic activities.
Pasture land is limited in the zone itself,
forcing households to graze animals outside of the zone most of the
year.
This situation not only diminishes the benefits to be gained
from manure but also limits a household's ability to sell animals
easily.
This area of the central Peanut Basin would be expected to need
food aid following crop shortfalls to a greater extent than the
northern basin where off-farm activities are better developed.
Food
for work programs to provide food security and encourage labor
intensive land improvements would also be appropriate.
Given the much
higher returns to millet production and the zone's preference for
millet consumption, research to increase millet yields per hectare may
be more indicated than improved peanut yields.
Relatively low levels of rice consumption mean that households in
this zone would suffer less than others from a sharp increase in the
price of imported rice.
The low levels of peanut production and low
returns to peanut labor mean, however, that the zone is also not
likely to benefit from any policy induced increases in producer prices
of peanuts.
Review of general findings of relevance across all zones
Imported rice is consumed in important quantities in many rural
and all urban households in the IFPRI/ISRA study zones; therefore, any
policy that increases the price of imported rice will harm rice
consuming households by reducing their real income.
The potential for price policy to reduce rice consumption in
rural zones does not appear to be strong given that rice was an
important share of cereals consumed even when coarse grains (including
processing costs) were considerably less expensive.
Rice consumption is likely to increase in the future as
commercialization of rural areas advances and nonmarket villages adopt
the same high rice consumption patterns currently exhibited by market
villages.
Farmers are unlikely to increase cereal production given
prevailing price incentives that result in higher returns to labor for
peanuts than cereals in most zones.
The anticipated positive income effect of policies that increase'
the producer price of cash crops will be dampened significantly by the
,’
.’
244
fact that cash crops provide considerably less than half of total
household income.
This dampening will be exacerbated if policies to
increase export prices in domestic currency are accompanied by removal
of current export price subsidies, as is likely to be the case.
Poorly developed infrastructure, marketing institutions, and
credit policies diminish the relative profitability of crop production
in high potential zones, pushing households into migration.
This is
most apparent in Senegal Oriental.
In all zones with reasonable cropping potential, growth in
agricultural productivity is constrained by a lack of liquidity that,
in the absence of reliable institutions for savings and credit, pushes
farm households further into off-farm production than would otherwise
be the case.
Failure to understand income diversification patterns can lead to
erroneous conclusions about the relative well-being (particularly the
food security status) of households in zones with poor agricultural
productivity.
This can lead to inappropriate food aid and food for
work programs.
The generally low levels of per capita income and the large share
of households failing to earn incomes that cover minimum needs,
seriously constrains investment in productivity enhancing agricultural
technologies.
The problem is compounded by poorly functioning savings
and credit markets.
Overall about 30 percent of sample households fell below the
posited "minimum needs level" in 1988/89. This remained true in two
zones following better harvest in 1989/90.
So long as income for a
large share of households fails to cover minimum needs, investment in
agriculture from deferred consumption is unlikely to take place.
There are surely limits beyond which it is not possible for
households to adjust their consumption further downwards without
incurring severe negative impacts on food security.
It is also hard
to see how households could expand their agricultural production,
without specific attention to improving the natural resource and
policy environments in which households make their production and
consumption decisions.
It is clear that rural households remain
extremely vulnerable to both weather shocks and the probability of
further austerity measures affecting their real incomes.
245
REFERENCES CITED
BADIANE, Ousmane and Sambouh Kinteh.
Aqricultural Trade Pessimism in
West AFrican Countries and the Possible Role of Reqional Markets.
Washington, D.C.: IFPRI, forthcoming.
HOPKINS, Jane and Thomas Reardon.
Aqricultural Price Policy Reform
Impacts and Food Aid Tarqetinq in Niqer. Washington, D.C.:
IFPRI,
April 1993,
MATLON, Peter.
Income distribution amonq farmers in northern Niqeria:
Empirical Results and Policy Implications. African Rural Economy
Paper No. 18, Lansing, Michigan: Michigan State University,
1979.
REARDON, Thomas.
“Income diversification in Burkina Faso, unpublished
mimeo". 1989.
REARDON, Thomas; Christopher Delgado and Peter Matlon. "Determinants
and effects of income diversification amongst farm households in
Burkina Faso".
Journal of Development Studies, January 1992.
.I
:
:
:
.
I
‘.
246
LIST OF PRINCIPAL PROJECT DOCUMENTS
Dakono, Astou. 1991.
Comparaison des revenus au capital et au travail
dans les activites aqricoles et non aqricoles au Nord du Bassin
Arachidier.
Internship Report, Project IFPRI/ISRA, Dakar.
Diagana, B. 1991. Quelques caracteristiques tendancelles du modele
de consommation en milieu urban seneqalais. IFPRI/ISRA, April,
Dakar.
Diagana, B., A. Fall, V. Kelly, and T. Reardon. 1990.
Caracteristiques des zones d'etude du projet IFPRI/ISRA:
Resultats des enquetes de reconnaissance des villaqes. IFPRI/ISRA
Project Document 3, October, Washington, D.C.
Fall, A. 1991.
Composition multisectorielle et distribution du revenu
dans le bassin arachidier:
Remuneration du capital et du
travail.
Tenure paper, ISRA.
IFRPI/ISRA Project Document 5,
January, Washington, D.C.
fall, A. and B. Diagana. 1992.
Les activites non aqricoles et la
capacite de financement de la consommation alimentaire par les
menaqes ruraux au Seneqal.
Paper prepared for regional workshop
on: activites informelles non agricoles et securite alimentaire
au Sahel, 25-28 novembre, Prisas, Bamako, Mali.
Fall, A., V. Kelly, and T. Reardon. 1989. Household-Level Survey
Methods Used in the IFPRI/ISRA Study of Cons umption and Supp1.y
Impacts of Aqricultural Price Policies, IFPRI/ISF IA Project
Document 2, September, IFPRI, Washington, D.C.
Kelly, V., B, Diagana, and A. Fall.
1991.
La consommation des
cereals en milieu rural seneqalais. IFPRI/ISRA, July, Dakar.
Kelly, V., J. Duncan, A. Fall, and B. Diagana. 1992. Transforminq
Survey Data in Demand Models.
IFPRI/ISRA/PRISAS-MSU Working
Paper, November, Washington, D.C. and E. Lansing.
Kelly, V. and T. Reardon. 1989.
Samplinq Methods Used in the
IFPRI/ISRA Study of Consumption and Supp1.y Impacts of
Aqricultural Pricinq Policies. June. IFPRI, Washington, D.C.
Kelly, V., T. Reardon, A. Fall, B. diagana, and L. McNeillv. Draft
9
Final Report for the IFPRI/ISRA Study of Consumption&dsly
Impacts of Aqricultural Price Policies. December II392, IFPRI,
Washington, DC, Kelly V., T. Reardon, A. F-..,
'~111. -.
R w,..J..,,..,
nianana , and L.
McNeilly. 1992.
Kelly, V., T. Reardon, A. Fall, B. Diagana, and L. McNeilly. P992.
Food Consumption in the Peanut Basin and Seneqal Oriental:
October 1988 - September 1989.
IFPRI/ISRA Project Document 7,
September, IFPRI, Washington, D.C.
.
.
;..
.:
,
.I’
:
I.
,‘,
.’
247
Kelly, V., T. Reardon, and L. McNeilly. 1991. Sales, Purchases, and
Gifts of Cereals and Pulses in the Peanut Basin and Seneqal
Oriental.
IFPRI/ISRA Project Document 6, May, JFPRI, Washington,
D.C.
Reardon, T., A. Fall, V. Kelly, C. Delqado, P. Matlon, and 0. Badiane.
1992a. ’ Is Income DiversjficationU'aqriculture-led'
in the WASAT?
Survey Evidence and Development Strateqy Implications. Invited
paper presented at the First Biennial International Conference on
African Economic Issues, June, Lome, Togo,.
Reardon, T., V. Kelly, B. Diagana, and A. Fall. 1992b. Potential
Welfare Impacts of Trade Reqime Chanqes on Households in Seneqal:
Focus on Devaluation.
Paper prepared for the seminar on regional
integration of agricultural markets in West Africa.
Saly
Portudal, Senegal, 4-7 December. IFPRI, Washington, D.C.
FINAL REPORT FOR THE IFPRI/ISRA STUDY
OF
CONSUMPTION AND SUPPLY IMPACTS
OF
AGRICULTURAL PRICE POLICIES IN THE PEANUT BASIN
A
N
D
SENEGAL ORIENTAL
BY
Valerie Kelly
Thomas Reardon
A. Abdoulaye Fall
Bocar Diagana
Lisa McNeilly
(USAID Contract No. PIO/T 685-0281-3-70042)
September 1993
International Food Policy
Insti tut Senegalais de
Research Institute (IFPRI)
Recherches Agricoles (ISRA)
.
i
VOLUME I
FINAL REPORT FOR THE IFPRI/ISRA STUDY
OF
CONSUMPTION AND SUPPLY IMPACTS
OF
AGRICULTURAL PRICE POLICIES IN THE PEANUT BASIN
AND
SENEGAL ORIENTAL
BY
Valerie Kelly
Thomas Reardon
A. Abdoulaye Fall
Bocar Diagana
Lisa McNeilly
(USAID Contract No. PIO/T 685-0281-3-70042)
September 1993
International Food Policy
Institut Senegalais de
Research Institute (IFPRI)
Recherches Agricoles (ISRA)
TABLE OF CONTENTS
I .
INTRODUCTION
Background of Study and Policy Context . . . . . . . . . . . .
1
Study Objectives and Research Questions
. . . . . . . . . . .
3
Research Methods . . . . . . . . . . . . . . . . . . . . . . .
5
Analysis Procedures and Conventions
8
.
.
.
.
.
.
.
.
.
.
.
.
.
Organization of the Report . . . . . . . . . . . . . . . . . .
9
2.
CHARACTERISTICS OF AGROCLIMATIC REGIONS AND STUDY ZONES
. . . 1 1
Location and Characteristics of Agroclimatic Regions . . . . . 1 1
Representivity and Characteristics of Study Zones
. . . . . . 16
Summary of Main Points . . . . . . . . . . . . . . . . . . . .
31
3.
SAMPLE CHARACTERISTICS . . . . . . . . . . . . . . . . . . . . 3 4
Introduction
. . . . . . . . . . . . . . . . . . . . . . . . . 3 4
Sample Demographics
. . . . . . . . . . . . . . . . . . . . .
35
Household Demographics . . . . . . . . . . . . . . . . . . . .
35
Socioeconomic Characteristics of the Household . . . . . . . . 40
Relevance of Household Characteristics . . . . . . . . . . . .
52
for Agricultural Price Policies
Summary of Main Points . . . . . . . . . . . . . . . . . . . . 54
4.
DESCRIPTION AND ANALYSIS OF RURAL AND URBAN INCOME
. . . . . . 56
Introduction
. . . . . . . . . . . . . . . . . . . . . . . . . 56
Research Methods . . . . . . . . . . . . . . . . . . . . . . .
57
Levels, Distribution, and Sources of Income
. . . . . . . . .
61
Intersectoral Links of Income Generating Activities
. . . . . 74
in Rural Areas
Importance of the Informal Sector in Urban Zones . . . . . . . 75
Patterns of Income Diversificat i o n . . . . . . . . . . . . . . 7 6
Summary of Main PO ints . . . . . . . . . . . . . . . . . . . .82
5.
CROP PRODUCTION . . . . . . . . . . . . . . . . . . . . . . . 8 5
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 85
Methodology . . . . . . . . . . . . . . . . . . . . . . . . . 8 5
Gross and Net Cropping Incomes . . . . . . . . . . . . . . . . 88
Land Cultivation and Input Use Patterns
. . . . . . . . . . . go
Returns to Family Labor
. . . . . . . . . . . . . . . . . . .
gI
Distribution of Income by Gender . . . . . . . . . . . . . . .
93
Cereal Production Sufficiency Levels . . . . . . . . . . . . .
g4
Summary of Main Points . . . . . . . . . . . . . . . . . . . .
96
i
.
:,
.;
APPENDIXES
VOLUME 2 PART I
4.1
-
Tables of Average Household Income by Sector, Village, Zone,
and Year
5.1
-
Tables of Analysis Results Describing Characteristics of
Crop Production and Income
6.1
-
Tables of Analysis Results for Sales, Purchases, and Gifts
of Agricultural Products
6.2
-
Comments on zones
7.1
-
Frequency Distribution of Net Income from Livestock
Transactions
7.2
-
Tables of Analysis Results Describing Characteristics of
Livestock Sales and Changes in Stocks
7.3
-
Gini-Coefficients of Cattle and Small Ruminant Distribution
from Ndiaye (1993)
7.4
-
Statement of Net Annual Income (in FCFA) Associated with
Small Ruminants for the Typical Household from Ndiaye (1993)
7.5
-
Summary, Conclusions, and Policy Recommendations from Ndiaye
(1993) Analysis of IFPRI/ISRA Data
8.1
-
Detailed Zone Specific Description of Food Consumption
Patterns
8.2
-
Note sur les Resultas de la Transformation du Sorgho et du
Mais
8.3
-
Tables and Figures of Analysis for Food Consumption Patterns
9.1
-
Consumer and Producer Prices
9.2
-
Comments on Quality of Price Data for Specific Products and
Zones
9.3
-
Consumer Price Trends for Cereals and Pulses
iii
.’
:
.
.;
:
.’ ~
..
LIST OF TABLES
Table 2.1 --
Demographic characteristics of agroclimatic..........17
regions and study zones
Table 2.2 --
Infrastructure and institutional coverage............19
of study zones suggested by reconnaissance survey
Table 2.3 --
Relative importance of various crops in 1987 and.....23
1988 as declared by village representatives
during reconnaissance survey
Table 2.4 --
Agricultural constraints, fallow, and animal.........25
traction patterns suggested by the village
reconnaissance survey
Table 2.5 --
Relative importance of animal husbandry and other....26
off-farm activities as suggested by the
reconnaissance survey
Table 2.6 --
Food security and the relative importance of.........28
different coping strategies (to compensate for food
production shortfalls) suggested by the
reconnaissance survey
Table 3.1 --
Characteristics of the rural sample by zone..........36
Table 3.2 --
Characteristics of the urban sample by zone..........37
Table 3.3 --
Age and gender distribution of rural sample..........38
compared with theoretical population
Table 3.4 --
Household characteristics for the rural zones........39
Table 3.5 --
Household characteristics for the urban zones........41
Table 3.6 --
Characteristics of household heads for rural zones...42
Table 3.7 --
Characteristics of household heads in the
urban sample........................ . . . . . . . ..*....... 44
Table 3.8 --
Access to land for persons >= 15 years in rural......47
zones
Table 3.9 --
Livestock assets in rural zones......................48
Table 3.10 --
Standard of living indicators for households in......50
urban sample
iV
,
.
.
.
.’
.
.’
:
.
Table 3.11 --
History of households in the urban zone..............51
Table 4.1
--
Classification of income-generating activities.......60
by sector
Table 4.2
--
Level and distribution of net household income.......62
per AE (FCFA) by zone and year
Table 4.3
--
Frequency distribution of average net household......65
incomes (FCFA/AE) by zone and year for rural zones
Table 4.4
--
Frequency distribution of average net household......66
incomes (FCFA/AE) in urban zones
Table 4.5
--
Net household income per AE by zone, year, and.......67
sector for rural sample
Table 4.6
--
Net household income per AE (FCFA) by town and.......68
sector for urban sample
Table 4.7
--
Average net total income per AE (FCFA) and shares....72
by sector and tercile for each rural zone and year
Table 4.8
--
Average net total income per AE (FCFA) and shares....73
by tercile and sector for each urban zone
Table 5.1
--
Cereal/pulse production sufficiency rations by.......87
zone and harvest year
Table 5.2
--
Comparison of average gross and net cropping.........88
income by crop for 1989/90
Table 5.3
--
Levels of net cropping income and the cost of........89
minimum needs
Table 5.4
--
Median returns to family labor in cropping...........92
activities for 1989/90
Table 5.5
--
Distribution of crop income by gender................93
Table 5.6
--
Average cereal production sufficiency ratios.........94
for 1988/89 and 1989/90
Table 5.7
--
Average kilos of surplus cereals per AE:.............95
1988/89 and 1989/90
Table 6.1
--
Share of sales revenue from sales of home............102
production by zone, harvest year, and
crop purchased
Table 6.2
--
Volume of cereal sales by zone and year..............104
,
w
.
.
.
:
.’
:
Tab1 e 6.3
--
Estimated volume of coarse grain sales by............104
zone and year
Tab1 e 6.4
--
Share of purchases for home consumption by zone,.....105
harvest year, and product
Tab1 e 6.5
--
Volume of cereals purchased by zone, year, and.......106
product
Tab1 e 6.6
--
Share of purchases for home consumption by zone ,..... 107
harvest year, and product
Table 6.7
--
Estimated volume of coarse grain purchases by........109
zone and year
Table 6.8
--
Estimated net vol ume of coarse grain transactions....109
by zone and year (quantity marketed minus
quantity produced)
Table 6.9
--
Share and total value of seed purchases by zone,.....110
year, and product
Table 6.10 --
Share and tota
value of gifts out by zone, year,.. ..llO
and product
Table 6.11 --
Share and total value of gifts received by zone,.....112
year, and product
Table 6.12 --
Share of total cereal production marketed by zone....124
and year
Table 6.13 --
Share of total peanut production marketed by zone....124
and year
Table 6.14 --
Distribution of households by kilos of cereals.......126
marketed in 1988/89 and 1989/90
Table 7.1
--
Summary of net income from animal transactions.......131
Table 8.1
--
Summary of findings on caloric intake levels.........155
and shares
Table 8.2
--
Share of minimum calorie requirements covered........159
by product group, zone, and year
Table 8.3
--
Product composit on of cereals consumed by zone......160
and year
Table 8.4
--
Product composition of pulses consumed by zone.......161
and year
Table 8.5
--
Classification of consumption
patterns...............164
\\
vi
Table 8.6
-- Share of cereals and pulses purchased by zone........164
and year
Table 8.7 --
Share of rice in cereal purchases by zone and year...166
Table 9.1
-- Income patterns......................................206
Table 9.2
-- Policy effects via production output side............207
Table 9.3
-- Production input expenditure patterns and policy.....209
effect
Table 9.4 --
Expenditure patterns.................................210
Table 9.5
-- Policy effects via decreases in real income on.......212
consumption side
Table 9.6
-- Calculation of net effects on real household.........214
1988/89
vii
‘.
.’
:
LIST OF FIGURES
Figure 2.1 --
Agroclimatic and Study Zones . . . . . . . . . . 12
Covered by the IFPRI/ISRA Survey
Figure 6.1 --
Relative Shares of Sales by Gender
113
.
for (B) Millet, (B) Peanuts, (C) Cotton,' ' l
(D) Cowpeas, and (E) Other for Harvest
Years 1988/89 and 1989/90
Figure 6.2 --
Relative Shares of Purchases by Gender . . . . 118
for (A) Millet, (B) Peanuts, (C) Cotton,
(D) Cowpeas, and (E) Other Products for
Harvest Years 1988/89 and 1989/90
Figure 8.1 --
Consumption by Product Group, Zone, and . . . 154
Year
Figure 8.2 --
Cereal and Pulse Consumption by Product, . . . 158
Zone, and Year
Figure 8.3 --
Composition of "Other Food" by Product,
. . , 162
Zone, and Year
Figure 8.4 --
Sources of Cereals and Pulses by Zone . . . . . 165
and Year
Figure 8.5 --
Product Composition of Cereal and Pulse . . . . 168
Consumption Compared for (A) NonMarket and
(B) Market Villages
Figure 8.6 --
Sources of Cereals and Pulses . . . . . . . . . 170
Compared for (A) NonMarket and
(B) Market Villages
Figure 8.7 --
Composition of Cereals and Pulses . . . . . . . 171
Purchased Compared (A) NonMarket and
(B) Market Villages
viii
,
:
:
:
:’
I
‘.
.I
‘.
Figure 8.8 --
Income Effects on Level and Composition . . . 173
of Cereal and Pulse Consumption for two
Harvest Years
Figure 8.9 --
Income Effects on Sources of Cereals . . . . . 174
and Pulses Consumed for Two Harvest Years
Figure 8.10 --
Income Effects on Composition of Cereals . . . 177
and Pulses Purchased
Figure 8.11 --
Seasonality of Cereal and Pulse Consumption . . 179
Figure 8.12 --
Seasonality of Cereal and Pulse Purchases . . . 184
Figure 9.1 --
Consumer Price Trends for Millet/Sorghum . . . 193
Figure 9.2 --
Consumer Price Trends for Maize . . . . . . . . 195
Figure 9.3 --
Consumer Price Trends for Shelled Peanuts . . . 196
Figure 9.4 --
Consumer Price Trends for Cowpeas . . . . . . . 197
ix
.:
.’
.I.
:
.
.
.
:
:
1 .
INTRODUCTION
BACKGROUND OF THE STUDY AND POLICY CONTEXT
In the early 1980s the Government of Senegal (60s) began a
structural adjustment program.
Some of the most important structural
adjustment reforms being considered at that time were the liberalization
of cereal markets, the privatization of agricultural input markets, and
adjustments in exchange rates.
These are reforms that can be expected
to change levels of real income, and prices of inputs, outputs and
consumption goods.
The IFPRI/ISRA
study, Supp7y and Consumption Impacts
of Agricultura7 Price Policies in the Peanut Basin and Senegal Oriental,
was designed to provide a better understanding of how households are
affected by and respond to these types of policy reforms.
In the mid-1980s, the GOS relied heavily on price policies to
elicit desired production and consumption responses.
I n
1 9 8 6
the
official price for both coarse grains and peanuts (the major cash crop
in most of the rainfed zones) was raised from 70 to 90 FCFA/kilo. These
increases in producer price were designed to compensate for increases in
input costs brought about by the elimination of subsidies and sharp
restrictions on access to credit -- measures introduced between 1980 and
1985 in conjunction with Senegalese structural adjustment programs.
A series' of market liberalization policies was introduced during
the second half of the 1980s.
Among the steps taken to liberalize the
cereals market was the replacement of the official coarse grain price
with a designated floor and ceiling price in 1987. The Commissariat de
Securite Alimentaire (CSA) was given the mandate to intervene in cereal
markets to keep prices within the designated ranges.
By 1988, CSA's
role was diminished from one of intervention to one of simple price
monitoring.
Other important changes were the lifting of the ban on
private sector coarse grain trading and the development of a market
information system that provided the public with weekly cereal prices
for major markets.
These actions were designed to reduce the consumer
price of local cereals by lowering marketing costs and increasing
competition.
Such a policy, if successful, would permit the consumer
price to fall without significant reductions in producer prices.
The potential demand augmenting impact of steps taken to lower
consumer prices of coarse grains was,
however, diminished by the GOS's
decision to maintain control of the imported rice sector and to lower
rice prices from 160 to 130 FCFA/kilo following political disturbances
after the 1988 elections.
This action reduced the price differential
between coarse grains and imported rice thereby reducing the consumer's
price incentive to consume local cereals.
Steps were also taken to reduce government intervention in the
peanut sector and permit licensed private traders to participate in ttie
first stages of peanut collection activities.
These actions were to
)
..
.’
.
2
make the exportation of peanuts more profitable by realizing economies
in marketing rather than by reducing producer prices.
As stated in the
NPA, the GOS was seeking policies that would encourage farmers to
maintain levels of peanut production realized in the early 1980s while
increasing coarse grain production and overall farm income.
In brief, the decade of the 1980s witnessed a vast array of policy
initiatives that would significantly change rural incomes (through both
production and consumption effects) and urban incomes (through changes
in consumer prices).
Unfortunately,
at the time the changes were
implemented it was extremely difficult for policy analysts to gauge the
potential impact of their interventions.
One of the most important types of knowledge needed to predict the
likely consumption and supply impacts of these policy changes is an
understanding of how producers and consumers respond to price and income
changes.
To fully analyze the potential impacts of these policy
initiatives one would need to understand rural and urban consumption
patterns for locally produced millet, maize, sorghum, and rice vs.
imported rice and wheat.
One would want to know how these patterns vary
with changes in prices, as well as such nonprice factors as income,
location,
and infrastructures for marketing and processing. In rural
households, one also needs to know how decisions concerning production,
stocking, and marketing of cereal and cash crops respond to the same set
of price and nonprice factors.
Very little is known about Senegalese consumption patterns, except
for short periods in specific zones (Kramer 1984). It is not clear how
the large demand for imported rice, or for that matter any other cereal,
is distributed over the population.
While it is clear that rice and
wheat are widely consumed in Dakar, various limited-extent surveys have
turned up evidence of significant rice consumption in rural areas (e.g.,
Benoit-Cattin 1987).
Moreover the responsiveness of these patterns to
changes in prices and incomes is not known.
In 1984 Kramer pointed out
that "income elasticities are largely unknown for Senegal. The same is
true of price elasticities of demand... there are almost no estimates of
cross-price elasticities for major cereals." Researchers have found the
lack of such estimates a severe constraint when they attempt to model
aggregate response to price policy options in Senegal (see Braverman and
Hammer 1986; and Abt Associates 1985).
More recently, Kite (1992) reviewed the current evidence on food
consumption patterns,
particularly estimates of price and income
elasticities.
He concludes that ' . ..there is very little "hard"
(recent,
or reliable) information about consumption parameters for
Senegalese families, so a quantitative measure of likely impacts of
pricing policy is difficult."
The picture is not much better on the crop production and supply
side.
Although there is substantial information available on crop
output response to various technologies, there remains a lack of "hard"
:
.
.’
.
3
data on how producers respond to changes in prices of both inputs and
outputs.'
Work by Martin (1988) uses a linear programming model to
predict farmer response to changes in relative prices given a set of
input/output relationships for several different levels of technology.
However,
his input/output data were developed in consultation with
knowledgeable individuals (agronomists, extension workers, and so on)
and are not based on household data.
The model also does not take into
account nonagricultural options for generating income that are available
to rural households.
The IFPRI/ISRA study was designed during the above described period
of rapidly changing agricultural policies.
The study was designed as a
partial response to the need for better data and analysis of the
potential impacts that changes in relative prices might have on the
consumption and supply behavior of Senegalese households. The response
is partial because the study covers only the Peanut Basin and Senegal
Oriental.
These agroclimatic regions do, however, cover 85 percent of
the rural population, 60 percent of the total Senegalese population and
90 percent of both peanut and coarse grain production.
The study may also be thought of as a partial response to data and
analysis needs because it provides only two years of household data,
limiting the inferences that can be drawn from the study to the specific
combination of prices and policies that prevailed during the study
period.
In recognizing these limitations, it is also important to view
this project in the broader Senegalese context where most household data
sets cover only one year.
The fact that the IFPRI/ISRA study collected
data on the same households over two years permits both cross-sectional
and inter-annual comparisons of responses to changes in policies.
STUDY OBJECTIVES AND RESEARCH QUESTIONS
The ISRA/IFPRI project was designed to improve the data base for
and the analysis of income generating and consumption behavior in rural
and urban households throughout the Peanut Basin and central Senegal
Oriental.
The emphasis in rural zones is on the strategies employed by
households to secure both income and food over time through a wide
variety of economic activities.
The study puts the agricultural sector
in the context of the overall rural economy by covering all aspects of
household production and consumption behavior rather than limiting
itself to the cropping sector (as is often done in rural household
studies).
The urban areas are studied to gain insights on levels of
income and consumption patterns , important determinants of urban demand
for crops produced by rural households.
By looking in detail at how farmers allocate time and resources
across agricultural and non-agricultural activities, the study provides
information concerning the factors that influence decisions to produce
different kinds of crops
and to undertake various non-cropping
activities.
By looking at expenditures and consumption of home produced
4
products,, the study provides a detailed picture of how households use
their income and the factors that influence consumption behavior.
Relative prices and income levels are but two of the explanatory
factors examined on both the production and consumption side. Others
are size of household, access to markets, access to credit, availability
of transportation services, access to land, livestock opportunities and
livestock holdings, to name but a few.
Survey data are used to describe household behavior prevailing
during the survey period.
The data permit one to examine how selected
policy interventions would be likely to influence behavior in the future
-- particularly crop production patterns, food security strategies, and
real incomes.
The study permits one to address the following questions:
Composition of Grain Consumption
0
Are households in the study regions consuming substantial
quantities of rice and wheat? To what extent have these products
"penetrated" extra-Dakar consumption patterns?
0
Is there substantial effective demand for maize (a crop thought to
have greater productive potential than millet)?
How does this
differ between households which produce maize and those which do
not?
0
How are decisions about the product composition of consumption,
production, and sales linked to the overall resource and cash flow
strategies of rural households? How do millet and sorghum fit in?
maize? peanuts?
cotton?
Source of Products
0
For rural households in these regions, how important is purchased
food in total consumption?
How important are transfers such as
gifts from relatives in Dakar, donor food aid, or gifts from
neighbors?
What effect is this having on the composition of
consumption?
0
Are the purchases and sales mainly taking place in the village
itself, in weekly markets, in regional markets?
Levels of Consumption
0
What is the average level of calories consumed per adult equivalent
per day in each study zone?
0
Are there entire zones, or at least specific groups within zones,
which are at
risk
in terms of food consumption?
What
characteristics identify these households?
., . .
:
‘.
I
.
_’
.;
5
Responsiveness to Prices
l
How sensitive is the consumption of millet in the face of changes
in its price?
to changes in income? in the face of changes in the
prices of rice, maize, or sorghum?
If millet consumption drops in
the wake of a millet price increase, toward what product do
consumers move?
l
How sensitive are production and sales of millet and maize to
changes in their own prices?
to changes in peanut or cotton
prices? to changes in income from noncropping sources and overall
income?
RESEARCH METHODS
The research reported here was a cooperative effort between staff
economists at IFPRI and ISRA. Project design began in 1987 as a follow-
up to a jointly sponsored IFPRI/ISRA regional seminar on cereal
marketing in West Africa.
Project funding was approved by U.S.A.I.D. in
April 1988 and fieldwork began the following June.
Sample Selection
A village reconnaissance was used to obtain a broad understanding
of how variables that normally influence economic behavior differed
within and across agroclimatic regions in the Peanut Basin and Senegal
Oriental.
Knowledge of differences and similarities among the
agroclimatic regions was used to assure that the purposive selection of
study zones and villages would produce as representative a sample as
possible, given resource constraints.
The reconnaissance survey collected data on zone characteristics
such as,
ethnic
composition,
marketing
and
transportation
infrastructure,
and access to various services
(health, credit,
education, extension, water).
The reconnaissance also provided data on
villagers'
perceptions
of their economic well being and income
generating capacity (cereal
sufficiency
levels,
crop production
constraints, degree and nature of income diversification).
The reconnaissance survey used an informal interview procedure, not
a structured questionnaire.
Data were collected in group meetings
attended, at a minimum, by the village chief and a group of elders.
Usually, other residents of the village (both men and women) attended
and participated in the discussions.
Reconnaissance results suggested that five agroclimatic regions
needed to be covered to obtain a
sample that was generally
representative of the entire Peanut Basin and central Senegal Oriental.
The five agroclimatic regions were divided into eight representative
study zones -- six rural and two urban as follows:
0
1
,
.
.’
:
6
Agroclimatic regions
Study Zones
Sample
Villages
1.
Northern Peanut
Sagatta
3
Basin
2.
Central Peanut
Niakhar
3
Basin
Colobane
3
Kaolack (urban)
1 town
3.
Southwest Peanut
Passy
3
Basin
4.
Southeast Peanut
Dioly
3
Basin
5.
Central Senegal
Missirah
3
Oriental
Tambacounda (urban)
1 town
In selecting study zones within each agroclimatic region, both physical
and socio-economic factors were considered to assure that study zones
were representative of the larger agroclimatic region.
Three villages were selected to represent each rural study zone.
One village was always the site of a weekly market, while the remaining
villages had poorer access to market infrastructure.
This
stratification was done to permit analysis of the influence that access
to market infrastructure has on household behavior. The three villages
selected were not more than 50 kilometers from each other, thereby
permitting coverage by a single enumerator. The populations living in
the villages selected had to reflect the full diversity of the study
zone (as revealed by the reconnaissance survey) in terms of economic,
agronomic,
infrastructural,
and socio-cultural characteristics.
In each village 12 households were randomly selected from a list of
all households in the village.
This provided a total sample of 36
households per study zone,
Due to attrition, the number of households
enumerated and analyzed declined in some zones.
In most cases, however,
the number of households did not drop below thirty,'
The urban study zones of Kaolack and Tambacounda were selected
because they are the urban focal points for the majority of rural zones
covered.
Forty households were randomly selected in each town by
' Readers desiring more information on sampling procedures
'
should refer to Kelly and Reardon (1989).
:
7
referring to detailed ma;ps of the two areas prepared for the national
census conducted in 1988.
Survey Methods
The study period covered two harvest years (1 October 1988 through
30 September 1990) for all rural areas but the Northern Peanut Basin
(October 1988 through September 1989). The latter zone was dropped in
the second year so more researcher time could be devoted to the
preparation of the urban study.
The urban study covered one year (1
July 1990 - 30 June 1991).
Data were collected using intensive, repeated, household-level
interviews.
The data collected fall into the following categories:
Household demographics;
Food consumption;
Prices and quantities for all purchases and sales of principal
crops and livestock;
Prices and quantities
for expenditures on services and
products other than principal crops;
Cash transfers;
Household assets;
Net income from all economic activities;
Gross income and input costs for most economic activities;
Full input/output data (including labor) by plot for 1989/90
and 1990/91 crop seasons.
An extensive set of analyses has been completed to date and
reported in numerous project reports, working papers, and conference
papers.
These analyses can be grouped into the following general
categories:
* In addition to the data collected on the zones covered by this
report, we have some supplementary data from both the reconnaissance
and the household surveys for the Central Peanut Basin (Ndioum
Gainthe), the Southeastern Peanut Basin (Mereto), and the Upper
Casamance (Diaobe).
These data can be used in the future as a control
on results obtained for other study zones in the agro-climatic
regions.
:
*’
‘Y’
:
:
.
‘.
:
8
Food consumption: Calculation of average levels of caloric
intake per adult equivalent and product composition of food
consumption, by zone, season, income group, and access to
commercial infrastructure.
Income levels:
Calculation of average levels, share by
sector,
and distribution (Gini coefficient) of income per
capita and per adult equivalent by zone and income tercile;
Income diversification: modeling of the determinants of income
diversification and development of a topology of income
diversification.
Crop production:
Calculation of the value of crop production
per adult equivalent, returns to labor in crop production, and
degree of cereal
self-sufficiency provided by household
production; calculations are differentiated by zone, by level
of commercial infrastructure, and by income category.
0
Crop and livestock transactions:
Calculation of average values and quantities of products
bought and sold per adult equivalent plus average transaction-
derived prices by month for each principal product;
description of frequencies for place of transaction, persons
involved (household status and gender), sources of financing,
uses of revenues, and reasons for transactions.
0
Price trends:
Development of a full consumer price series for
agricultural product purchases during harvest years 1988/89
and 1989/90 based on transaction derived survey prices
supplemented with data from the Commissariat de la Securite
Alimentaire (CSA) and Michigan State University.
0
Price analysis: Simulation of the short run impact that
changes in the relative prices of cereal and cash crops will
have on real household income.
ANALYSIS PROCEDURES AND CONVENTIONS
Analyses to date have been done separately for each zone, no
attempt has been made to generalize results across all zones.
Within each zone, observations are weighted by household size (AE)
and a coefficient to adjust for the intentional over sampling of market
villages required for analyses of the influence of infrastructure on
household behavior.
The results presented for each individual chapter are based on the
data for all households having a complete set of data for that topic.
This means that the results presented in different chapters are not
s
.’
t
.
.’
.:
. -.
‘.
9
based on the exact same set of households.
This procedure was used
because:
(1) Data entry and analyses were performed sequentially, topic by
topic, so that interim reports on selected themes such as food
consumption and crop transactions could be issued.
When the
consumption data were analyzed, for example, it was not yet known
which households would also have a full set of consistent data for
income or for crop transactions.
(2) Given the relatively small sample size in each zone,
it was
considered best to use the maximum number of observations possible
for each type of analysis.
Once all the data were entered and the households with a full set
of data for all topics were identified, a few analyses were rerun with
the reduced set of households.
A comparison of these results with those
obtained in the earlier analyses suggests that the results are not very
sensitive to changes in sample size.
Given time constraints, it was not
feasible to rerun all analyses with the same set of households.
ORGANIZATION OF THE REPORT
The physical and socio-economic characteristics of the agroclimatic
regions and study zones are described in Chapter 2. The characteristics
of sample households are described in Chapter 3.
Chapters 4 through 9 contain the study's principal findings
concerning household income, crop production, crop disposal, livestock
transactions, food consumption, and prices.
Chapter 10 contains a resume of the principal findings and a
discussion of the policy implications.
.,,
:
.
.J
.’
:
10
REFERENCES
Abt Associates, Inc. 1985.
"Senegal agricultural policy analysis,
agricultural
policy
analysis project", April
(submitted to
USAID/Senegal).
Benoit-Cattin, M.
unpublished.
"Explaining rice consumption in the
Senegalese peanut basin", in Delgado, Reardon Matlon (eds.)
Commodity priorities and changing food patterns in West Africa.
mimeo, IFPRI, Washington, D.C.
Braverman, A. and J.S. Hammer. 1986. "Multimarket analysis of
agricultural pricing policies in Senegal", in I. Singh, L. Squire,
J.
Strauss, eds.
Aqricultural households models, Baltimore,
Maryland, USA: Johns Hopkins University Press for the World Bank.
Kelly, V. and T. Reardon. 1989. Samplinq methods used in the IFPRI/ISRA
study of consumption and supply impacts of aqricultural pricinq
policies. June. IFPRI, Washington, D.C.
Kramer, C. 1984.
"Food consumption, food demand in Senegal: Summary of
literature and suggestions for needed research", submitted to
USAID/Senegal.
Kite, Rod. 1992. "Evidence on food consumption patterns and behavior in
Senegal:
Implications
for the food policy dialogue", mimeo,
USAID/Dakar, October.
Martin, F. 1988. Food security and comparative advantaqe in Seneqal: A
micro-macro approach. Ph.D. diss., Michigan State University, E.
Lansing, U.S.A.
.
.
;
..i.
!“.
:
11
2.
CHARACTERISTICS OF AGROCLIMATIC REGIONS
AND STUDY ZONES
This chapter describes the agroclimatic regions and study zones
covered by the IFPRI/ISRA survey.
The information helps the reader to
interpret the report's descriptive and analytical findings in the
context of the physical and socioeconomic characteristics of the study
zones.
The first section describes the characteristics of the agroclimatic
regions while the second section describes the characteristics of the
rural and urban study zones selected to represent each agroclimatic
region.
Information concerning the agroclimatic regions is drawn
primarily from secondary sources.
The description of IFPRI/ISRA study
zones draws primarily on the village reconnaissance surveys and not at
all on the IFPRI/ISRA household surveys. The picture of the study zones
obtained from the reconnaissance data provides a general idea of how the
interviewers' perceive of their zone.
These data permit one to develop
hypothesis about the zones that are later tested with much more detailed
household survey data.
LOCATION AND CHARACTERISTICS OF AGROCLIMATIC REGIONS’
Figure 2.1 is a map that locates each of the agroclimatic zones and
study zones covered by the survey.
Northern Peanut Basin
The Northern Peanut Basin covers the administrative region of Louga
and all its arrondissements, except Darou Mousty and Dahra.
It covers
10,746 square kilometers (urban centers not included) and had a rural
population of 264,136 persons in 1988, and a rural population density of
' All the demographic statistics on the agroclimatic regions come from
preliminary 1988 census data provided by the Bureau National de
Recensement (September 1988).
The area data come from Secretariat
d'Etat a la Decentralisation (1986). The description of soils,
climate, and cropping systems of the regions draws heavily from
l/Atlas du S&-&al Editions Jeune Afrique, 1983, second edition. The
rainfall data come from l/Atlas du Senegal (Jeune Afrique 1983) and Le
Borgne (1988).
.
12
Figure 2.1 -- Agroclimatic and study zones covered by the IFPRI/ISRA
survey
.,’
.
.
.
i
..‘,,
.
:.
.
.’
:
:
.
.
.
.
:
13
25 persons per square kilometers.'
The main ethnic groups are the Wolof
in the west and the Fulani in the east.
The profile of the region is very uniform, with sandy plateaus and
dunes. The soils called dior by the Wolofs are ferric and not leached.
The Northern Peanut Basin is found in the most arid zone of the Sahel.
The area is prone to the harmattan, a hot, dry wind. The average annual
rainfall varies between 300 and 500 mm and rises as one goes from north
to south.
Agriculture is characterized by a traditional cropping system with
a biannual rotation of millet and peanuts.
Toward the middle of the
198Os, there was substantial growth in the production of cowpeas because
its short cycle is well adapted to the low rainfall and short rainy
season.
The stock of agricultural equipment dates from the Programme
Agricole (1960-80). Animal traction (horse and donkey) is common.
Climate and technical factors affect the type and extent of animal
husbandry practiced in the region.
It is mainly pastoral, based on
transhumance (seasonal migration). In the rainy season, the herds stay
in the region around temporary watering holes; in the dry season they
are led south and west where there are pastures, wells, and markets.
This animal husbandry focuses on cattle and is practiced mainly by the
Fulani.
Sedentary animal husbandry is practiced by the Wolof in the
western part of the region.
The infrastructure network is not well developed. A dearth of
correctly functioning wells creates a lack of fresh water.
The medical
and sanitary infrastructure is very insufficient. The dispersion of the
population aggravates poor access to medical care. Louga (the regional
capital), Kebemer,
and Linguere are the administrative and urban
centers.
The road and rail networks are limited to joining
administrative centers.
The embryonic
industry that exists is
concentrated in Louga.
The rural institutions in the region are Societe de Developpement
et de la Vulgarisation Agricole (SODEVA) (extension of cereal production
technologies and administration of the cowpea program) and some private
voluntary organization (PVO) activity such as World Vision's dry-season
vegetable gardening projects.
Central Peanut Basin
The Central Peanut Basin covers a part of the administrative region
of Louga (i.e., the arrondissements of Darou Mousty and Dahra), part of
the administrative region of Thies,
and part of the administrative
region of Diourbel, and partially includes those of Fatick, Kaolack, and
Tambacounda. It covers 33,958 square kilometers and had a population of
' The total population (including urban areas) was 336,570 persons in
1988, comprising 31,811 compounds and 35,357 households.
.
14
2,254,355 persons
in 1988 of which 1,740,233 are rura1,3 with a
population density of 66 persons/square kilometers (of which 51 are
rural).
The main ethnic groups are the Wolof in most of the region, and
the Serer, especially in the administrative regions of Thies and of
Fatick. In the eastern parts of the zones, there are also Fulani,
Toucouleur, and Mandingue.
Apart from the cliffs of Thies in the far west of the study region,
the land is even and flat.
Dior soils are predominant in all the zones
except for the ferric, leached soils in the northern part. The region
is in rainfall isohyets of 500 to 700 mm. The climate is hot and subject
to the harmattan.
Despite ethnic and rainfall variation over the region, the cropping
system everywhere is dominated by millet and peanuts. Millet continues
to be mainly a subsistence crop.
Peanuts are both a food crop and the
main cash crop.
These two are followed by cowpeas and sorghum in the
center, and by maize as one goes east. Cropping practices are relatively
homogeneous,
although there is a modest diversity linked to differences
in ethnic groups, to the degree of land constraint, and to the degree of
integration between cropping and animal husbandry.
The infrastructure is relatively well developed, having benefitted
from the important peanut economy.
The large towns (Kaolack, Thies,
Diourbel) grew with the peanut economy. The main industrial facilities
are based on phosphate mining and on peanut processing (Societe
Electrique et Industrielle du Basl--SEIB--and Societe Nationale de
Commercialisation des Oleagineux du Senegal--SONACOS). These provide
substantial local employment. There is also intense commercial activity,
especially in the Kaolack region, where there are a number of well-
developed permanent and seasonal markets. The road and rail network in
this region is much more dense than in the other study regions.
Many firms and other rural institutions are present in the region
for marketing, or for technical training and organization of producers
(SODEVA, Caisse de Securite Alimentaire, Societe de Developpement des
Fibres Textiles). The marketing of peanuts brings with it a strong
presence of SONACOS, of cooperatives, and of private storage units
(OPS), which manage peanut collection points. Tourism is developing,
mainly on the Petite Cdte, which attracts substantial foreign exchange
and stimulates local industry (particularly cottage manufacturing).
There are also many PVOs with development projects in the region.
Southwest Peanut Basin
This region covers a part of the administrative regions of Fatick
and Kaolack. It covers 6,765 square kilometers and had a population of
3 In 1988, 195,178 compounds, subdivided into 243,585 households were
registered in the census.
1
,’
.;
.
.
:
15
398,396 persons in 1988, and a density of 59 persons/square kilometers.
The main ethnic groups are the Serer and the Wolof.
As one goes from north to south in this region, the soils go from
being ferric/leached to red-ferric and semileached in the lower plateaus
of the northern Saloum.
In the western coastal area, the soils are
salinized and of the mangrove type because of seawater. The sea climate
affects the overall climate of the zone, and the annual average rainfall
is between 800 mm in the north and 1000 mm in the south.
The main crops are millet and peanuts, although there is some crop
diversification into maize and sorghum.
Use of animal traction is
widespread; horse-drawn plows predominate but cattle-driven traction is
more prevalent than in the center and north of the Peanut Basin.
Vegetable gardening is important in the dry season.
In contrast to the North and Central Peanut Basin, there are no
large towns; instead there are a number of semiurban areas where the
administrative infrastructure is located. The delta of the Saloum River
is an important tourist spot.
SODEVA operated a maize program in the
rural community of Keur Samba Gueye.
Southeast Peanut Basin
The region is bordered in the south by the frontier with The
Gambia;
the west-east boundaries go from Nioro (Kaolack) to the
arrondissement of Koumpentoum (Department of Tamba). This is a mainly
rural region with an area of 6,659 square kilometers and a population of
214,829 persons in 1988,
and a density of 32 persons per square
kilometers.
The Wolof are the main ethnic group, followed by the
Fulani.
There are some Manding in the eastern part of the region.
The Southeast Peanut Basin has plateaus with rocky soils. The
climate is typically sudanian, characterized by a hot, dry season and a
rainy season, with average rainfall varying between 800 and 1000 mm per
year.
It is an extensive-agriculture region. Rotation between millet and
peanuts is common, as is fallowing. Maize and sorghum are also
important.
Cotton is grown in the east of the region, which is the
beginning of the cotton zone.
Cattle-drawn traction is common in the west of the region, but in
general horse traction predominates.
Livestock husbandry is important,
especially among the Fulani.
The region has relatively less infrastructure than the Central
Peanut Basin.
However, there are fairly dynamic weekly markets along
the Gambian border. There are no large urban centers. Apart from the
4 The total population (rural and urban) of the region was
574,151 persons in 1988, hence a density of 85 persons/km' (BNR 1988).
16
national highway (Dakar-Tamba) that runs through this region (as does
the rail), the road network is quite poor.
The region was once the beneficiary of the New Lands program
(Terres Neuves), now defunct. The program was designed to stimulate and
organ-ize the colonization of the empty lands in the Koumpentoum zone by
colonists coming from the Central Peanut Basin where there are severe
land constraints.
Societe de Developpement des Fibres Textiles (SODEFITEX) supplies
technical services in this zone, as well as a market outlet for maize
and cotton. It also builds roads and helps form Farmer Associations
(ABP). SODEVA is present in the west of the region.
Central Senegal Oriental
The region covers the departments of Tamba (except for the
arrondissement of Koumpentoum) and of Bakel. Its area is 16,748 square
kilometers and it had low population density (five persons/square
kilometers in 1988) and a small population of 82,342 persons in 1988.
There is great ethnic diversity.
Fulani are in the majority, but there
are also the Mande groups (Manding, Diakhanke, Soninke) as well as
migrants from the Peanut Basin (Wolof and Serer).
The region is composed of plateaus cut by valleys with heavy bush
cover. The average rainfall was 770 mm during 1961-85 with a minimum of
460 mm and a maximum of 1,250.
Agriculture is substantially diversified in the region, with cotton
as an important cash crop, and substantial production of maize and
sorghum.
There is little use of agricultural equipment and animal
traction because the Programme Agricole did not operate in this zone.
There is some irrigated rice and banana production. Vegetable gardening
is important in the dry season. Animal husbandry and migration income
are important in the region.
SODEFITEX is the main development institution in the region. Its
activities are focused on cotton and maize, as well as on road-building
in some remote areas.
The region has relatively little infrastructure
(roads, schools, and so on).
Tambacounda is the principal semiurban
center (a secondary city).
Its distance from urban centers in the west
or center makes transport costs an important problem.
REPRESENTIVITY AND CHARACTERISTICS OF STUDY ZONES
Representivitv of Study Zones
As mentioned in the introduction, study zones were selected to
represent each of the agroclimatic zones covered by the IFPRI/ISRA
study.
Table 2.1 shows the relationship between the IFPRI/ISRA study
zones and the agroclimatic regions that they represent.
The table
1 7
Table 2.1 - Damographic charactaristicr of agroclimatic ragions and study zanas
Uorthrn
Central
Southwattarn
Southaastarn
Cantral
Agroclimatic tagfons’
Paenut Basin
Peanut Basin
Peanut Basin
Peanut Basin
Sanagal Oriantal
Rural CommunitieslCommunes
32
131
29
15
9
Rural Villages
1,606
4,492
1,313
713
454
Urban Communes)Quartiers
3124
131202
4147
II?
1115
Area (sq.km.)
10.748
33,058
6,765
6,659
16,748
Population
345,336
2,269,401
573,985
215,264
124.227
Households
35,935
246,599
59,626
20,841
14,401
Population Density
Total (peoplelsq.km)
32
67
85
32
7
Rural only
26
52
59
31
5
Ethnic composition
Wolof, Fulani
Wolof, Serer, Fulani
Wolof. Sww
Wolof, Fulani
Fulani, Diakhanke,SarakolC, Manding
Study zonas covatad
Sagatta
Niakhat
Colobana
Passy
Urban
Oioly
Miasirah
Urban
Number of Villages Covered by
reconnaissance
24
28
15
27
25
16
Study villagas(towna covarad by household
Sagatta (mkt)
Niakhar (mkt)
Colobane (mkt)
Passy (mktl
Kaolack
Dioly Mandakh mkt
NiaoulelTouba
lambs
survey
Khelcom Peulh
Mboltogne
TeckMar
K.Aly Gueye
T . ThieckenelMDioukoul
Medina Alylhfoukouty
Darou Cisse
Mboyene
Khayane
Ndiathiang
Kouthia
ffousahold rurvay raprarantivity’
Rural Community)Communes
29
26
66
26
1
14
8
1
Rural Villages/Ouartiers
1,606
649
3,047
1,313
35
379
454
15
Area (sq.km)
10,746
4,604
14,833
6,765
2,965
16.748
Rural Population
People
275,335
373,312
958,819
399,277
150,861
144,229
82,342
4,1885
Households
28,181
41,103
102,295
38,141
18,638
13,082
8,585
5,816
Percent People in market
_...:
villages’
14.3
20.5
17.1
13.7
15.1
14.9
Approximata rhara of total rural (or urban)
6
8
21
8
5
3
2
1
population taprasantad by study zone
Source:
The data used in the table come from the Secretarial d’Etat 6 la DBcentralisation 1986 (number of rural communilies, villages, and area), from the BNR (population in 1988) end from the village reconnaissance sm-~~eys of the )FPRt/lSRA project
(ethnic composition).
Notes:
’ Data in this section describe the entie agroclimatic region. Parts of these regions are not covered by the IFPRIIISRA study (see section on study zone representivity below)
’ Results from the IFPRIIISRA reconnaissance surveys suggest that each study zone reflects the general pattern of household behavior for the areas and populations indicated in this sub-section of the table.
’ Market village is defined as a village with commercial infrastructure. Villages where the rural community or arrondissement is located fall into this category, as do all villages that host weekly markets.
1 8
illustrates the relative importance of the population and land areas
represented by the sample households in each study zone.
Sampling was not proportional to population, therefore, it is
important to keep in mind that the behavior of households in the zones
with larger populations and land areas will have a greater impact on
aggregate
consumption
and production
behavior than behavior of
households in less populated areas.
For example, households in the
areas represented by the Colobane sample in the central Peanut Basin
account for 21% of the total Senegalese rural population.
While the
study zone in the southeastern basin and that in central Senegal
Oriental account for only 3% and 2% of the rural population,
respectively.
Characteristics of Rural Study Zones
The following description of the rural study zones is based on the
village reconnaissance done before the household survey began. Hence,
the behavioral points (such as importance of nonagricultural income) are
general points made b.y villagers and merely serve as hypotheses to be
tested usinq the more detailed observations from the household survey.
The description of rural zones is organized according to the
following categories: demographics, infrastructure and institutions,
cropping systems, animal husbandry, nonagricultural activities, and food
situation.
In each category, the different rural zones are compared.
Following the discussion of rural zones is a brief description of the
two urban zones that were added in July 1991.
Demographics
Table 2.1 presented above shows the area, population, and
population density of each study zone,
The most densely populated zones
(>50 persons per square kilometer) are found in the western parts of the
central and southern Peanut Basin represented by the Niakhar and Passy
study zones.
The least densely settled area is central Senegal Oriental
(5 persons per square kilometer).
The Wolof constitute the principal ethnic group in four of the five
agroclimatic regions.
They are the predominant ethnic group in all
rural study zones but the zone of the Niakhar (central basin), which is
predominantly Serer, and in central Senegal Oriental where the Fulani
and Mandingue are more highly represented.
The Fulani are the second
largest group in
all zones dominated by the Wolof, except the
southwestern Peanut Basin where the Serer are in second place.
Institutional Presence
Table 2.2 shows the types of public or private organizations that
distribute inputs, purchase outputs, supply credit, or provide extension
services
to farmers
in the study zones.
Three categories of
institutions are examined: (a) parastatals; (b) village sections (VS)'
:
:
:
.,;
.
,‘,
19
Table 2.2 --
Infrastructure and institutional coverage of study zones suggested by reconnaissance survey results.
Northern
Central
Central
Southwest
southeast
Central
Peanut
Basin
Ba.Sin
Peanut Basin
Peanut
Senegal
Basin
Niakhar
Colobane
Basin
OliEittd
Characteristics
n=16
II=15
n=24
Percent of Villages surveyed in reconnaissance having indicated attribute
Institution5 uresent
. SDR/other
79
85
22
44
68
100
agencies
NGOs
37
31
1village groups (VS)
21 46
29 79
87 14
30 78
40
80
Access to credit
.Credit VS/RDS/NGOh
62
29
20
52
52
100
of which VS
1 2
1 8
7
52
8
1 2
FZD9
54
18
7
7
48
100
Marketiue and diet.
u
Store in village
1 7
44
73
30
36
31
.Markt 5 5km from
33
61
47
56
12
31
village
.C?olIectn point
62
82
67
89
76
50
-< 5km from village
Processsine Center
Mill in village
1 7
25
27
30
16
40
Access to education
and health
. School in village
1 2
43
38
48
1 2
50
>50%’ of children
21
68
13
70
4
57
in school
health care in
4
7
27
22
4
37
village
Water SUDDI~
Problem very
75
40
63
7
72
25
important
Source: IPPRI/ISRA village reconnaissance survey (1988)
(4
The data in this table represent the percentage of villages among all the villages surveyed in the zone that declared the presence of the given
characteristic.
(b)
This line gives the percentage of villages in each zone having declared receipt in 1988 of agricultural credit obtained by a Village Group (VS),
a Rural Development.
Society (RDS) or an NGO
This represents the share of villages having declared receipt of credit obtained by VS.
;;
This represents the share of villages having declared receipt of credit obtained by RDS.
Note that “a” is not always equal to “b+c”
because certain villages received credit by VS et RDS at the same time.
(4
This line shows the percentage of villages surveyed in each zone having declared that more than half of the village’s children go to French-
speaking school.
(0
The question asked was “where do the village residents go for primary health care?”
\\
In many villages, there is a health post that is not used by the villagers because of lack of medicines or personnel. We counted these cases
as ones where access was absent.
”
2 0
which are cooperatives that have legal status; and (c ) non-governmental
organizations (NGOs).
Parastatal activity is ubiquitous in central Senegal Oriental
(Missirah).
It consists of activities by SODEFITEX for production of
maize and cotton.
By contrast, in the zones of the Peanut Basin
parastatals are less present (22 to 85 percent of the villages depending
on the zone).
The cowpea program was the reason a large number of
villages were working with SODEVA in the northern and central basin; the
maize and cotton programs were the reason for many villages working with
parastals in the southeastern Peanut Basin.' The relative dearth (due
to gradual withdrawal over the past decade) of parastatals is noteworthy
in the central Peanut Basin (particularly Colobane) and in the southwest
Peanut Basin (Passy).
More than 70 percent of villages in the central
and southwestern basin and in Senegal Oriental participate in VSs.' One
finds a smaller share (40-45 percent) of villages in the northern and
southeastern Peanut Basin with functional VS.
NGO's are uniformly
present in 15-30 percent of villages across all rural study zones.
Access to credit is generally positively correlated with the
presence of parastatal, NGOs and VSs.
Senegal Oriental has the highest
level of institutional presence and highest access to credit while
Colobane in the central basin has low institutional presence and low
access to credit.
The Niakhar zone in the central basin is, however, an anomaly. It
has substantial institutional presence but a low level of declaration of
credit access.
Among villages saying they received credit, the most
important sources are parastatal institutions, followed by VSs.
Infrastructure
Table 2.2 also shows infrastructure in the study zones. We examined
the availability of the following resources: water, education, medical
care, markets,
supply and/or processing of agricultural products and
other consumption goods.7
In interpreting these institutional and
infrastructural findings, keep in mind that half of the population of
Senegal is located within 200 km of Dakar.'
5 The maize programs existed at the beginning of our survey but was
phased out during the survey period.
6 Only the VSs that declared to have undertaken activity for their
members (sale of peanuts, provision of seeds, etc.) during the
previous or then current growing season (1987/88) are counted. In
Senegal Oriental, APBs organized by SODEFITEX are classed as VSs.
7 A discussion of the content of the concept of infrastructure can be
found in Ahmed and Hossain (1988).
' See Enders (1989) for a more detailed discussion of population
'
density and distribution in Senegal.
I
:
:
I
,’
)
*
,.
1.
:
21
Access to water used for household consumption, for animal
husbandry, and for dry-season gardening is a serious problem in several
zones.
The problem is
especially common in
the northern and
southeastern basin where two-thirds of the villages say that access to
water is a very serious problem.g
The problem is somewhat less common
but still important (40-50 percent of villages) in both zones of the
central Peanut Basin.
Long distances and poor roads aggravate the
problem of getting to water sources and transporting water back to the
village.
In villages where access to water poses a problem, a great
deal of labor is expended to transporting water and is, therefore, not
available for alternative activities.
Some villages declared that
children were not sent to school because they had to transport water.
Access to formal education (government or private school taught in
French)" is considered easier when the village has or is close to a
school. If a school is in a village, the attendance rate is higher, as
expected. The survey showed that Senegal Oriental, the southwestern
basin and Niakhar in the central Peanut Basin have the largest share of
villages with schools."
More than half (57 to 70 percent) of the
villages in these three zones claim an attendance rate of more than 50
percent.
By contrast, the southeastern basin zone has the fewest
villages with schools (only one of eight); in almost all of these
villages (96 percent), respondents said that less than 50 percent of the
children go to French-speaking school. Similar results were obtained for
the northern basin and the Colobane zone of the central basin.
Access to primary medical care is a general problem in the study
zones.
Over all study zones, more than two-thirds of the village
respondents declared that the villages do not have primary health
facilities in the village, and villagers have to travel more than 5 km
to get care.
In this situation of general dearth, Senegal Oriental is
relatively well off; a third of its villages said that they had health
units in the village.
However,
our secondary data and the local
authorities gave us a dimmer view of health facilities in Senegal
Oriental. In this matter the study villages may be better equipped than
the average in the zone.
There is a great diversity of access to marketing and processing
infrastructure for agricultural products across zones.
Less than 50
percent of the surveyed villages have a shop, except in the Colobane
zone of the central basin where three-quarters have one or more shops.
' The question posed to the chief and/or notables was whether water
supply was sufficient; if not sufficient, was water a minor, serious,
or very serious problem.
" Another source of formal education that was not included in the
survey is the Koranic schools where Arabic and the Koran are taught.
" To determine the attendance rate for children, the question posed %
was whether school-attending children represent at least half the
children of school age in the village.
22
Fifty to sixty percent of villages in the southwestern Peanut Basin and
the Niakhar zone of the central basin were less than 5 kilometers from
a weekly market.
The high population density in these zones favors the
denser market coverage.
By contrast, the southeastern basin has many
weekly markets, but they are relatively spread out, and most villages
are more than 5 kilometers from a market.
Access to collection points for peanuts is similar to that for
weekly markets.
More than 80 percent of villages interviewed in the
southwest and in the Niakhar zone of the central basin have fewer than
5 kilometers to travel when selling peanuts.
In Senegal Oriental 50
percent of the villages surveyed had to travel more than 5 kilometers.
Mills for processing coarse grains are most common in Senegal
Oriental (40 percent of villages) and least common in the northern and
southeastern Peanut Basin (16-17 percent of villages surveyed).
Agriculture
Crop Mix
The village respondents' views on the structure of their
agricultural production are presented in Table 2.3.12.
In all the
Peanut Basin, millet and peanuts are the main crops. The reconnaissance
survey suggests that millet dominates in the central and southwestern
basin, while peanuts predominate in the northern and southeastern Peanut
Basin.
In Senegal Oriental maize dominates, followed by millet and cotton.
Cowpeas are important as the third ranked crop in the north and
center of the Peanut Basin, mainly because of the promotion campaign by
SODEVA.
Dry-season cropping is only important in Senegal Oriental where
about 60 percent of villages grow dry season crops.
Fewer than 25
percent of villages in
other zones cultivate dry-season plots.
Vegetable gardening is the only dry-season farming in the Peanut Basin.
In Senegal Oriental irrigated rice and bananas are grown in the dry
season, with the help of SODAGRI and OFADEC.
'* To classify crops, we asked each village chief to identify and
categorize the crops grown in the village in order of importance. The
latter was implicitly based on land dedicated to each and the number
of producers growing it.
.
.’
:
23
Table 2.3 --
Relative importance of various crops in 1987 and 1988 as declared by village
representatives during reconnaissance survey.
Northern
C@lWal
central
southwest
Southea.5t
CC3ltKd
Peanut
BaSill
Basin
Peanut
Peanut
Senegal
Ba.SiU
Niakhar
Colobane
BaSill
BaSiIl
OIifXttal
Characteristics
(N=16)
(N=26)
(N=lS)
(N=24)
(N=Zl)
(N=16)
PriucipaI rainfed crop
peanuts
71
47
22
68
6
millet
29
100
53
78
28
31
sorghum
-
-
cowpeas
maize
4
38
cotton
25
rice
TOTAL
100
100
100
100
100
Secondary rainfed crop
peanuts
21
100
63
78
24
19
millet
67
40
22
64
6
sorghum
7
8
31
cowpeas
12
maize
25
cotton
4
19
rice
-
none
TOTAL
100
100
100
100
100
100
Tertiary rainfed crop
peanuts
4
8
6
millet
-
7
4
sorghum
4
43
33
18
36
25
cowpeas
88
36
63
4
maize
7
52
44
19
cotton
-
4
8
60
rice
11
none
4
11
other
21
TOTAL
100
100
100
100
100
100
Principal dry-season crop
vegetables
7
7
26
8
19
rice
12
banana
12
sorrel
-
none
92
93
93
74
84
67
other
4
-
8
TOTAL
100
100
100
100
100
100
Source:
IFPRI/ISRA village reconnaissance survey (1988).
Remark:
The interpretation of the table is as follows: In the Northern Peanut Basin, 71% of the
villages surveyed declared peanuts as the most important crop and 29% placed millet in
first place. In the same zone 67% ranked millet in second place while 21% placed
peanuts second and 12% placed cowpeas second. Third place is dominated by cowpeas
(88% of the villages).
.*
,
:
.
.:.
.
.
*
‘.
.
.
.
.
24
Crop Production Constraints
Table 2.4 shows the constraints to crop production declared by the
village representatives.13
Land constraints are considered most important in the Niakhar zone
of the central basin (about half of the villages). The next most severe
case is Colobane, still in the central Peanut Basin. These zones are
among the most densely populated. In the other zones, the constraint was
seldom mentioned (only one in 10 declared a problem). These two land-
constrained zones plus the southwestern Peanut Basin had the smallest
share of villages (15-30 percent) claiming to fallow land regularly.
The low prevalence of fallowing in these zones, supports the villagers
claim that a land constraint exists.
Other zones had 50 to 75 percent
of villages claiming to use fallows regularly.
It is surprising how few villages in all zones declared a labor
constraint, despite the size of the migration phenomenon in the Peanut
Basin and in Senegal Oriental. In fact, fewer than 10 percent of the
villages said they were seriously affected by labor constraints. One
possible reason is that the labor that migrated is surplus labor, with
marginal productivity lower than the average productivity, and so these
laborers can leave the rural areas without reducing output. Another
possible explanation is that much of the migration takes place in the
dry season and does not conflict with cropping activities.
A third
hypothesis is that the principal bottleneck is at harvest time when
temporary laborers from Guinea are available.
An agricultural equipment constraint is declared by 10 to 20
percent of villages in all zones but the southeastern Peanut Basin where
the constraint is considered important in 40 percent of villages. The
"Programme Agricole",
which provided equipment credit throughout the
1960-80 period, was not as well developed in the southeastern basin.
Villages in this zone are also not extremely active in the SODEFITEX
cotton program, which is the only other major source of credit for
annual traction equipment.
Horse-drawn traction is declared to be the most widely used
form
of traction, except in Senegal Oriental where donkeys are more common.
Oxen traction predominates in about 15 percent of villages surveyed in
the southern Peanut Basin zones and in Senegal Oriental.
Animal Husbandry and Off-Farm Activities
Table 2.5 provides an extremely simplified picture of the relative
importance of income sources for men and women as revealed in the
l3 Here the question was whether a given constraint exists; if it
does, was it of medium severity or great severity? Despite the "moral
hazard" of this type of question,
it provided broadly indicative
I
information and grist for hypotheses to be tested with the household
survey data.
25
Table 2.4 --
Agricultural constraints, fallow, and animal traction
patterns suggested by the village reconnaissance survey.
Northern
Central
Central
Southwest
Southeast
Central
Peanut
Basin
Basin
Peanut
Peanut
Senegal
Basin
Niakhar
Colobane
Basin
Basin
Oriental
Characteristics
N=16)
(~=26)
(N=15)
(N=24)
(N=Zl)
(~=i6)
Animal traction type:
(Percent of villages declaring category as principal form of traction)
principal'
ox
horse
donkey
18
16
19
96
100
87
82
84
37
TOTAL
4
13
44
100
100
100
100
100
100
Animal traction type:
(Percent of villages declaring category as second most important form of
traction)
secondaryb
Zrse
4
13
11
76
25
donkey
13
1 9
16
31
none
aa
100
67
70
8
44
a
7
TOTAL
100
100
100
100
1 0 0
100
Fallow
76
28
17
ia
50
6 9
(Percent of villages claiming to use fallows to restore soil fertility)
Constraints:
(Percent of villages declaring the constraint to be important)
. Land
a
46
21
4
1 2
12
Equipment
17
20
40
12
Labor
8
4
12
Source:
IFPRI/ISRA village reconnaissance survey (1966)
a)
The question asked was "what is the principal type of animal traction used in the village?"
b)
The question asked was "what is the second most used type of animal traction?"
2 6
Table 2.5 --
Relative importance of animal husbandry and other
off-farm
activities as
suggested by
the
reconnaissance survey
E
Northern
Central
Central
Southwest
Southeastern
Central
Peanut
Basin
Basin
Peanut
Peanut
Senegal
Basin
Niakhar
Colobane
Basin
Basin
Oriental
E
Characteristics
(N=16)
(~=26)
(N=15)
(N=24)
(N=21)
(N=I~)
Sources of Incane'
iPercent of villaqes Declarinq Activity as Principal, Secondary, or
Tertiary)
E
For Hen
Principal activity
crop
crop
Crop prod
crop prod
crop prod
crop
% villages
prod
prod
100
100
100
prod
E
Secndry activity
92
100
hsbndry
hsbndry
hsbndry
9 4
% villages
hsbndry
hsbndry
67
89
52
mig/ext
Tert. activity
62
8 9
Migrat.
mig/Sen
comnerc
2 5
% villages
comnerc
Mig/Sen
33
41
12
hsbndry
E
42
65
2 5
For Women
E
principal activity
crop
crop
crop prod
crop prod
crop prod
crop
% villages
prod
prod
87
100
100
prod
secndry activity
a3
100
commerce
hsbndry
hsbndry
7 5
% villages
hsbndry
hsbndry
27
8 9
4 0
hsbndry
tert activity
4 2
54
hsbndry
conrnerc
other
19
E
% villages
commerce
mig/Sen
2 0
26
8
comerc
8
3 5
12
E
Husbandry
(Percent of villaqes where husbandry is declared
important)
E
. cattle herding
4
25
27
56
24
50
sheep herding
29
68
73
9 3
8 0
9 4
. head herding
29
71
6 0
9 3
7 2
9 4
Cattle fattening
4
11
7
11
0
12
E
sheep fattening
4 6
18
15
26
3 6
--
17
Source:
IFPRI/ISRA village reconnaissance survey (1988).
The reconnaissance results presented here provide a very general picture of zone tendencies.
Detailed
i
enumeration of income for every member of the household study suggests that the actual composition of income
is somewhat different than the picture obtained from the reconnaissance data.
See Chapter 4 for household
survey results.
a)
The declared income sources for men and for women are classed in order of importance.
The figures below each activity title show the proportion of villages surveyed in each zone mentioning the
activity in a given rank.
Example: In the northern Peanut Basin,
92% of the villages surveyed declared that crop prod agriculture is
the principal source of income for men in their village; hence for 8% of the villages the principal source
of income was declared not to be agriculture.
As the second scurce ofincome 62% cited animal husbandry and as the third source of income 42% declared
commerce.
"Mig/Sen" means migration inside Senegal while
"Mis/ext" means migration outside Senegal; "Migrat"
includes both Senegalese and foreign migration.
27
reconnaissance data.
Analysis of household survey data (Chapter 4)
shows that these general perceptions of income sources are not always
confirmed by detailed enumeration of income earned by each household
member. Nevertheless, the reconnaissance data did permit us to identify
zones
where off farm incomes from migration and commerce were
potentially important.
The table 2.5 shows that in all the zones, for men and for women,
rainfed cropping is declared to be the main source of income. (As shown
in Chapter 4, this contrasts with household survey findings in some
zones.) The table also shows the importance of animal husbandry and
fattening activities in each zone.
Cattle raising was declared important in more than 50 percent of
villages in the southwestern basin and in Senegal Oriental, but in fewer
than 25 percent of villages in other zones.
Small ruminants were
important in more than 70 percent of villages in all zones but the
northern basin.
Animal fattening (primarily of small ruminants) was
most common in the northern and southern Peanut Basin.
For men, sales of animals and animal products were declared to be
the main noncropping income sources throughout the Peanut Basin. In
second place is petty commerce (especially in the zones of Sagatta and
Dioly) and migration.
Migration is declared more important than
livestock income in Senegal Oriental.
Migration is also declared an
important income source in the central and southwest basin.
For women, animal husbandry was declared the main noncropping cash
generator in 5 of 6 zones.
Only in the central Peanut Basin zone of
Colobane was petty commerce declared to be more important than animal
husbandry.
Women's migration was a commonly declared source of income
in the central basin zone of Niakhar.
Food Situation
Table 2.6 shows the declarations of the village representatives
concerning the village's general food situation in 1987 and 1988. The
table also shows the relative importance of methods used to redress
deficits.
The data indicate how many months at least half of the
households in the village could feed themselves based only on their 1986
and 1987 harvests.14
This is a rough measure of self-sufficiency.
Three ranks were created based on the declared number of self-
sufficient months: (1) self-sufficient or not very deficient households
can eat from own production for at least nine months of the year; (2)
moderately deficient, from five to eight months; and (3) very deficient,
fewer than five months.
Analysis of the results led to the following
conclusions.
l4 Again, we stress that these results are only broadly indicative and
serve as hypotheses for the household survey data.
-*-
28'
Table 2.6 --
Food security and the relative importance
of different coping strategies (to
compensate food production shortfalls)
suggested by the reconnaissance survey.
Northern
Central
Central
Southwest
Southeast
Central
Peanut
Basin
Basin
Peanut
Peanut
Senegal
Basin
Basin
Basin
Oriental
Niakhar
Colobane
Characteristics'
(~=16)
(~=26)
(N=15)
(N=24)
(N=Zl)
(N=I~)
(Percent of villages surveyed falling into each category)
1987 Harvest
2 9 months
5-6 months
8
68
40
81
50
25
I4 months
21
28
60
15
46
75
71
4
4
4
1988 Havest
19 months
5-8 months
< 4 months
8
71
27
81
38
25
25
25
73
19
54
75
67
4
8
Means to cover deficit
(Percent of villages declaring activity as principal, secondary or tertiary means)
. print. means
hsbndry
hsbndry
hsbndry
hsbndry
hsbndry
hsbndry
67
73
73
67
89
81
. second. means
migrat
migrat
credit
migrat
misc
comnerc
37
11
40
30
26
31
. tert. means
conxnerc
comnerc
comnerc
off-farm
29
7
16
8
Source:
IFPRI/ISRA village reconnaissance survey (1988).
a)
The indicator for degree of self-sufficiency used is the "number of months covered by
the cereal harvest for more than half of the households in the village.
29
The southwest basin and Niakhar in the central basin have the most
villages declaring themselves self-sufficient (81 and 68 percent,
respectively).
In these zones, one notes a large share of villages
ranking millet as their first crop.
These are also zones with a large
share of Serer households.
The food (production) situation is poorest in the northern Peanut
Basin where,
70 percent of the villages declared themselves very
deficient in 1987.
Almost all the other zones were moderately
deficient, with the southeast having about the same number of villages
each in the first and second categories.
A comparison of the 1986 and
1987 declarations indicates that the food (production) situation was
more or less stable. Some downturn was declared in the northern and
southwestern Peanut Basin zones.
Most villages declared that the first means that they used to buy
cereal to redress their deficits was animal sales (especially sheep and
goats).
Sales of poultry and milk were also mentioned frequently. In
second place was the use of migration income, especially in the northern
basin,
in Niakhar,
and in Senegal Oriental.
Earnings from petty
commerce were the second source in Senegal Oriental and the third in the
northern and southern basins.
Consumption credit (from merchants or
family) was important in Colobane (second after animal sales).
Characteristics of Urban Study Zones
Demography
Kaolack, founded in 1853, is located in the heart of the Peanut
Basin, approximately 200 km southeast of Dakar. Forty-eight percent of
Kaolack's 151,000 inhabitants are men and 42% are women, divided into
14,600 concessions and 18,600 households (BNR 1988). Kaolack is also
the capital of the region and is comprised of 21 districts and 5
villages, officially incorporated into the town. The predominant ethnic
groups are the Serer.
Tamba is situated 500 km east of Dakar and is the most important
city in the eastern part of the country.
It is comprised of six
official districts, plus there are six affiliated villages which make up
the commune of Tamba, the administrative capital of the region. It has
a population of 41,900 (11% of the total population of the region) of
which 51% are men and 49% are women.
The population forms 4330
concessions and 5820 households (BNR 1988). The ethnic composition is
diverse,
with Mandh being the predominant group (Manding, Bambara,
etc...), followed by the Fulani group residing in upper Casamance and
Guinee, and finally the Wolof group from the Peanut Basin.
Demographic evolution of the two zones
1955
1965
1976
1988
Kaolack
46600
81600
135473
150961
Tamba
4600
12700
25780
41885
Source: Atlas JA (1983) et BNR (1988).
3 0
Kaolack was three times more populated than Tamba in 1988, while
three decades before, the proportion was ten times greater.
Infrastructure and Economy
The following table presents information which is available for the
two cities in terms of health, education, communication, and leisure
infrastructure.
Note that the two cities are relatively well endowed with schools,
especially at the elementary level.
Access to water for household
consumption appears to be adequate.
For those who do not subscribe to
SONEES (public utility company providing water) have access to public
faucets where one must pay.
At this level, the problem lies in the
number of public faucets and the establishment of a price scale for
water for the local populations.
Almost all of the districts are
covered by one or more paved roads.
Electric power is available to
households near main roads; but in reality equipment is often in a state
of disrepair.
Also, in half of the districts of the two cities, there is either
a small clinic or a medical center.
However, due to dilapidation or
lack of material and/or medication, which even affects the hospitals of
these two cities, the efficient operation of these infrastructures is
diminished.
Infrastructural Capacity in the two Zones
Kaolack
Tamba
Percentage of neighborhoods
having:- elementary school
81
8 3
- junior high school
3 8
3 3
- public water faucet
95
8 3
- clinic
4 7
5 0
- paved roads
57
6 6
- roads with electricity 95
- markets
33
ii;
- mosque/church
100
100
- cinema
14
17
Does the city have:
- a harbor?
yes
II0
- an airport?
yes
yes
- a hotel?
two?
two
- a hospital?
yes
yes
- a high school?
two?
one
Source: IFPRI/ISRA urban survey data (1990).
31
Each district of the two cities has a religious center, generally
a mosque. Kaolack has three cinemas, whereas Tamba only has one for the
entire city.
Caution must be used when evaluating data for markets. In looking
at the table, it would seem that Tamba is better off than Kaolack.
However, Kaolack,
being geographically larger and more populated,
actually has more markets than Tamba, which apart from the central
market only has smaller, less lively markets. Kaolack is known for its
commercial nature; the strength of this sector is due to the peanut
trade and to the proximity to The Gambia.
For the other infrastructure, Kaolack is better endowed than Tamba,
which has been affected by its isolated position.
Both cities have an
airport, but Kaolack, situated on the ocean, has a port (primarily used
for transporting salt products).
The telephone system is now automatic
in the two urban centers and each has regional broadcasting station.
There is much more industrial activity in Kaolack: processing of
peanuts,
salt harvesting and processing, textiles, etc.
In Tamba,
ginning of cotton by SODEFITEX remains the main industrial activity of
the city.
There is also a banking sector in the two cities where the
regional agencies of the principal banks of the country are located.
SUMMARY OF MAIN POINTS
Results presented in this chapter come from the village
reconnaissance
survey
conducted prior
to the household survey.
Reconnaissance data provides a broad picture of how respondents (village
chefs and elders) perceive the physical and socio-economic environment
in which they live.
Reconnaissance data were used to select
representative study zones and villages.
The most important points made in the chapter are:
a The IFPRI/ISRA sample is considered representative of
approximately 50 percent of the Senegalese rural population and 6
percent of the urban population.
o Fifteen to 20 percent of the rural population in the study zones
live in villages with commercial infrastructure (that is, market
villages and seats of the arrondissement or rural community).
a Access to markets is easiest in the central and southwestern
Peanut Basin (50 percent of villages less than 5 kilometers from a
market) and most difficult in the southeastern basin (only 12
percent of villages within 5 kilometers).
l Crop production is considered the principal source of income for
greater than 95 percent of villages in all zones.
.
32
l Cropping productivity is most hindered by land constraints in the
central Peanut Basin and by equipment constraints in
the
southeastern Peanut Basin and the Colobane zone of the central
basin.
l Labor was rarely mentioned as a constraint, but most cases were
in the zones of high migration (northern basin and Senegal
Oriental).
l Adequacy of cereal production
for household needs in the two
years prior to the IFPRI/ISRA study was best in the Niakhar zone of
the central basin and in the southwestern basin; it was poorest in
the northern basin and Senegal Oriental.
l Income from animal husbandry activities was considered the most
frequently used source of money to cover cereal deficits across all
zones.
l The highest rates of school attendance appear to be in the
Niakhar zone of the central basin and the southwestern Peanut
Basin.
l The lowest rates of school attendance appear to be in the
southeastern Peanut Basin and the Colobane zone of the central
basin.
l Access to water is a serious problem in 70 percent of villages in
the northern and southeastern Peanut Basin and in 40 to 50 percent
of villages in the central basin.
l Institutions providing formal credit are most commonly found in
Senegal Oriental and least common in the central Peanut Basin.
The extent to which the reconnaissance survey results are
consistent with data obtained at the household level is examined in the
next chapter.
33
REFERENCES
Bureau National du Recensement.
1988.
Donn6es preliminaires du
recensement demoqraphique.
September.
Jeune Afrique.
1983.
Atlas du Seneqal.
Serie de 1'Atl as Jeune
Afrique Editions, Second edition. Paris.
Le Borgne, Jean. 1988.
La pluviometrie au Seneqal et dans la Gambie.
Ministere Fransais de la Cooperation et ORSTOM. April. Paris
..;
.’
3.
SAMPLE
CHARACTERISTICS
INTRODUCTION
This chapter describes sample and household characteristics for
the six rural and two urban study zones covered by the report.
The
analysis is drawn from the full set of households that participated in
the study during the first year of data collection for each zone --
those interviewed from the outset as well as those brought in as
substitutes for households that dropped out.'
Both sample and household characteristics are generally based on
the situation prevailing when interviews began in each zone (September-
October 1988 for the rural sample and June 1990 for the urban sample).
The only exception is that the number of adult equivalents reported is
the average number present in sample households during the entire survey
period. This was done to adjust for migration and demographic changes.
First the demographic characteristics of the sample are described:
size,
age
and gender distributions,
and ethnic composition. A
description of the typical household in terms of size and structure
follows,
The next section describes socioeconomic characteristics of
the typical sample household -- age, gender, and educational profile of
the household head; children's education, religion, economic activities
(both local and migratory), rural assets, urban living standards, and
urban
settlement patterns.
The final section points out the
implications of selected findings for agricultural price policy.
As might be expected, a sample covering as many agroclimatic
regions and sociocultural situations as this sample does is bound to
exhibit a great deal of diversity among households within each zone as
well as across different zones.
The discussion in this chapter
generalizes rather than reporting the details of this diversity -- a
"typical" household is described.
In doing this, we relied primarily on
median rather than mean values for households within each zone. In
generalizing about the "typical" household across zones, we report the
modal value of the individual zone medians, pointing out zones where
characteristics differ substantially from the 'typical' picture being
presented.
' Sample households were selected randomly from a list of households
developed by project personnel.
The number of households originally s
selected was larger than the required sample size, thereby providing a
number of households available as substitutes for dropouts.
35
SAMPLE DEMOGRAPHICS
The study sample analyzed in this chapter includes 296 households
(226 rural and 70 urban) which contain 3,452 people or 2,636 adult
equivalents (AE).
Each of the six rural and two urban zones is
represented by 35-40 households (see Tables 3.1 and 3.2).
Aqe and Gender Distribution
The average age of the sample population is about 20 years while
the median is 15 (Table 3.1 and Table 3.2). Forty-five to fifty percent
of the population of each zone is considered "active" (between the ages
of 15 and 59, inclusive).*
Forty-five to forty-seven percent of the
population is under 15 while 4-7 percent is 60 years or more. Males
(51-52 percent) outnumber females slightly in most zones. There is no
difference in age or gender distribution between the urban and rural
zones.
The age/sex pyramid for the sample population was compared with
the theoretical population structure for each rural zone using 1986 data
(Table 3.3).3
This type of comparison allows one to verify whether the
population structure of the IFPRI/ISRA sample is similar to the
structure reported by the national statistical service.
There was only
one statistically significant difference between the sample and the
theoretical population.
The difference occurs in the southwestern
Peanut Basin (Passy zone) where only 40 percent (rather than 45-50
percent) of the sample falls in the "active" category.
Ethnic Composition
The ethnic composition of the sample reflects what one would
expect in each zone (see Chapter 2 for a detailed discussion of zone
characteristics) so details are not repeated here. Table 3.1 shows the
percent of the major ethnic groups represented in each zone.
HOUSEHOLD DEMOGRAPHICS
The typical rural household consists of 11 people or 8 adult
equivalents (Table 3.4).
This typical household is a synthesis of
households ranging in size from 2 to 43 people (2 to 36 adult
* We use the ISRA definition of "active" which includes all persons
215 or t60 years of age.
3 The data for the 1986 theoretical population is based on
information contained in Situation Economique 1986, issued by the
Direction de la Statistique, Senegal.
..
.
.:
36
Table 3.1 --
Characteristics of the rural sample by zone
Central Peanut Basin
Northern
Southwest
Southeast
Central
Peanut
Peanut
Peanut
Senegal
Characteristic
Basin
Niakhar Colobane
Basin
Basin
Oriental
Number of
40
37
37
38
36
38
households
Number of
519
378
438
437
469
431
people
Number of AE
394
290
346
325
350
331
Average age
2 0
21
2 2
2 0
19
2 1
Median age
1 5
1 5
1 8
12
15
1 6
Percent
Active'
46
4 5
5 2
40
48
49
< 15
4 7
47
4 2
54
48
47
2 60
7
8
6
6
4
4
Percent male
4 9
52
5 1
5 1
5 2
5 1
Percent female
5 1
48
49
49
48
49
E-t:~~~~ol;mp,~~~~~~~
.:.:.:; . . >:.:.::::.:::.:..,.:.: . . . . . . . . . . ..A.....
Dominant group
Wolof
Serer
Wolof
Serer
Wolof
Diakhanke
Percent
(63)
(100)
(77)
(47)
(86)
(45)
Second group
Fulani
Serer
Wolof
Fulani
Fulani
Percent
(31)
--
(19)
(33)
(13)
(38)
Third group
Toucouleur
Tout.
Fulani
Serer
Other
Percent
( 2)
--
( 3)
( 5)
( 3)
(17)
Source: ISRA/IFPRI project survey.
Notes:
Information presented in this table describe the sample at the beginning of the survey; changes
in household size and composition are taken into consideration in the analyses but not reported
here.
a)
We use the ISRA definition of active worker which is 115 years and
~60
37
Table 3.2 --
Characteristics of the urban sample by zone
Characteristics
Kaolack
Tamba
No. of households
3 5
3 5
Total population of the households
441
339
% active population
(>=15 ans et ~60 ans)
4 9
50
% population less than 15 years
4 6
4 6
% population 60 or more
years
5
4
% males
4 7
53
% females
53
47
No. of adult equivalents
340
260
Average age (years)
21
2 0
Median age
16
16
p~ijr"i"li":i;ic.ii~~:t-li;ic:i'i~~."'$'i~~bn::::I-~r
Bambaraf
Zone
Yol of
S&&e
Peul
Toucoul cur
Handi ngue
D i a k h a n k e
Sarakol e
Kaolack 51 18
9
15
5
2
Tamba 17 1
4 5
3
16
2
14
Source:
ISRA/IFPRI project survey (1990 - 91).
:
;.
.
.
. .
38
Table 3.3 --
Age and gender breakdowns of sample compared with theoretical population
Percent by age group
ZONE
Age
Percent
Percent
Group
Males
Females
Sample
Theoretical
Northern
< 15
24.3
25.2
49.5
46.6
Peanut Basin
15-54
21.1
22.4
43.5
46.3
>55
3.3
3.6
6.9
7.1
Percent
Sample
48.7
51.2
100.0
100.0
by gender
Theoretical
4 9
51
100.0
100.0
Central Basin
NIAKHAR
< 15
26.8
22.4
49.2
46.5
15-54
20.6
22.2
42.8
46.3
>55
4.3
3.7
a.0
7.1
Percent
Sample
51.7
48.3
100.0
100.0
by gender:
Theoretical
49
5 1
100
COLOEANE
< 15
22.6
19.4
42.0
46.5
15-54
24.2
25.6
49.8
46.3
>55
4.1
4.1
a.2
7.1
Percent
Sample
50.9
49.1
100.0
100.0
by gender:
Theoretical
49
5 1
100
Southwestern
< 15
28.3
25.9
54.2
46.5
Peanut Basin
15-54
17.3
20.0
37.3
46.3
>55
5.0
3.5
8.6
7.1
Percent
Sample
50.7
49.3
100.0
100.0
by gender:
Theoretical
4 9
51
100
Southeastern
cl5
25.4
2 3
48.4
46.5
Peanut Basin
15-54
23.9
21.9
45.0
46.3
>55
2.8
3
5.3
7.1
Percent
Sample
52.1
47.9
100
by gender:
Theoretical
50.2
49.8
Central
< 15
23.8
22.9
46.7
46.9
Senegal
15-54
24.0
22.2
46.2
47.2
Oriental
>55
3.1
4.0
7.1
5.9
Percent
Sample
50.9
49.1
100.0
100.0
by gender:
Theoretical
49.8
50.2
100
Source: Compiled from ISRA/IFPRI project survey and Situation Economique, 1986, Direction de
la Statistique, Senegal.
,
:.
,’
:
.,
.
.’
39
Table 3.4 --
Household characteristics for the rural zones
Central Peanut Basin
Northern
Southwest
Southeast
Central
Household
Peanut
Niakhar Colobane
Peanut
Peanut
Senegal
Characteristic
Basin
Basin
Basin
Oriental
Number of people
Average
13
10
12
12
13
11
Median
12
10
11
11
11
11
Nunber of AE
Average
10
a
9
9
10
9
Median
9
a
a
a
a
9
Composition'
Percent
Nuclear
61
60
60
61
54
56
Extended
38
40
36
39
31
39
Other
1
--
4
v-h
15
5
Education
Percent of HH with
children attending
French school
n.a.
68
26
71
0
52
Migration
Percent
Ory Season
All migrants
13
10
13
4
5
7
Navetane as percent
of all migrants
5
0
25
m-b
aa
38
Rainy Season
4
5
1
1
2
7
Source:
ISRA/IFPRI project survey.
Notes:
Information presented in this table describes the sample at the beginning of the survey; changes in
household size and composition are taken into consideration in the analyses but not reported here.
1
Nuclear family includes: household head, wives, children.
Extended family includes others related to the
household head or his wives (siblings, parents, grandparents, aunts, uncles, nieces, nephews,
grandchildren).
Others includes all not falling in the above categories; they are primarily navetane
(agricultural laborers), apprentices, and Koranic students.
b
This zone does have navetanes, however, none were present at the beginning of the survey so they were not
enumerated until they returned to the sample households in 1969.
Approximately 20 percent of unrelated
household members were navetanes in 1989.
..’
_,
.
:
*
.
.
*.
.
i
_’
4 0
equivalents).
Households in the Serer part of the central Peanut Basin
(Niakhar) tend to be somewhat smaller than households in other zones.
Urban households are smaller than their rural counterparts (Table
3.5); 8-9 people and 6-8 adult equivalents. Households in Tambacounda
are somewhat smaller than those in Kaolack.
Structure
Sixty to seventy percent of the people in the typical rural
household are members of the nuclear family (household head, wives, and
unmarried children) (Table 3.4).
Thirty to forty percent are members of
the extended family (relatives of the household head or his wives).
Five to ten percent of the people in the typical household are not
related to the household head.
In this latter category we find most
frequently navetanes (agricultural laborers), apprentices, and Koranic
students.
This type of household structure is fairly uniform across all
rural study zones.
In the Southeastern Peanut Basin and Senegal
Oriental, however, the nuclear family accounts for a smaller share of
the household population (54 and 56 percent).
In the southeast the
difference is taken up by the "other" category which is 88 percent
agricultural laborers.
In Senegal Oriental, the share of extended
family is larger than in other zones.
The Tambacounda urban sample exhibits the typical
rural
composition described above (Table 3.5).
The Kaolack urban sample,
however, has the smallest share of nuclear family (47 percent) and the
largest share of extended family (53 percent) found in any sample zone.
SOCIOECONOMIC CHARACTERISTICS OF THE HOUSEHOLD
Profile of the Household Head
The typical household head is a male in his late forties or early
fifties.4
More than 95 percent are married; 50-55 percent have one wife
and 30-40 percent have two wives.
Cases of three and four wives are
rare; these cases are more common in the southeastern Peanut Basin where
about 40 percent have one wife, 30 percent have two, and 20 percent have
three (Table 3.6).
Monogamy is more common in the Serer zone of the central Peanut
Basin (Niakhar study zone) where slightly more than 70 percent of
household heads have only one wife.
Formal education in the French system is rare among household
heads;
only 5 of the 226 rural household heads have attended school,
4 In rural zones less than 1 percent of household heads are female.
.
.
:
41
Table 3.5 --
Household characteristics for the urban zones
Characteristics
Kaolack
Tamba
s&$&#g&l$E~@;:( kii.&&)
‘:i.:...... ..,.:. I,. . . . . . :,.,........., . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . : ,.,.,...,.i,.,..,.,
.
Average population at start of study
12.0
10.0
Median population at end of study
9.0
8.0
Average AE at start of study
9.7
7.4
Median AE at end of study
7.5
5.7
Average annual population
12.0
9.0
Median annual population
6.0
8.0
Average annual AE
9.4
7.0
Median AE
6.5
5.9
~~~iSirjii’,oFi.r~~~&~:~PXt
..,..,.........,...,.,.,,..,
.,.. :..:.:.:,:
. . . . . :. ..;,.,,_,.
.,... .. ,.,
,.>>,.::.:..::.:::
,,........
“‘.....’
Nuclear family
46.5
6 1
Extended family
53.0
36.0
Others (not clear)
0.5
3.0
EnoeAri~.:i:fi’E~~~~~~~~E~,T
. . . . . . . . . ..I.. .
..,.. ..‘.,.. .......,, ,, .,,.:. . . . . . . . . .., .,.;. .., .,,.,
% of total household population who have:
- attended French school
23.0
30.0
- received formal vocational or technical
10.0
1.0
training
- served a traditional apprenticeship
6.0
9.0
- a current source of income
17.0
19.0
);Il&pJzf~
~qJ&jz ~‘:~&&~~~~@j~~
:.... : :.:. :. :::.::..:.:::.:.:..
: ..:.:.+: : ;:. .:.. . . .,.,
Dry season
0.7
1.0
Rainy season
4.0
4.0
Source:
IFPRI/ISRA project survey (1990-91).
42
Table 3.6 --
Characteristics of household heads for rural zones
Central Peanut Basin
Northern
Southwest
Southeast
Household
Peanut
Niakhar
Co1 obane
Peanut
Peanut
Characteristic
Basin
Basin
Basin
Oriental
&++g&.&
.:..
‘A6’e;as’e
.;..,.
47
53
49
51
--
48
Median
45
51
51
52
40
50
h&r
.:..:“‘Phrce”t ma, e
100
97
97
100
100
100
Percent female
0
3
3
0
0
0
..,
&j-&l i::.&$$.js
., . .:: . . . .’ .:.:..: . . . . . .,
Percent
Unmarried
0
a
5
3
0
1 0
One Wife
50
69
47
64
35
59
Two Wives
31
15
37
28
38
21
Three Wives
17
a
5
5
24
10
Four Wives
2
0
5
0
3
0
Number with
formal schooling
~6 years
n.a.
0
1
0
0
1
6-9 years
n.a.
1
0
1
0
0
9-12 years
n.a.
0
0
0
0
1
Number with
literacy training
n.a.
0
3
0
0
3
Source: ISRA/IFPRI project survey.
Notes:
Information presented in this table describes the sample at the beginning of the
survey; changes in household size and composition are taken into consideration in
the analyses but not reported here. n.a. means not available.
43
most for 6 or fewer years.
Literacy programs in local languages have
been no more popular;
only 6 household heads have attended classes.
There are some distinct differences between "typical" rural and
"typical" urban household heads.
Urban zones have a much larger share of female household
heads (26 percent in Kaolack and 14 percent in Tamba) than
rural zones (<l percent) (See Table 3.7).
About 30% of urban households are headed by unmarried
persons (t5 percent in rural zones).
Urban household heads have a greater exposure to formal
education than their rural counterparts, with 20-30 percent
having attended school at some time. Another 20-30 percent
have had some technical or apprenticeship training. This
leaves, however, about 60 percent with no type of education
or formal job training.
Profile of the Household
The rate of formal education among children is generally higher
than that of household heads; however, great disparity exists across
zones.
In two study zones (the Niakhar zone of the central Peanut Basin
and the southwestern Peanut Basin) about 70 percent of households have
at least one child attending school.
Thirty to thirty-five percent of
households are sending children to formal schools in Senegal Oriental
and the Colobane zone of the central Peanut Basin,
In the southeastern
Peanut Basin (Dioly) not a single child in the entire sample is
attending school.
Disaffection for formal education is typical of this
zone and is not a sampling anomaly.
In the urban sample, 20-30 percent of the total population has
some formal schooling (the rate is higher in Tambacounda).
About 40
percent of the sample in each urban zone has had some type of formal
schooling, apprenticeship, or formal job training.
Reliqion
Households are predominantly Muslim in both rural and urban
samples.
Only in the Niakhar zone of the Central Peanut Basin do we
find some Catholic households.
The presence of Catholics in this zone
undoubtedly contributes to the higher rate of monogamy noted above.
Educational infrastructure provided by the Catholic church in the zone
helps to encourage the higher rate of school attendance found in this
area.
.
. .
44
Local Economic Activities
Almost all rural households declared crop production as their
major source of income.
As Chapter 4 reveals, this was not found to be
true in all zones and all years covered by the ISRA/IFPRI survey.
Table 3.7 -- Characteristics of household heads in the urban sample.
Average (years)
52
4 6
Median (years)
51
42
$*:+,::.
% Male
74
86
% Female
26
14
.,
~~yAtp;ciii’$~j$~&:Qi(,X]:::
. . ,.,: . . . . :.:.: . ,.:.:.:.>,.: :.,.:.:.:.:.:.:.:.,.:.: ,,._.,.,., ,.
% Unmarried
31
28
% Married with 1 wife
4 9
54
2 wives
11
11
3 wives
6
3
4 wives
3
3
.,
,. . . .
EnucAT~a#AL.::.sErr~~~,~~~,
: ,.,., ,.:... ,. .):.:.~,.:., ,.. :.: ,., ..:.:.: :,,
- Formal education
Primary school
9
12
Secondary school
3
6
University
6
12
Other
82
70
- Formal and formal vocational or professional training
3
6
- Formal instruction and traditional apprenticeship
9
- Traditional apprenticeship only
24
12
- Formal instruction only
15
14
- None of the above
58
59
Source:
IFPRI/ISRA project survey (1990 - 91).
4 5
Most rural households have several other sources of income in
addition to their crop production; the importance of these activities
varies across households and across zones.
Very few urban households draw income from cropping and livestock
activities.
The most common activities are in the service and commerce
sectors.
Only 17 percent of the Kaolack and 19 percent of the Tambacounda
population declared that they were earning some type of income when the
survey began; this means that just 35-40 percent of the "active"
population considered themselves to be employed at that time.
Miqration
Migration is a phenomenon of varying importance across zones.
The percent of household members 115 years of age leaving the village
during the dry season ranges from 4 to 13 percent. The highest rates
are found in the northern Peanut Basin and the Colobane zone of the
central Peanut Basin.
The lowest rates are in the southwestern and
southeastern Peanut Basin.
Outmigration during the dry season in three zones includes a large
number of 'navetanesn
returning to their own homes (25 percent of
outmigrants in the Colobane zone of the central Peanut Basin, 38 percent
in Senegal Oriental, and 88 percent in the southeastern Peanut Basin).
The remainder of the outmigrants are predominantly males, both married
and single; many are household heads.
In terms of gender, the Niakhar zone of the central Peanut Basin
presents an exception to the "typical" migration profile. Slightly more
than 50 percent of the dry season outmigrants from this zone are
unmarried women going to Dakar to work as maids.
Rainy season
migration is generally lower than that during the
dry season.
It is more common in the northern Peanut Basin and Niakhar
in the Central Peanut Basin where poor rainfall and land constraints
make it increasingly difficult to farm.
Senegal Oriental, with its low population density and relatively
good rainfall, surprisingly has the highest rainy season outmigration.
In this zone there are several households with adult males on long-term
migration to Europe or to other African countries.
These individuals
are considered to be household members in those cases where they have
left spouses and children in the village or if they are single but
regularly send remittances to their parents.
Migration from urban households during the dry season is about 1
percent of total population; this increases to 4 percent during the
rainy season when some individuals return to rural areas to participate
in agricultural production activities.
46
Household Assets in Rural Zones
Rural assets considered in this discussion are land and
livestock.5
Farm size varies considerably across households.
The
sample contains farms as small as 1 hectare and as large as 60 hectares.
In the Peanut Basin zones, farms are typically 5-10 hectares; they are
somewhat smaller (3-5 hectares) in Senegal Oriental, where traction
equipment is not as common.
Access to household land by different members of the household is
subject to some variation across zones (see Table 3.8). Across-zone
comparison shows a range of 62 to 90 percent of males 215 years having
their own field(s).
The lowest access is in central Basin (Niakhar),
which has a serious land constraint.
The northern Basin also exhibits
a relatively low rate of "own fields" for males.
This may be due more
to the low productivity of agriculture in this zone than to the
availability of land.
Men most likely to not have fields are unmarried
sons, brothers, or uncles of the household head.
Participation by women ranged from a low of 44 percent in the
northern Peanut Basin to a high of 82 percent in Senegal Oriental,
Access to land was quite high in the southeastern Peanut Basin (77
percent) and in both zones of the central Peanut Basin (60-66 percent).
This simple picture of access to land without specifying the crop
or the size of the plots farmed fails to provide a true picture of the
economic
returns generated
by different groups having access.
Examination of net returns to agricultural production during the survey
shows that in the zones where the largest percent of women have access
to land, women's cropping income is a larger share of agricultural
income for the zone.
The typical share of net cropping income going to
women is about 10 percent.
It is highest (23 percent) in Senegal
Oriental, where a relatively large share of women stated that they had
access to land.
It is lowest in the northern Peanut Basin - a zone
where only a small share of women had access to land.
The typical rural household owns about 20 animals (including
traction animals and poultry) with an approximate value of 120,000 to
220,000 francs (CFA) per household or 13,000 to 21,000 FCFA per adult
equivalent.
Livestock holdings are extremely variable across households within
each zone (see Table 3.9).
For example, the range between the minimum
and maximum number of animals owned in a given zone varies between 55
and 107.
The range in value of holdings per household is more than
5 Although most rural households own animal traction equipment, it
was generally purchased in the sixties and seventies.
After
depreciation of this equipment, the net present value would be zero or
very close to it.
Only in Senegal Oriental, where traction equipment
has been available more recently through the cotton parastatal, is
there relatively new equipment. For this reason, traction equipment
has not been analyzed.
.
.
47
4,000,OOO FCFA in one zone (Senegal Oriental) and there are no zones
where it goes below l,OOO,OOO FCFA.
Even standardized values in FCFA/AE
continue to exhibit broad ranges between the smallest owners, who have
less than 2,000 FCFA/AE and the biggest owners, who
have 175,000 to
500,000 FCFA/AE.
Table 3.8 -- Access to land for persons 2 15 years in rural zones
Percent faming own field 1988/89 season
Zone/vi 11 age
Ma1 es
Females
Northern
70.1
44.4
Peanut Basin
Sagatta
68.6
36.0
Khelcom
83.8
41.0
Darou Cisse
55.2
60.0
Central Basin
62.1
61.2
(Niakhar)
Niakhar
58.1
54.1
Mboltogne
64.5
69.0
Mboyene
63.6
62.5
Central Basin
90.0
66.0
(Colobane)
Colobane
84.6
57.5
Teck Mar
94.6
62.2
Khayane
91.9
80.0
Southwest Peanut
75.5
54.2
Basin
Passy
70.3
27.8
Keur Aly Gueye
76.7
57.1
Ndiathiang
80.0
77.8
Southeast Peanut
83.0
77.0
Basin
Dioly Mandat
78.0
74.0
T. Thieckene/M. DiocKoul
77.0
83.0
Kouthia
95.0
77.0
Central Senegal
86.0
82.1
Oriental
Missirah
73.0
74.0
NiaoultYTouba
80.0
76.0
Medina AlylMoukouty
77.0
70.0
Source:
ISRA/IFPRI project survey.
..!
:”
..
48
Table 3.9 -- Livestock assets in rural zones
Central
Characteristic
North
Southwest
Southeast
Senegal
Basin
Niakhar
Colobane
Basin
Basin
Oriental
Number of animals
Mean
22
17
15
20
22
26
Median
Minimum
1
1
1
1
1
1
Maximum
86
92
5 5
70
6 7
107
@$g&ijitii
.:‘. a
Me.;i...
294
567
212
283
461
533
Quartile values
Lower 25% own
less than
this amount:
49
4 7
7 5
65
a7
2 2
Median value:
129
154
122
191
339
221
Upper 25% own
more than
this amount:
274
645
278
423
649
566
Minimum
6
6
2
10
14
.5
Maximum
1895
4065
1334
1433
2866
3102
gJ&y&@ g&y&&<
.j:.: .;.: . . . . . . “.tiGa;, ,,.,., ,.,,.,.j .,.,.,,,,
3 2
7 1
26
30
5 3
46
Quartile values
Lower 25% own
less than
this amount:
7
9
11
12
11
5
Median value:
13
25
1 8
20
28
21
Upper 25% own
more than
this amount:
39
96
2 7
40
70
7 6
Minimum
.9
.9
.6
1.9
1.4
.1
Maximum
177
509
175
172
312
249
ource:
ISRA/IFPRI project survey.
otes:
Information presented on this table is based on livestock inventory data collected shortly before the beginning of the
1990/91 cropping season (Feb.-April, 1990). Animals are valued by category (cattle, sheep, goats, etc. without respect
to age, sex. or weight) using producer prices obtained from ISRAIIFPRI transactions data and average over the entire
survey period.
Note that values are reported in thousands of FCFA. Nia. Niakhar; Colo. Colobane.
:
.
:
.
.’
~
49
When comparing the typical value of holdings per AE across zones,
the southeastern Peanut Basin comes out on top, followed by the Niakhar
zone of the central Peanut Basin (median values of 28,000 and 25,000
FCFA/AE).
The zone with the lowest value of holdings is the northern
Peanut Basin
( d
me ian value of 13,000 FCFA/AE), where livestock
contributes substantially to overall household income (see Chapters 4
and 7).
Details concerning the mean and median size of livestock herds and
the mean, median, and quartile values of holdings by zone are presented
in Table 3.9.
Income from livestock transactions is discussed in
Chapter 7.
Indicators of Urban Livinq Standards"
Tables 3.10 and 3.11 present characteristics of urban households
giving the proportion of households in the two towns owning real estate,
household appliances or a means of transportation. The proportion with
access to water and electricity is also shown.
Habitat: Type and ownership
More than 75 percent of the households in both towns own the house
that they occupy (79 and 76 percent respectively in Kaolack and Tamba).
About half of these houses
are built of durable materials
(cement/concrete) in Kaolack and about one fourth in Tamba where housing
is typically made of "banco" (clay mixed with straw), crintin (straw
mats) or straw. In Kaolack, 42 percent own empty lots in the same or
other towns compared to 12 percent in Tamba.
Means of Transport
The Tamba households seem better equipped to meet their transport
needs: 41 percent of them own at least a 2 wheeler (bicycle, moped,
motorcycle) and 12 percent own a car or a truck.
In Kaolack, the
proportion is only 9 and 3 percent respectively for both types of
vehicles.
Household appliances
In general, urban households own few appliances: less than 33
percent of households in both zones combined own an appliance. Two times
as many households have a sewing machine in Kaolack as in Tamba (18 and
9 percent); however, the opposite is true in the case of fans and air
conditioners (12 and 24 percent). Fifteen percent of the households in
the sample in each town have at least a fan.
6 The following paragraphs, as well as any other sections of this
draft report presented in French, will be translated to English for
the final report.
:
.
..:
:
50
Table 3.10 -- Standard of living indicators for households in urban sample
Characteristics
X of households possessing
Kaol ack
Tamba
., . . . . . . ,.,. ., : : : ,. ,. ./ ,.
A. ~$j$;~“pi:~Pa
P....ngDorr+tt-ari’!’
Bicycle/motorcycle
9
4 1
Automobile/truck
3
12
Cart
3
6
B .
j$.$K;gtiGIB'i$ .:':?? ..>:.:pki
earn
Sewing machine
18
9
Refrigerator/freezer
15
1 5
Fan/air conditioner
12
2 4
Others: camping stove,
gas
- '
3 5
peanut mill
c .
@&ond cs
Radio/cassette player
9 7
88
TV/Video
9
2 6
Telephone
3
0
0.
m
Owner occupies house
79
7 6
Owner of another house
6
0
Building constructed in
4 8
21
cement
Owns parcel of bare land
4 2
12
Have electricity
3 6
3 5
Have water faucet in compound
4 8
6
Have well in compound
3
8 8
Source: IFPRI/ISRA project survey.
' not available for Kaolack.
.’
:
51
Table 3.11 --Settlement profile of households in the urban sample
Characteristics
Kaol ack
Tamba
&G!ige. . .
-.... ::dur$$ .I,... . . . .
,. f.tjii.‘ii:j ,,.,.,.,.,.................
,.I.~.~..~/,,j ..:, ,:,.. ,.
:.:., :
..,.:. . .
,:
‘.,,.:.: ::, ..:I..:.:.):.:.),.~
n’.rilfo~.:iilb~~~,~~~.~: ,..:.:.:...- ,.:.>::.:.:.: .,.,.,.,..,,,...,.....,.,.,,.
30
24
_,., . . . . . . . ., .. . . . . . . . .
P’liac~:..t3f.:i:arr,g:n:l,:~~~:~~~:~a~~~.~~:~~~~,~~ha~
.::.,...:..:
.,.,.
.,.,.,..;:.:..:.y
y::,,.. .
:;.:.:.:..:. :
,:,:,:,,
,,,,.:
:,:,.::,:,:.:,:,..:,,.,, d
.:,:,:,.
_,.,.,..... ,, :.:.:...:.:.:.:.:.:.:.:.:.:.:.:.:
. ., ,.
:.:.:
.,.,.:
. . . . . . . . . .
- Same town (% of households)
9
3
- Other urban semiurban centers (%)
22
30
- Rural zone in the same region (%)
33
37
- Rural zone in another region (%)
33
10
- Outside the country (X)
3
20
- Born there (%)
9
3
- Work (%)
54
71
- Rejoin family/marriage (%)
34
26
Source:
IFPRI/ISRA project survey (1990-91).
Almost all households in Kaolack have a radio and/or tape recorder
compared to 88 percent in Tamba.
However, it is surprising to note that
many more households in Tamba than in Kaolack own a television set or a
video: 26 percent compared to 9 percent.
A telephone is a luxury in
both zones; only one household in Kaolack subscribes to SONATEL.
Direct access to water and to electricity
Slightly more than one third of the households in each zone has
direct access to electricity through SENELEC. Kaolack has more difficult
access to water than Tamba; only half of the households from Kaolack
have a faucet or a well inside their compounds while the other half has
access to water mostly from public water fountains.
Almost all
households in Tamba (94 percent) have a private water source;
subscribers to SONEES are not numerous (6 percent) because wells
predominate as the main source of household water (88 percent) in Tamba.
Settlement of urban households in towns
The average length of urban settlement of survey households is 30
years in Kaolack and 24 years in Tamba (Table 3.11). In both towns, the
proportion of household heads born in the town of current residence is
low: 3 percent in Tamba and 9 percent in Kaolack. Households from rural
zones in the same or other regions constitute almost two thirds of the
Kaolack sample and 38 percent in Tamba. However, those who are native of
other urban or semi-urban centers within the country make up almost one
.
.
.
.
:
.
.!
”
.’
.
52
fifth of the sample in each town. There is also an important share of
households in Tamba coming from outside Senegal (about 16 percent
compared with 3 percent in Kaolack), especially from Guinea and from
Mali.
The search for work, commercial factors or transfers initiated by
one's employer are the reasons most often mentioned to explain the
settlement of a household in town: 51 percent in Kaolack and 59 percent
in Tamba.
Other motivations
are
strictly personal
("to join
family/husband or "to be entrusted to a family member") and make up 34
percent of the cases in Kaolack and 22 percent in Tamba.
RELEVANCE OF HOUSEHOLD CHARACTERISTICS FOR AGRICULTURAL PRICE POLICIES
Two important reasons for looking at sample demographic and
socioeconomic characteristics are (1) to ensure that our sample
resembles the general population and (2) to examine factors that might
cause households or zones to respond differently to agricultural price
policies.
From the above discussion,
we conclude that the sample is
generally representative of the population it is to reflect in terms of
age, gender, and ethnic distribution.
The large share of the population
(15 years of age has important economic implications because this puts
a greater burden on the "active" population.
Data for the urban zone
show that 17-19 percent of the entire sample population had a source of
income at the time the survey began.
This means that each "employed"
individual must earn enough to support more than five people.
The data confirm what has been observed elsewhere -- that
education and literacy levels in both rural and urban Senegal are
extremely low.
This finding has important implications for economic
modeling exercises that assume farmers are able to use available price
information
and
their
knowledge of agricultural
input/output
relationships to make rational, profit-maximizing or cost-minimizing
production and marketing decisions.
When farmers are unable to keep
records of economic returns to different input combinations across years
in an environment as climatically fickle as Senegal, it is extremely
difficult to sort out the relative profitability of different
activities.
Village reconnaissance results reported in Chapter 2 suggest that
cropping activities provide the principal source of women's income in at
least 75 percent of villages in each zone. Information provided in this
chapter shows, however, that women's fields produce a relatively small
share of total household crop output.
These results imply that while
income from personally managed plots accounts for an important share of
women's total income,
the value of women's crop production is a
relatively small share of total household production.
. .:
53
Much of the women in development literature on Africa portrays
African women as the primary agriculturalists responsible for producing
most of the food crops.
Information provided above shows that this is
not true
for Senegal, where
household heads have the primary
responsibility for providing household cereals and female headed
households are rare.
Although Senegalese women in the IFPRI/ISRA study
zones manage only a small share of the area cultivated, they play a very
important supporting role by providing labor to fields managed by other
members of the household.
Understanding the extent of female
participation in agricultural production is extremely important when
designing agricultural policies to increase production and marketed
surplus.
In the urban zones, there is a significantly higher level of
female-headed households (14-26 percent). The implications of this for
household well-being (levels of income in particular) and consumption
habits (preferences among cereals,
for example) warrant further
examination.
The wide range of local and migratory activities practiced by
sample households reflects the transition from a purely agricultural to
a mixed economy that is taking place throughout Senegal. In many cases,
this is a transition that is being forced on an agrarian population that
is becoming less and less able to provide its basic needs through
agricultural production due to short and unreliable rainy seasons, a
deteriorating natural resource base, and a growing population. In other
cases,
it can be the result of poor infrastructure or inappropriate
agricultural policies.
The role played by the diversification of income sources in
providing food security and liquidity for improving agricultural
production is one of the most important issues to be addressed by this
study.
The data suggest that agricultural policies in Senegal cannot
be adequately designed without paying greater heed to the broader policy
issues of rural development that take into account the links between
agricultural and nonagricultural sectors.
Chapter 4 deals with these
issues in depth.
One generally anticipates that households with larger asset bases
enjoy better standards of living and tend to be more economically
productive.
We use information on access to land and livestock holdings
to examine the extent to which these anticipated relationships can be
corroborated.
Assets are frequently used to classify households. Knowing the
effect that agricultural policies have on households with different
asset bases and the extent to which asset levels determine the type of
response a household makes to a particular policy can greatly improve
policy design.
Our survey is unique in that it contains data on assets
as well as income and expenditure.
We are able to use all of these
variables to classify households, examine the relationship between the
different indicators,
and the relationship between levels of the
indicators and impact from or response to policy interventions.
.
.’
:
54
SUMMARY OF MAIN POINTS
Chapter 3 presents descriptive information about the sample
households and household heads.
Both demographic and socioeconomic
characteristics are discussed.
The main points to retain about the sample are:
Average age is 20 years while median age is 15.
The active population is about 45-50 percent of the total
sample.
The main points to retain about the sample households are:
Household size is generally 11-13 people and 9-10 adult
equivalents;
it is smaller in the urban zone of Tamba (9
persons and 7 AE) and the Niakhar zone of the central Peanut
Basin (10 persons and 8 AE).
The nuclear family accounts for 55-60 percent of the
household population except in Kaolack where it is 47
percent.
Households in the southwest Peanut Basin and the Niakhar
zone of the central basin have the highest rates of school
enrollment (about 70 percent of households with children in
school) while the southeastern basin households have no
children in school at all.
Zones with the highest outmigration during the dry season
are the northern basin and Colobane in the central basin;
Senegal Oriental has the highest rates during the rainy
season.
Seventy to 90 percent of adult males have access to farm
land in all zones but Niakhar in the central basin where
only 62 percent have access;
45 to 60 percent of females
have access in the north, center, and southwestern basin
while 75 to 80 percent have access in the southeastern basin
and Senegal Oriental.
Households in Niakhar in the central basin, the southeastern
basin and Senegal Oriental have the largest value of
livestock holdings per AE (roughly 45 to 70,000 FCFA/AE),
while households in the northern and southwestern basin, and
Colobane in the center have the smallest values (25 to
30,000 FCFA/AE).
.!
.
:
.
55
The main points to retain about household heads are:
Rural household heads are almost exclusively males (only two
females in the entire sample).
Urban household heads are 26 percent female in Kaolack and
14 percent female in Tambacounda.
Only 2 percent of rural household heads have any formal
education (French system) while 20 percent of the Kaolack
sample and 30 percent of the Tambacounda sample have some
formal education.
In most cases, the characteristics of the sample reflect the
characteristics that the reconnaissance survey suggested one would find
in each zone.
:
-,
,
4. DESCRIPTION AND ANALYSIS OF RURAL AND URBAN INCOME
INTRODUCTION
The level of income is an important determinant of the level of
household expenditures in general and the composition of the food
basket in particular.
It can also influence production behavior,
particularly a household's ability to make capital investments or to
purchase costly inputs when credit markets do not function well.
The share of income earned in different activities can also
influence consumption and production behavior.
In most studies, data
on total household expenditures are used as a proxy for income because
it is so difficult to enumerate income data directly.
Unfortunately,
data sets based on expenditures cannot address questions about income
diversification and what this implies for household response to policy
changes.
For example, assuming equal levels of income, a Senegalese
household earning 30 percent of income from cropping and purchasing 80
percent of cereal needs will have a different reaction to a rice price
increase than a household that earns 80 percent of its income from
cropping and only purchases 10 percent of its cereals.
These two
households could also be expected to respond differently to programs
designed to improve
agricultural productivity by decreasing soil
erosion
if such programs
competed
with nonfarm activities for
household labor.
The IFPRI/ISRA data set provides a full enumeration
of household income by source,
thereby permitting one to better
understand household consumption and production behavior.
Objectives
Levels,
distribution,
and sources of income in rural Senegal
differ across zones due to differences in the natural resource
endowment (climate, soils), the infrastructure (transportation,
communication),
and the institutional environment (savings, credit,
markets).
The overriding objective of this chapter is to show how
levels, distribution, and sources, of income differ across study zones
and to understand what this
implies for policy design and
implementation.
The chapter responds to the following specific objectives:
1.
Compute the level and distribution of rural and urban income for
each study zone.
2.
Examine the degree
of disparity in incomes
using frequency
distributions and Gini coefficients.
.
I’
:
.’
*
:
:
.
.
57
3.
Analyze the nature and degree of income diversification by zone
and by income strata.
4.
Develop a topology of diversification patterns.
5.
Model the relative importance of factors that explain higher
shares of nonagricultural incomes.
6.
Discuss the relevance of income levels and diversification
patterns for design of rural development policies.
Research questions
The objectives mentioned above can be summarized in the form of
research questions that we attempt to answer in this chapter. The
questions are as follow:
1.
What is the importance in absolute terms (level) and relative
terms (share) of cropping and non cropping incomes in the total
income of rural households by agroclimatic zones?
2.
What is the total level of income in urban sectors by study zone
and by strata?
3.
What different activities contribute to the composition of rural
and urban income,
and what is the share of each sector of
activity in the total income of the household?
4.
What are the most common income diversification patterns?
5 .
What are the factors that determine which type of diversification
pattern evolves in a particular zone?
6.
What is the policy relevance of the different levels and
diversification patterns found?
RESEARCH METHODS
Zones and time periods covered
Incomes for the rural sample are reported for harvest years
1988/89 and 1989/90 for five study zones (Niakhar and Colobane in the
central Peanut Basin, the southwestern and southeastern basin, and
Senegal Oriental).
For the northern Peanut Basin, only harvest year
1988/89 is covered.
A harvest year begins with the cereal harvest
(October 1) and ends just before the following harvest begins
(September 30).
We use a harvest year rather than a calendar year or
a cropping year to analyze income.
The harvest season provides an
excellent point of departure for the income analyses, because the
success of the harvest tends to influence household expenditure
.
”
.:
.
:
.,
.
:
58
patterns as well as decisions on allocating labor and capital to both
cropping and non-cropping activities throughout the rest of the year.
The urban analysis covers the towns of Kaolack and Tambacounda
for the period July 1, 1990 through June 30, 1991. These dates were
dictated by funding considerations.
As urban incomes are not subject
to the pronounced seasonal patterns of rural incomes, the beginning
and ending months of the survey period do not pose methodological
problems.
The lack of overlapping periods for the rural and urban
surveys does, however, limit the possibility for modeling rural and
urban demand concurrently.
Data collection methods
Fall, Kelly, and Reardon (1989) describes in detail the types of
questionnaires and procedures used to collect the income data.
The
following paragraphs summarize the most important aspects of data
collection procedures used.
l Cropping income was collected using a plot-level questionnaire
that enumerated all inputs and outputs for each field.
@ Data for on farm livestock income come from a questionnaire
that enumerated all purchases, sales, and home consumption of
livestock.
The questionnaire was administered monthly.
l Data on locally earned off-farm wages, cash transfers, and
borrowing were collected monthly.
l Data on migration remittances were collected by enumerating all
cash and inking remittances received by household members and
corroborating this information through in-depth interviews with
migrants when they returned home.
l Data on off-farm activities were collected using a variety of
techniques.
For local off-farm activities, interviewers used a variable
interview
schedule with more frequent interviews for persons
exercising activities that were difficult to enumerate (for example,
shop owners or petty traders with rapidly changing inventories) and
longer recall periods for less complex or less frequently practiced
activities (for example, preparing and selling donuts once a week at
the market).
Questionnaires on income from off-farm activities were
introduced about six months later than other questionnaires and a long
recall used to cover the first six months.
The recall was usually
conducted by a researcher rather than by the interviewer.
This was
done because we felt that project personnel had to develop the
necessary rapport with respondents before asking detailed questions
about a topic as sensitive as personal income.
.*
:
”
:
59
Analysis procedures
The activities are organized into the following nine production
sectors: own cropping activities
livestock production
wages from off-farm cropping
commerce
services
transport
food preparation
gathering
handicrafts
Table 4.1 provides a list of the activities classified in each of
these nine sectors.
Three additional 'Isectors" are used to report
income from cash transfers, borrowing, and remittances from household
members on migration.'
The methodology used for the analysis is based on the computation
of all sources of remunerative activities (cash and/or in-kind) taking
place locally/or in migration by members of a household during the
year. The income of each activity is calculated by deducting variable
and fixed costs (depreciation) from the gross receipts.
Cropping income is estimated in terms of net returns to family
labor following methods described in Chapter 5.
Livestock income includes the households' net receipts from
transactions of animal products and by-products and the income from
activities related to animal husbandry (shepherding).
Capital gains
and losses associated with changes in the size of the household's
total livestock holdings are not considered.* Details of estimation
methods used are described in Chapter 7.
Net transfers and borrowing reported in this chapter cover only
t r a
cash n s a c t i o n s .
No attempt has been made to value and incorporate
inking transfers and borrowing into the income analysis.
The household income for a given sector constitutes the sum of
net income earned by all household numbers from activities in the
sector.
Total household income is the sum of income across all
sectors plus net transfers and borrowing. Net transfers and borrowing
' The definition of all the sectors appears in Fall (1991).
2 See Ndiaye 1993 for a more complete analysis of the IFPRI/ISRA
livestock data for Niakhar in the central basin and the southwestern
basin.
Ndiaye combines the stock and transactions data to estimate
capital gains and loses.
.’
1
.’
Table 4.1 -- Classification of income generating activities by sector
CM-I
Livestock
Agricultural
Cannerce/
Service
Transportation
Food
Gathering
Cottage
:
Cropping
Labor’
Trade
Processing
I&tries
Crop
Herding
Cropping labor
Shop owner
Tire repair
By cart
Peanut processing Gathering/sale
Tailoring
Production
to produce oil
of wood
Cattle sales/
Fruit/vegetable
Petty-trade
Bicycle repair
By motorized
Errbroidery/
Fruit/
trading of
gardening, labor
vehicle
Sale of Koumpa/
Gathering/sale
Hand-
Vegetable
household
Livestock
Apprenticeship
beignets
of fruit
sewing
Gardening
animals
trade’
Various
Maid
transportation
Sale of roasted
Gathering/sake
BLacksmith
Sale/processing
Fish trade
Peanuts
of birds
animal by-
Civil Servant
Basketry
products (mi lk,
Kola nut trade
Bread making
leather)
Guard
Leatherwork
Drink trade
Restauration
Village Chief
Cloth
Tools, parts,
weaving
t i r e r e p a i r
Marabout
Jewelry-
Hardware trade
Praise singer
making
Tobacco trade
Veterinary
Fence
Assistant
construction
Assistant trader
Hechani c
Peanut trade
Hairdresser
Blacksmith
Retirement
Pension
Source:
IFPRI/ISRA survey data.
’ For a third party outside the household.
’ This involves the buying and reselling of cattle that are not considered part of household herds.
61
do not originate in a specific sector, but constitute supplemental
sources of income.
Total household income was standardized on the basis of household
size (income per capita) and on the basis of household composition by
age and gender (income per adult equivalent). These results were then
compared to estimates of household expenditure per adult equivalent as
a control.
Only those households where income per AE was at least 70
percent of expenditure per AE were retained for further analysis.
Given the difficulties inherent in collecting income data the
assumption was that the expenditure data were more reliable than the
income data.
The average income per adult equivalent was then calculated for
each village,
permitting one to examine inter-village differences
within each zone as well as the influence of market infrastructure
within and across zones.
Individual household observations were
weighted by household size (AE) in calculating village averages.
Tables with the village and zone averages per AE are presented in
Appendix tables A4.
Most of the discussion in this chapter draws on estimates of
average income per AE or average income per capita for each zone. In
these estimates individual household observations were weighted by
household size (AE) and a coefficient to correct for intentional over
sampling of market villages.3
For the 1988/89 harvest year data for 189 rural households were
analyzed.
In 1989/90, the number of households with complete income
data was 158.
Some analyses (primarily ones making inter-annual
comparisons) use only those rural households for which two full years
of income data are available; for these analyses, the sample size is
141 households.
The urban analysis covers 60 households (31 in
Kaolack and 29 in Tambacounda).
LEVELS, DISTRIBUTION, AND SOURCES OF INCOME
Levels of Income
Table 4.2 shows that average household income in rural zones
ranges from 27000 FCFA ($95) per AE in the central Peanut Basin zone
of Niakhar in 1988/89 to 72000 FCFA ($250) per AE in the southeastern
Peanut Basin during harvest year 1989/90. Average income for urban
3 See Kelly and Reardon, 1989, for a full discussion of sampling
methods and stratification procedures.
.
,.
:
.
,’
-.
Table 4.2 --
Level and distribution of net household income per AE (FCFA) by zone and year
~
Northern
Central Basin
Central Basin
Southwest Peanut
Southeast Peanut
Central Senegal
Peanut Basin
Ni akhar
Co1 obane
Basin
Basin
Oriental
a
c
Kaol
k
Tambacounda
Y e a r
a8189
88189
89190
88189
89190
88189
89/90
88189
89190
88189
89190
go/91
go/91
1. Rev. per AE
42509
26975
32696
33319
56246
43459
67208
49806
72051
39163
42487
107760
102625
N.Obs.
-29
25
3 1
33
35
30
2 7
36
36
36
29
3 1
29
C.V. (%)
70
4 1
37
60
49
60
a4
41
68
7 1
145
125
64
2. Rev. per Capita
32072
20298
20297
25744
42190
32675
49309
36685
51897
29340
30893
83625
79950
N.Obs.
2 9
25
3 1
33
35
30
27
36
36
36
2 9
3 1
29
C.V.(%)
7 3
42
39
6 2
5 1
57
a2
39
67
69
7 5
140
67
.
3. Gini
Coefficient
.34
.30
.23
.30
.26
.30
.43
.I8
.26
.38
.49
.48
.35
.,
Source:
ISRA/IFPRI Project survey, 1988-1991.
Notes:
In rural zones, the data are presented by harvest year (from October 1 - September 30) for 1988/89 and for 1989/90. However, in urban zones,
the period is from July 1 - June 30, 1991.
..(I.
:i
Sagatta was only covered in 1988/89.
The Gini Coefficient is calculated using income per capita.
6 3
households is higher -- about 105,000 FCFA per AE ($365) in both
Kaolack and Tambacounda in 1990/91.4
The level of average income per AE varies significantly across
zones.
Two of the three study zones with high agricultural potential,
the southwestern and southeastern Peanut Basin, have the highest
levels of income both years.
Average incomes in the third zone with
high potential (Senegal Oriental) are, however, at the lower end of
the spectrum both years.
The Niakhar zone in the central Peanut Basin
exhibits the lowest average incomes across all zones in both survey
years.
Average urban incomes in 1990/91 are about double most of the
averages for rural zones in the year following the above average
1989/90 harvest.
The highest net income per capita found in rural areas during the
survey period (52,000 FCFA in the southeast of the Peanut Basin} is
about one fourth of the value of gross domestic product per capita
estimated from macroeconomic datae5
The urban incomes are about half
of the GDP estimates.
This illustrates, as one would expect, that GDP
figures fail to provide an accurate picture of purchasing power and
living standards for an average Senegalese household.
A comparison of household income and the cost of minimum needs in
each zone helps one evaluate the adequacy of incomes earned by rural
households."
Following the poor 1988/89 harvest, income did not
cover minimum needs for 50 percent of households in the Niakhar zone
of the central basin.
In the Colobane zone of the central basin, the
southwestern basin and Senegal Oriental, 20 to 30 percent of
households had incomes less than minimum needs.
Following generally better 1989/90 harvests, Niakhar (central
Peanut Basin) and Senegal Oriental still had a substantial share of
households with combined income from all sources less than "minimum
needs" levels (35 and 21 percent, respectively).
Interannual differences in levels of rural income are large.
Following the better 1989/90 harvest, average household incomes (from
4 The per capita equivalents of the Niakhar and southeastern
Peanut Basin figures cited above are 20,000 and 52,000 FCFA,
respectively ($70 and $180); the urban per capita figure is 82,000
FCFA ($280).
During the survey period, US$l.OO was worth
approximately 290 FCFA.
5 See, for example, DPS, 1988, or IMF, 1992, both of which report
GDP per capita of about 180,000 FCFA.
6 Minium needs is defined as the cost (using zone-specific
prices) of 1900 kcal/AE/day of coarse grains plus 20 percent to cover.
other essential food and nonfood expenditures (see Kelly, 1993 for
details)
..!
:
64
all sources) increased by 20 to 70 percent throughout the Peanut
Basin.
The central Peanut Basin zones had both the lowest (Niakhar)
and the highest (Colobane) increases.
Most of the increase came from
better cropping outcomes which increased by 50 to 100 percent between
the two years, but some also came from higher off-farm incomes.
Only Senegal Oriental failed to experience an important increase
in income in 1989/90.
This was due in part to poor rainfall but also
because farmers in the zone boycotted cotton production the second
year.
Distribution of Income
The degree of concentration of income in the households for each
study zone is estimated using the Gini coefficient.
This measurement
provides information on the equality (coefficient nearing 0) or the
lack of equality (coefficient nearing 1) of income distribution for a
given population.
The Gini coefficients calculated on the basis of
average income per capita are shown for each zone and year in Table
4.2.
Gini coefficients range from .18 to .49, but are generally in
the .25 to .35 range.
Most zone estimates are similar to those found
in other Sahelian countriese7
The distribution of income in central Senegal Oriental and in the
southwest Peanut Basin is more unequal (Gini about .4) than in the
other rural zones.
This is due in part to the level of income of
certain migrants (in Senegal Oriental) and large merchants (southwest)
which are considerably higher than the average
income
in their
respective zones.
The urban distribution appears to be more skewed
than the rural with Gini coefficients of .35 in Tambacounda and .48 in
Kaolack.
One also notes that in the zones with greater agricultural
potential,
Gini coefficients increased following the better 1989/90
harvest.
This result suggests that something is preventing the poorer
households from fully realizing the benefits that accrue from better
rainfall.
Tables 4.3 and 4.4 show the frequency distribution for average
net household income by zone using seven categories.
The modal
category for rural zones in both survey years is 30,000 to 60,000 FCFA
per AE.*
This category accounts for approximately 45 percent of
rural households in the sample both years.
Table 4.3 shows that the
two central Peanut Basin zones and Senegal Oriental have the highest
number of households with incomes below 20,000 FCFA per AE.
Incomes
at this level generally fail to cover minimum needs.
7 Nigeria , Niger and Burkina Faso, for example, are .28, .30 and
.33 (Matlon, 1979; Reardon, 1989; Hopkins and Reardon 1993).
a This is approximately $100 to $200 per AE per year.
.
.
:
I
I
65
Table 4.3 --
Frequency distribution of average net household incomes (FCFA/AE) by zone and
year for rural zones
Part A: Harvest Year 19BB/B9
Incaiie
Northern
Central Basin
Southwest
Southeast
Central
Category
Peanut
Peanut
Peanut
Senegal
Basin
Niakhar
Colobane
Basin
Basin
Oriental
(number of households)
1100,00
5
0
1
1
1
2
60-99,999
0
0
2
6
1 3
5
30-59,999
1 7
10
1 5
13
1 9
1 3
20-29,999
5
7
9
a
3
8
15-19.999
2
6
6
1
0
4
lo-14,999
0
2
0
1
0
2
<
10,000
4
Number of
29
25
3 3
30
36
38
Households
Part B: Harvest Year 1989/90
Income
Northern
Central Basin
Southwest
Southeast
Central
Category
Peanut
Peanut
Peanut
Senegal
Basin
Niakhar
Colobane
Basin
Basin
Oriental
(number of households)
I100,OO
na
0
3
5
9
3
60-99,999
na
1
1 2
4
1 5
2
3659,999
na
1 9
1 7
1 2
1 1
1 0
20-29,999
na
6
3
4
0
9
1s19,999
na
4
0
2
1
3
lo-14,999
na
1
0
0
0
2
Number of
na
31
35
2 7
36
29
Households
.
.
_’
. :
66
Table 4.4 -- Frequency distribution of
average net household
incomes (FCFA/AE) in
urban zones
Incune
Kaol ack
Tarnbacounda
Category
(number of households)
I100,00
13
12
60-99,339
11
9
30-59,999
5
6
20-29,999
1
1
1s19,999
1
1
10-14.999
Number of
31
23
Households
The modal group for urban households in 1990/91 is the highest
income category --
2 100,000 FCFA per AE.'
The modal group includes
about 40 percent of the urban sample.
Sources of Income
One of the major strengths of the IFPRI/ISRA data set is the
manner in which it permits household income to be disaggregated by
sector in both rural and urban zones.
Table 4.1 at the beginning of
the chapter illustrated how income generating activities were mapped
into 9 economic sectors.
Table 4.5 shows average income by sector for
each rural zone and year of the study while Table 4.6 presents the
urban results.
Table 4.5 reveals that cropping activities contributed from 20 to
70 percent of rural household income in 1988/89 and 40 to 80 percent
in 1989/90.
An extreme case of only 20 percent cropping income was
found following the locust invasion of the northern Peanut Basin in
1988/89.
Consistently large shares of cropping income (greater than
70 percent) were found only in the southeastern Peanut Basin.
Most
zones were in the 40 to 50 percent range in 1988/89 and 70 to 80
percent range in 1989/90.
' This is approximately $345 per AE per year.
67
Table 4.5 - Net household incom per AE (FCFA) by zone, year, and sector for rural sample.
No
Own
Ag
Food
cot
Ttl
Obs
Crop
Lvsk
labor
Cm
Serv
Tv
Prep
Gath
Ind
Trf
Bar
Mgrt
Unk
Rev
NORTHERN PEANUT BASIN
88-89
2 9
7340
7899
578
4004
3919
0
1766
0
8810
1640
491
2762
3300 42509
Pet
17
19
1
9
9
0
4
0
2 1
4
1
6
8
100
c.v
.92
2.00
1.40
4.55
4.01
3.21
2.91
2.03
3.27
1.62
1.61
.70
CENTRAL BASIN - NIAKHAR
88-89
25
12642
1505
7 5
4699
4427
0
119
0
2386
87
0
805
232 26977
Pet
47
6
0
1 7
1 6
0
0
0
9
0
0
3
1
100
c.v
.68
2.10
1.06
1.69
1.94
4.72
2.32
3.52
2.42
3.69
.41
89-90
31
25931
1125
44
2260
1086
0
40
0
987
-20
119
929
1 9 5 3 2 6 9 6
Pet
79
3
0
7
3
0
0
0
3
0
0
3
1
100
c.v
.51
3.81
1.81
2.49
3.29
5.84
4.37 -11.90
6.03
2.75
3.98
.37
CENTRAL BASIN - COLOBANE
88-89
33
18154
7512
168
1190
1250
814
1597
1 2
1178
-88
123
1409
0 33319
Pet
5 4
2 3
1
4
4
2
5
0
4
0
0
4
0
100
c.v
.74
1.19
1.15
3.64
2.94
4.74
3.76
5.38
3 . 8 2 - 1 3 . 0 8
3.42
1.82
.60
89-90
35
42624
2615
169
864
2716
2521
1192
6 1
1469
5 2
180
1782
0
56246
Pet
7 6
5
0
2
5
4
2
0
3
0
0
3
0
100
c.v
.69
2.55
1.20
3.82
3.65
4.75
1.40
4.52
3.61
33.49
8.60
1.76
.49
SOUTHWESTERN PEANUT BASIN
88-89
3 0
20738
1810
1180
4027
10292
1250
323
724
1686
6 3
1350
0
15
43459
Pet
48
4
3
9
2 4
3
1
2
4
0
3
0
0
100
c . v
.53
1.82
2.85
2.47
2.15
3.82
7.42
5.08
3.50
4.19
1.36
4.91
.60
89-90
27
33426
2380
1257
16746
9117
1176
833
593
1772
1234
-1474
140
7 67208
Pet
5 0
4
2
25
14
2
1
1
3
2
-2
0
0
100
c.v
.49
2.68
3.00
3.08
2.24
4.27
3.54
4.60
3.50
3.57
-2.89
3.07
4.19
.84
SOUTHEASTERN PEANUT BASIN
88-89
36
33954
6465
1 5 1
2551
2699
54
253
182
677
-119
2660
114
165
49806
Pet
6 8
1 3
0
5
5
0
1
0
1
0
5
0
0
100
c.v
.48
1.23
1.85
2.85
1.54
5.17
4.80
5.63
2.97
-7.43
1.74
6.15
4.47
.41
89-90
36
51003
12865
126
5652
2827
414
189
155
1216
-1343
-1436
2
380
72051
Pet
71
1 8
0
8
4
1
0
0
2
-2
-2
0
1
100
c.v
.73
1.12
1.87
2.62
1.96
6.12
5.01
5.66
3.25
-3.85
-4.06
25.94
3.84
.68
SENEGAL ORIENTAL
88-89
36
18688
5931
687
2579
2147
9
1447
276
764
2209
331
4115
0
39163
Pet
48
1 5
2
7
5
0
4
1
2
6
1
11
Cl
100
c . v
.70
1.56
1.39
2.80
2.42
4.07
3.43
3.56
3.64
2.02
2.16
3.43
.71
89-90
2 9
15673
9326
656
967
1536
7
-92
20
2250
2670
-39
9504
0
42487
Pet
37
22
2
2
4
0
0
0
5
6
0
22
0
100
c.v
.64
2.09
1.69
3.59
2.11
4.98
-14.97
4.77
3.13
2.89
-48.60
3.43
.81
Source:
IFPRI/ISRA survey data.
Note:
See Table 4.6 for legend to sector abbreviations.
:
,
.’ 2
,’
:
.
Table 4.6 --
Net household income per AE (FCFA) by tom and sector for urban sample.
Z O N E
SECTOR
Own
Lvsk
corn
Serv
Tw
Food
Gath
cot
Trf
t30r
Mgrt
Total/
Crop
Prep
Ind
AE
KAOLACK
Rev/se
660
300
24085
49380
0
14570
0
4570
8740
640
2865
107760
%
1
2 2
46
0
1 4
0
4
8
1
3
100
c v (%)
368
1353
270
146
0
379
0
277
192
1752
416
129
Rev/se
485
0
11875
77115
175
5285
0
5910
2785
575
-440
102625
%
0
1 2
7 5
5
0
6
3
-1
100
c v (X)
673
0
180
102
244
282
0
260
374
914
314
6 4
Source: ISRAIIFPRI Project survey, 1990/91.
Legend for sector abbreviations:
Own crop = onfarm cropping income
Lvsk
= livestock
colwl
= commerce and trade
Tv
= transportation
Food Prep= Food processing/restaurant
Gath
= gathering plant/forest products for resale
Cot Ind = cottage industries
Trf
= cash transfers/gifts
B0r
= cash borrowing
Mgrt
= migration
,Unk
= unknown
69
Interannual differences in levels of cropping income are quite
large.
In the central Peanut Basin, levels of cropping income doubled
between 1988/89 and 1989/90.
In the southern Peanut Basin zones, the
level increased by 50 percent.
In Senegal Oriental the level of
cropping income dropped between the two years by about 20 percent (due
largely to the cotton boycott and poor rains in 1989/90).
Net income from livestock transactions was an important share of
total income (13 to 23 percent) in all zones but Niakhar in the
central basin and the southwestern basin (5 percent) following poor
1988/89 harvests."
The following year, livestock income accounted
for less than 5 percent of income except in the southeastern basin and
Senegal Oriental where it represented about 20 percent. The tendency
for livestock income to vary inversely with cropping income and,
therefore, smooth total income is clearly evident.
The final columns of Table 4.5 that concern income earned in the
agricultural sector (broadly defined as all cropping, forestry, and
livestock activities) are agricultural wage labor and gathering
activities.
Income is classified as agricultural wages only when it
is earned by working off-farm.
Both cash and inking payments are
included.
Gathering includes all activities where forest products
were gathered and sold; gathering for home consumption was not covered
by the survey.
It is noteworthy that only a very small share of total
household income (0 to 5 percent) is generated in agricultural wages
and gathering activities.
Low agricultural wage income is anticipated
in situations where there is not a large class of landless households.
When the demand for agricultural labor is high, everyone is working
their own fields and not hiring out their labor".
Gathering (particularly leaves for sauces and firewood) for home
consumption is common, but the resale of gathered products is not a
major source of household income in any of the IFPRI/ISRA study zones.
In summing the four columns that report income earned in the
agricultural sector (broadly defined), we find that households in most
zones earn from 60 to 90 percent of total income in these activities
except in the northern Peanut Basin where less than 40 percent of
income was earned in agriculture and animal husbandry.
Haggblade,
Hazel1 and Brown (1989) found that 25 to 30 percent of rural household
income in
selected countries
of sub-saharan Africa came from
" The extreme reluctance of the Serer households in the Niakhar
zone of the central Peanut Basin to sell their livestock is evident in
the IFPRI/ISRA data.
This zone has the largest livestock herds, the
lowest income levels, and the lowest caloric intake in 1988/89 yet
they sold very little livestock to supplement income.
" In most of the study zones, supplementary laborers
ar,e
temporary migrants from Guinea.
This is particularly true at harvest
time.
.
.
I
.
.‘.
:
:
:
.
.
I
. .
70
nonagricultural activities, while 70 to 75 percent came from cropping
and
animal
husbandry.
This suggests
that the northern and
southwestern basins and Senegal Oriental -- zones with less than 70
percent of income earned in agriculture (broadly defined) -- are
somewhat atypical of the sub-saharan region in general.
Although agricultural incomes remain an important share of total
income, many households still fail to meet minimum needs with incomes
earned directly in agriculture,
livestock and forestry activities.
Nonagricultural incomes, therefore, play a critical role in assuring
household food security and well being.
For this reason, it is
important to understand how sources of income vary by income group.
Nonagricultural income has a variety of sources; commerce and
service activities account for the largest share in most zones while
handicrafts (northern Peanut Basin) and migratory remittances
(northern basin and Senegal Oriental) tend to be more zone specific.
Household income from net transfers (gifts) and borrowing account
for less than 1 percent of income in both the Niakhar and Colobane
zones of the central Peanut Basin.
In the remaining zones, transfers
and borrowing together usually account for 5 to 7 percent of total
income.
Net transfers and borrowing are generally positive for sample
households suggesting that funds may be moving into sample zones from
other areas of Senegal or beyond national borders.
The only exception
was found in the southeastern Peanut Basin following the good 1989/90
harvest when sample households' net transfers and borrowing reduced
total income by almost 3000 FCFA per AE on average.
Relatively low shares of income from borrowing (either positive
or negative) suggest that although credit markets do exist, they are
not
p;oviding
important sums
of money over extended periods of
time.
Table 4.6 shows that the service and trade sectors take first and
second positions,
respectively in the two urban zones studied in
1990/91. The service sector represents 75% of the total income per AE
in Tambacounda, while trade accounts for 12%. The other sectors are
relatively less important.
In Kaolack, however, the service sector generates almost half of
the total income by AE (46%), and trade represents 22%. It is
important to note that Kaolack is reputed to be a trade center as a
result of its dynamic peanut trade and its proximity to Gambia which
stimulates cross-border trade.
Due to the role that Kaolack plays as
12
" Matt Warning, a Ph.D. student at UC Berkeley, is
currently doing an independent study of credit using the ISRA/IFPRI
sample in Niakhar and Passy.
He will collect additional data on
access in Niakhar and Passy.
This data will then be combined with
existing data and used to model demand and supply of credit.
,
.
‘:
.’
.’
,
71
a transport and commercial crossroads, the town registers an
impressive number of snack bars and restaurants. Results show that
income from the food preparation sector is quite important (14% versus
5% for Tambacounda).
Cottage industries provide about 5 percent of urban income. This
is comparable to the share in most rural areas.
The migratory sector provides 3% of the total household income in
Kaolack.
It is negative
in Tambacounda suggesting that sample
households sent more money to members on migration than they received.
This occurs when a migrant fails to find employment quickly.
Transfers are more important in Kaolack than in Tambacounda (8%
and 3%).
The fact that transfers are positive in urban as well as
rural areas - and account for a larger share of income in urban areas
- fails to support the conventional wisdom that cash transfers are
flowing from urban to rural areas.
More in-depth analysis of the data
is required to sort this out.
Tables 4.7 and 4.8 illustrate the manner in which the share of
income from different sectors varies across income strata.
For each
study zone and year households were ranked by total net income per AE.
Each household was then assigned to an income tercile according to its
ranking in the zone.13
Due to large differences in levels of income
across zones and nonproportional sampling procedures, we have made no
attempt to lump households together as the aggregate results could be
misleading.
Table 4.7 reveals that following the poorer 1988/89 harvest, the
share of cropping income is neqatively correlated with the level of
total income in all zones but Colobane (central basin). Households in
the lowest tercile have the highest shares of cropping income while
households in the highest tercile have the lowest share.
This
supports the hypothesis that households with more diversified incomes
are better able to compensate for crop shortfalls following poor
harvests because their total incomes are higher.
Following the above average 1989/90 harvest, the negative
correlation between the share of cropping income and the level of
total income remains unchanged in the two zones of the southern basin.
In the central basin, however, the middle and higher income households
have a larger share of cropping income than the poor.
One also notes (data not shown) that the absolute value of net
cropping income tends to increase more between the two harvests for
l3 In other words,
the third of households with the lowest
incomes in the zone were placed in the lowest income tercile, the
middle third of households in the zone in the middle tercile, and the
highest third din the top tercile.
.-,
;
.
I
,
.1’
:
:
72
Table 4.7 --
Average net total incane per AE (FCFA) and shares by sector and tercile for each rural
zone and year.
S H A R E O F
I N C O M E B Y
S E C T O R :
TOTAL
FOOO
COTTAGE
ZONE
TER INCOHE/AE
CROPS AGLAB LVSTK COI4 SEW TRPT PREP GATH INDUST MIGRA TRNFS BOR UNK
Northern
low
24839
36
4
24
0
2 0 2
0
1
11 2
1
1 8
Peanut
mid
35492
2 9
1
2 1
2
1 0 14
0
7
9 7
2 7
Basin
top
75336
1 1
0
1 4
1 7
17
0
0
0
3 1
3
1
1 4
Central
low
15481
5 7
1
7
6
16 0 0
0
1
9 1
0 3
Basin
mid
26909
5 3
0
a
2 7
8 0 0
0
3
1
0
0 0
Niakhar
top
41219
38
0
4
1 3
21
0 1
0
18
3 0
0 0
Central
low
20799
56
1
1 7
1 5
10 3
0
1
4 1
0 0
Basin
mid
30838
5 9
1
28
0
60 1
0
3
3
-1
0
0
Colobane
top
56919
6 1
0
17
2
3 4 6
0
4
4 0
1
0
Southwest
low
23721
7 0
6
a
5
0 5 0
1
0
0 0
5 0
Peanut
mid
37130
5 9
1
7
6
14 3 3
0
4
0 0
3 0
Basin
too
77613
42
2
0
1 3
31 2 0
2
6
0
0
2 0
Southeast
low
34002
7 6
7
3
6 0 0
0
1 0
2
1
Peanut
mid
55025
75
9
3
5 0 0
1
0 0
7 0
Basin
tOD
77331
70
15
6
20 1
0
0 0
5 0
Central
low
17883
66
9
0
8
0
0
1
1
7
2 0
Senegal
mid
34193
5 9
1 3
6
5
0
7
0
5 2
1
0
Oriental
top
72789
5 0
1 3
a
30
1
1
16 5
0 0
S H A R E O F
I N C O M E B Y
S E C T O R :
TOTAL
FOOD
COTTAGE
ZONE
TER INCOME/AE
CROPS AGLAB LVSTK COM SERV TRPT PREP GATH INDUST MIGRA TRNFS BOR UNK
Central
low
20960
68
0
12
1 0
6
0
1
0
2
2
0
0
0
Basin
mid
34205
aa
0
2
1
3
0
0
0
0
5
0
1
1
Niakhar
top
48636
7 8
0
-1
1 2
2
0
0
0
8
2
0
0
0
Central
low 3 5 3 8 0
68 1
3
8
60 3 1
4
7
1
-
l
0
Basin
mid 5 6 8 6 2
77 0
3
1
8 0 1 0
6
3
1 0
0
Colobane
top 100114
a3 0
4
0
2 7 2
0
0
1
1 1
0
Southwest
l o w 27988
a5
i
i
i
a
i 3 1
0
0
2 -2
0
Peanut
mid 48642
56 1
10
3
12 5 5 0
2
1
5
0
0
Basin
top
140647
40 2
2
39
14 1 0 1
4
0
0 -3
0
Southeast
low 42285
a2 0
6
3
5
0
0
1
5
0
-1 0
0
Peanut
mid 71963
a0
0
20
2
4 2 1
0
2
0
-5 -6
0
Basin
top 126319
74 0
15
9
2 0 0
0
a
0
0 0
1
Central
low 18951
47 8
1 9
0
4 0 0
1
1 2
8 0
0
Senegal
mid 3 0 6 2 2
57 1
27
2
6 0 -1
0
:
0
3
2
0
Oriental
top 141131
64 0
8
2
10
0
o
3
20
2 0
0
Source:
IFPRI/ISRA
survey
1988-90.
.
Note:
See Table 4.6 for legend to sector abbreviations.
73
Table 4.8 --
Average net total incane per AE (FCFA) and shares by tercile and sector for
each urban zone
S H A R E O F
I N C O M E B Y S E C T O R :
TOTAL
FOOD
COTTAGE
ZONE
TER INCOHE/AE
CROPS AGLAB LVSTK COH SERV TRPT PREP GATH INDUST HIGRA TRNFS BOR UNK
Kaolack
low
46142
0
0
0
I5
53
0
6
0
6
1
18
-2
3
mid
92319
0
0
1
1 2
5 6
0
8
0
7
8
6
-2
3
top
245092
0
1
0
30
38
0
1 9
0
2
1
5
2
1
Tabma
1OW
36046
0
3
0
32
30
1
2 1
0
10
0
4
-1
0
mid
95407
0
0
0
22
62
0
9
0
6
0
1
0
0
top
177737
0
0
0
3
90
0
0
0
5
-1
3
-1
0
Source:
IFPRI/ISRA survey 1988/89
Note:
See Table 4.6 for legend to sector abbreviations
households in the higher income terciles than for those in the lower
terciles.
This suggests that the wealthier households are better able
to realize the benefits of good rainfall than the poorer households.
The mechanism by which this works is not clear, but it is reasonable
to hypothesize that the wealthier households have greater liquidity to
purchase inputs (peanut seed, in particular) and are better able to
hire labor to get around bottleneck periods.
There is no clear pattern across zones and terciles concerning
the sectors in which off-farm incomes are earned.
For example,
livestock and migration are important sources of income for the poor
in the northern Peanut Basin and an important source of income for the
rich in Senegal Oriental.
Cottage industries are important to the
wealthy in the Niakhar zone of the central basin and in the north.
Poor households in the lowest tercile account for most of the
agricultural wage income earned in the north and the southwest.
The
commerce and service sectors provide an important share of off-farm
income across all terciles.
As Fall (1991) has noted, it is the
capital intensity of different activities within each sector that
tends to differentiate the lower and upper income terciles.
For
example, the poor in the commerce sector tend to be petty traders
while the rich are shop owners.
The above discussion has' stressed the critical role that
nonagricultural incomes play in rural Senegal.
It must be stressed,
however, that most of the nonagricultural sources of income are not
independent of the agricultural sector.
The amount of income earned
in nonagricultural activities is often determined by how successful
agriculture is in a particular zone.
This is true because many
nonagricultural activities are tied to agriculture through production
and consumption linkages.
.I
:
”
:
:‘.
I
,
74
INTERSECTORAL LINKS OF INCOME GENERATING ACTIVITIES IN RURAL AREAS
This section describes in greater detail the nature of activities
by sector and by zone to facilitate the examination of intersectoral
linkages in rural areas.
We are particularly interested in
identifying upstream and downstream flows between activities in the
agricultural and the nonagricultural sector.
As was mentioned in the previous section, cottage industry, trade
and services constitute the principal sectors of nonagricultural
activity in rural zones. One notes, however, that the nature of these
activities differs by zone.
Cottage industry in the north of the Peanut Basin includes
embroidery, basketry and shoe making.
In the other zones where cottage
industry is
less important,
the sector is dominated by sewing,
blacksmithing,
weaving and jewelry making. Some of these activities
are closely related to those upstream of agriculture (for example,
with blacksmithing as an input). Other activities of the sector are
likely to exhibit strong consumption linkages with the agricultural
sector.
For example, growth of the agricultural sector would lead to
an increase in the demand for cottage industry goods.
Trade is oriented toward the marketing of condiments, fish and
goods of first necessity in the northern Peanut Basin and the Niakhar
zone of the central Peanut Basin.
These products are, for the most
part, not directly linked to the local agricultural economy.
On the
other hand, in the Colobane zone of the central basin and the southern
Peanut Basin zones trade in cola nuts, cereals and peanuts is more
common. These activities exhibit downstream linkages with agriculture.
Animal trading ("tefanke") dominates in the northern basin and the
center of Senegal Oriental where animal husbandry is very important,
This activity has primarily downstream links to agriculture, but also
exhibits some upstream links.14
The service sector in all zones is made up primarily of "griot"
(praise singers) and "marabout" services, security and domestic work.
The latter activity is usually conducted in Dakar or in other regional
capitals. The income from these activities does not have production
links to the local agricultural economy. In the center of the Peanut
Basin and in Senegal Oriental, the service activity "repair of
equipment" does exhibit production linkages (as an input) to local
agriculture.
l4 Traders use agricultural by-products such as peanut hay to
feed animals, which in turn are potential inputs for agriculture (as
traction animals and sources of manure)
:
;
75
Food preparation is tightly tied to agriculture in all the zones.
This concerns production and sale of grilled peanuts and doughnuts15;
and extracting oil from peanuts (principally, in the Colobane zone of
the central basin). In the southeastern Peanut Basin, women living in
the market .village earn a large share of their personal income by
selling water to thirsty people and animals on market day.
Therefore, it is possible to conclude that many of the
nonagricultural activities in the zones covered by the study exhibit
some upstream or downstream production linkages to the agricultural
sector.
The nonagricultural activities upstream of agriculture are
limited almost exclusively to animal husbandry and blacksmithing
(production and repair of tools, for example). This reflects the
relatively low use of inputs in Senegalese agriculture.
Downstream
production links are primary food processing or marketing activities.
In addition to the production links, one must also consider the
indirect consumption links.
As agricultural incomes increase, the
demand for many locally produced goods and services will also increase
(tailors, crafts, and "griots", for example).
In
conclusion,
the
increasing
importance of
income
diversification in the rural areas doesn't exclude the need to improve
agricultural productivity.
Results presented above highlight the fact
that many nonagricultural
activities are tied directly (through
production linkages) or indirectly (through consumption linkages to
the agricultural sector.
The intersectorial links described here
require
additional
study to better understand and quantify the
interaction of the agricultural and nonagricultural sectors in the
IFPRI/ISRA study zones.'"
IMPORTANCE OF THE INFORMAL SECTOR IN URBAN ZONES
In urban zones, formal activities are defined as originating from
a public or private enterprise, with a legal statute, that allows the
State to obtain information on the firm's operating budget. These
activities are therefore more easily accounted for in the national
accounting system and estimates of national income. On the other hand,
many informal activities escape the national accounting system, due to
the fact that they are not legally structured and can easily avoid
reporting income and paying taxes.
l5 In town, doughnuts are generally prepared with wheat flou r;
but in rural zones,
it is possible to find doughnuts made from mill et
flour.
l6 More systematic analysis of indirect consumption linkages
the IFPRI/ISRA study zones is currently being conducted w Enh
supplemental funding obtained from USAID/Washington.
7 6
As was mentioned in the preceding section, services and trade are
key sectors of activity in the composition of urban income. However,
the nature of activities in the two towns in the study present
variations.
It is possible to note that the service sector in Kaolack is
composed essentially of room rental, "marabout", "griot", tailors,
electricians,
guard services,
and transport (driver) services. The
majority of these activities are in the informal sector (76% of the
service sector). Trade activities in Kaolack are dominated by shop
tending,
sale of food, fish, ice, cloth and used clothing. The trade
activities in Kaolack are almost exclusively informal (95%). This
shows clearly that in Kaolack the informal sector dominates income
generating activities of households.
In
Tambacounda,
the
service
sector is
represented by
administration activities,17 transport (drivers),
"marabout", maids,
and room rental. More than 70% of these activities are from the
informal sector.
Trade is composed especially of activities in the
informal sector such as trade of condiments, charcoal,
fruit and
tobacco.
PATTERNS OF HOUSEHOLD INCOME DIVERSIFICATION
Knowing what factors are driving households to diversify income
and the long-run
economic
implications
of existing patterns of
diversification is an important prerequisite to designing rural
development policies that will foster growth in rural incomes and
increased agricultural production.
Furthermore,
understanding income
diversification strategies of rural households permits one to better
judge household-level response to proposed policy changes.
To better
grasp these elements, a topology of income diversification patterns
and a
model
of factors driving
households to diversify were
developed."
" Receipt of retirement benefits, teachers, and other civil
servants.
'* There is an extensive literature on income diversification in
Asia,
but evidence from Nigeria (Matlon), Burkina Faso (Reardon,
Delgado, and Matlon), and Niger (Hopkins and Reardon), suggests that
there are some important differences
between Asian and Sahelian
diversification patterns.
The topology and model presented here are
specific to
Senegal
and
based on
survey
findings;
the
conceptualization of the different factors to consider in the topology
and model drew heavily on the analyses of the other West African
research cited above.
‘I
.
*
I
.
:
.’
:
.
.
:
.:
*
77
Toooloqv of diversification patterns
The topology provides a framework that facilitates the tasks of
(1) diagnosing the underlying causes of the diversification in each
zone, (2) understanding what factors are motivating production and
consumption behavior in each zone, and (3) anticipating the type of
household response that different policy interventions are likely to
elicit.
Three major styles of diversification have been identified:
(1) an outward-oriented
"copinq" pattern characterized by:
declining levels of cropping income over time due, in
large part, to a poor, deteriorating natural resource
base;
a dearth of activities with production and consumption
linkages to agriculture
an important share of household income coming from off-
farm activities only in emergency situations following
severe crop shortfalls
consistently low total household income due to poor
cropping outcomes and limited off-farm opportunities;
(2) an outward-oriented, lonq-run, adaptive pattern characterized
by:
declining levels of cropping income over time (as in the
"coping"
case) or low levels of cropping income because
technological
transformation has not yet taken place
and/or price risk is high due to poorly developed markets
and physical isolation from centers of demand.
nonfarm income often earned outside of the zone of
residence;
little reinvestment, if any, of nonfarm income into
cropping activities;
total incomes that are low but generally adequate (higher
than those for the "coping" case)
in the long-run, households may move out of crop
production entirely;
(3) an inward-oriented, lonq-run, adaptive pattern, characterized
by:
good
cropping
performance
stimulating
growth of
production and consumption linkages;
*
..’
*.
,
.
:.
.
.
.
:
.
‘.
_’
78
off-farm
activities
that
complement
rather
than
substitute for cropping income;
income from off-farm activities that is reinvested in
crop production;
the presence of well developed transportation and market
infrastructure
a policy environment that fosters development of factor
markets (particularly credit and labor).
The "copinq" pattern is found in the Niakhar zone of the central
Peanut Basin.
Households in this zone had the lowest average incomes
in the study during both survey years.
Following the extremely poor
1988/89 harvest, off-farm income was 53 percent of total income. It
was earned primarily from petty commerce of nonlocal products (sugar,
tobacco, beer) and migration to Dakar.
Off-farm activities in this
zone do not have many production linkages to local
cropping
activities.
Following the better 1989/90 harvest, cropping income
doubled and off-farm income dropped to only 20 percent of total
income.
Even with the doubling of cropping income, however, total
income remained low.
Households in this "coping"
zone have not yet developed an
adaptive strategy.
Crop land is a serious constraint preventing
households from expanding crop production through extensive practices
and also limiting the size of livestock holdings that can be kept in
the zone year-round.
Rainfall has declined over time and soil
degradation has increased, eroding the zone's agricultural resource
base.
Overall low levels of income do not permit productivity
enhancing investments in either farm or off-farm activities.
If nothing is done to improve the physical environment in this
zone, households will be forced to develop a more aggressive type of
outward-oriented diversification through migration to urban areas or
other zones.
The extent to which future policies should encourage the
continuation or intensification of cropping and livestock activities
in this zone cannot be answered by the IFPRI/ISRA data base as one
needs to evaluate the capacity of the zone's natural resource base to
support agricultural intensification.
An example of the outwardly-oriented, lonq-run, adaptive pattern
is found in the northern Peanut Basin, where cropping potential is
extremely
low
due to
poor
rainfall
and
degraded
soils.
Diversification is based to a large extent on migration and provision
of products and services to Dakar.
There are, however, some cases of
"inwardly-oriented"
activities that appear to have production and
consumption links to the local livestock sector.
An important share
of income comes from sales of household livestock.
Income from sales
of animal byproducts is also important as are commissions to sales
agents who transport and market local livestock in Dakar.
79
Incomes are generally adequate in this type of zone and most
households manage to assure food security,
even in years of severe
crop failure.
Another zone that appears to be moving toward an outwardlv-
oriented strateqy is Senegal Oriental.
In this case, however, the
strategy does not appear to be well adapted to the zone.
Senegal
Oriental has promising cropping and livestock potential given high
average rainfall,
less degraded soils than the Peanut Basin, and
abundant supplies of crop and pasture land due to low population
density.
Nevertheless, households earn a relatively low cropping
income ,per AE and a high share of "outwardly-oriented" off-farm
income.
There are many indications that market failures and lack of
infrastructure
are
constraining
agricultural
productivity
and
unnecessarily pushing Senegal Oriental's households into migration (22
percent of household income in 1989/90).
The low 1989/90 cropping
income was due in part to a cotton boycott.
The boycott was to
protest high input prices and low net returns for cotton, suggesting
the existence of some market or price policy failure.
*
Cereal sufficiency ratios in Senegal Oriental were higher in
1988/89 than those in several zones where larger shares of total
cereal production were sold,
This suggests the possibility of market
failure or inadequate transportation infrastructure.
Because migration remittances are being invested mostly in
housing and mosques and rarely in agricultural production,
one
surmises that households do not view agriculture as a good investment.
The lowest overall returns to family cropping labor were found in
Senegal Oriental.
This implies that migration is probably the best
income strategy at present.
Senegal Oriental clearly has agricultural potential. The policy
issue here is how to design policies that will encourage households to
invest in the intensification of agriculture rather than in outward-
oriented off-farm activities, particularly migration.
An inwardly-oriented adapted strateqy is currently found in the
two zones in the southern Peanut Basin and Colobane in the central
basin.
These zones have relatively low levels of migration income.
Much of the off farm income in these three zones is linked to crop
production activities
(peanut marketing
and processing, cereal
marketing, restaurants).
Animal husbandry is also an important source
of revenue in Colobane and the southeast but not in the southwest.
" The zone had the lowest average cropping income in 1989/90 and
the third lowest in 1988/89.
1
.I
80
The challenge
in these zones is to provide the incentives
necessary to improve agricultural productivity.
Without growth in
agriculture the impetus for growth in the local nonfarm sector will be
lost.
If this happens, the prevailing credit constraints will be
magnified because reinvestment of off-farm income in crop production
will decline.
Determinants of off-farm income shares
Understanding why diversification off-farm is taking place is
important to policy
analysts because households will
respond
differently if the factors pushing them to diversify are different.
Information about factors driving diversification can be used to
develop nonprice policies that will speed up and increase market
response to price signals.
It is relatively easy to understand the role of climate and
natural resource constraints in pushing households into "outwardly-
oriented" off-farm activities in zones such as the northern basin. It
is more difficult to assess the role of inadequate infrastructure,
credit constraints and other market failures.
Regression analysis was
used to gain a better understanding of the relative importance of
these factors.
Households located in market villages earned 12 percent more of
their income from off-farm sources than households in other villages
(approximately a 60 percent increase over average shares of 18 percent
for nonmarket villages).
The larger share for market villages can be
explained by "pull" factors -- access to market infrastructure draws
households into off-farm activities because the infrastructure makes
it easier to find employment and easier to sell goods and services.
Factors associated with the agroclimatic zone also determine the
share of off-farm income.
Households in the southwestern Peanut
Basin, earned a statistically significant greater share (-t 14 percent)
of their total income from off-farm activities than other zones. This
represents a 70 percent increase over average non-farm shares of 20
percent for other zones.
Higher nonfarm income in the southwest is explained by a number
of "pull" factors.
Proximity to the very dynamic markets in Kaolack
and the Gambia encourage individuals to engage in petty commerce
activities.
The relatively dynamic agricultural sector in the zone
increases the opportunity for employment at peanut collection points
and in cereal marketing activities.
This zone also has the second
highest level of average household income in the IFPRI/ISRA study; the
higher income is no doubt stimulating the demand for nonagricultural
goods and services."
" Coefficients for market and zone were significant at, the .01
and .06 level, respectively.
._
.
.
.
.
.
:
81
"Push" factors such as poor access to land and credit, shortfalls
in cereal production,
and inadequate levels of household savings
(represented by livestock assets) were also statistically significant
variables explaining the share of off farm income in total household
income."
Elasticities estimated for these variables suggest that:
*
A 10 percent increase in the cereal sufficiency ratio decreases
the share of off-farm income by 7 percent;
*
A 10 percent increase in land assets (proxied by cultivated area)
decreases the off farm share by 5 percent;
*
A 10 percent increase in livestock assets decreases the off-farm
share by 3 percent;
*
A 10 percent increase in agricultural and food credit decreases
the off-farm share by 1 percent.
The elasticity for the cereal sufficiency ratio is large and
negative,
as expected.
Households better able to meet food security
needs from home production have
less incentive to seek off-farm
income,
all else being equal.** This suggests
that programs or
policies which reduce food insecurity (food aid, food banks, or
credit,
for example) could encourage households in zones with good
agricultural potential to concentrate more on agricultural production.
The same types
of programs in zones
with an outwardly-oriented
adaptive pattern of diversification or a coping pattern would be
unlikely to elicit greater agricultural productivity.
The elasticity for land assets is also large and negative as
anticipated.
All else equal, more land permits households to earn
more cropping income by producing the desired mix of cereals and cash
crops; higher cropping income reduces the pressure to move off-farm to
supplement income.
This result is not easy to translate into policy
because it involves land tenure issues that were not covered by the
study.
The implication is, however, that larger farms should be
encouraged if the objective is to reduce unnecessary diversification
that reduces productive efficiency.
Larger herd size also decreases the share of off-farm income
because livestock provide savings that are easily converted to cash
when households need liquidity for food or input purchases, reducing
the need to move off-farm.
Policies to improve animal health and
water supplies (a major constraint to animal husbandry in several
zones)
when the carrying capacity permits larger herd sizes would be
2' These effects were all significant at the .Ol level or better
except the credit variable which was significant at the .08 level.
22 Cereal sufficiency is defined relative to beginning stocks of
cereals, avoiding problems of reverse causality.
.
.
_ .
.
i
I
*
*
,’
82
most likely to diminish off-farm diversification and promote more
integrated cropping/livestock systems in zones with inwardly-oriented
diversification patterns.
Such policies would be unlikely to improve
crop productivity in zones with coping or outwardly-oriented, adapted
strategies.
Although the size of the elasticity for credit is smaller than
that for other factors, it is an extremely important variable from a
policy perspective.
The credit elasticity implies that better access
to credit permits households to increase the share of cropping income
in total income (that is, it diminishes the need to seek off-farm
income).
This suggests that improving access to credit might permit
households to use labor which is now engaged in off-farm activities to
improve farm productivity -- planting trees or building windbreaks,
for example.
In sum,
the significant coefficients on three of the above
variables -- cereal sufficiency ratios, livestock holdings, and access
to credit -- provide ample support for the hypothesis that growth in
agricultural productivity is being constrained by liquidity problems
that push farmers into greater income diversification than would be
the case were institutions for credit and savings better developed.
Although these regression results provide some insights into the
nature of rural income diversification in the Peanut Basin, the model
did not directly address the links between risk and the share of off
farm income in total income.
Additional work on the risk aspects of
diversification is clearly needed.
SUMMARY OF MAIN POINTS
The discussion in this chapter has stressed the importance of
both the agricultural and nonagricultural sectors in the income
generating strategies of rural households.
While the agricultural
sector continues to provide the majority of income in all zones but
the northern Peanut Basin, income from cropping alone is frequently
not sufficient to meet minimum needs.
The food security and general well being of rural households in
the IFPRI/ISRA study zones depends on a balanced program of rural
development that stimulates economic activity in both the agricultural
and nonagricultural sectors.
The extent of production and consumption
linkages between the two sectors found in most zones suggests that
improved agricultural productivity can have positive repercussions on
incomes earned in nonagricultural activities.
Inadequate credit and
infrastructure (commercial and transport) have been identified as
factors that are preventing households from realizing their full
agricultural potential, particularly in Senegal Oriental.
Household income generating strategies were found to differ
significantly
across
zones.
The
existence of
different
*.,j
.
.‘
.j
:
.
.
:
‘_
83
diversification strategies implies that successful rural development
policies will need to respond to the particular circumstances found in
each zone.
This point will be developed further in the concluding
chapter of this report.
The most important quantitative findings presented in this
chapter follow:
*
Across all zones, rural household income is most often (i.e., the
modal range) in the 30,000 to 60,000 FCFA/AE ($100-200) range
following both good and poor harvests; 36 percent of households
in HY88 and 22 percent in HY89 earned less than this amount -
most of these households failed to cover minimum needs.
*
The modal income category for urban households was 2 100,000
FCFA/AE; approximately 40 percent of households were at this
level; only 7 percent of urban households earned less than 30,000
FCFA/AE.
*
Equality of income distribution across households within a zone
is typical of other West African cases
analyzed (Gini
coefficients in
.30 to .35 range) except in the southwestern
basin,
Senegal
Oriental,
and
Kaolack where
the
income
distribution is more skewed (Gini > .40).
*
Highest rural household incomes are found in the southern Peanut
Basin during both survey years.
*
Lowest rural household incomes are found in the Niakhar zone of
the central basin and in Senegal Oriental.
*
The share of nonagricultural (noncrop, nonlivestock) income in
total income is highest in the northern Peanut Basin (about 70
percent) and in the southwestern basin (about 50 percent);
*
The share of agricultural (cropping and livestock sectors) income
in total income is highest in the southeastern Peanut Basin and
the Colobane zone of the central basin (about 80 percent).
84
REFERENCES
Fall, A.A. 1991.
Composition multisectorielle et distribution du
revenu dans le bassin arachidier: Remuneration du capital et du
travail, Memoire de confirmation a l'ISRA, et document no. V du
projet ISRA/IFPRI, janvier, Dakar.
Fall, A.A., V. Kelly, and T.A. Reardon.
1989. Household-level survey
methods used in the IFPRI/ISRA study of consumption and supply
impacts of agricultural price policies in Senegal. Document 2,
IFPRI, Washington, D.C.
Haggblade, S., P. Hazell, and J. Brown. 1989. Farm-nonfarm linkages in
rural Sub-Saharan Africa, World Deve?opment, Vol. 17, No. 8.
Hopkins, Jane and Thomas Reardon. Aqricultural Price Po1ic.v Reform
Impacts and Food Aid Tarqetinq in Niqer. Washington, DC, IFPRI,
April 1993.
Reardon, Thomas.
“Income Diversification in Burkina Faso, unpublished
mimeo". 1988.
Ndiaye, 0. 1993.
Micro-level Small Ruminant Transaction Anal.vsis and
Livestock Policv in Seneqal.
Masters Paper,
Michigan State
University, E. Lansing.
Kelly,
v. 1993.
"Aspects Economiques de la Production et de la
Commercialisation des produits Agricoles au Niveau des Menages du
Bassin Arachidier et du Centre du Senegal Oriental".
Workshop
paper presented at l'atelier du Project IFPRI/ISRA/BAME sur
"Politiques Agricoles et Comportements des Menages dans le Bassin
Arachidier et au Senegal Oriental", LNERV/Dakar Hann, 29-30 Mars.
Reardon,
Thomas,
Christopher L.
Delgado,
and
Peter
Matlon.
"Determinants and Effects of Income Diversification amongst Farm
Households in Burkina Faso".
Journal of Development Studies.
January 1992.
I
.,
:
.
I
:
.:
.
.
85
5. CROP PRODUCTION
INTRODUCTION
Chapter 4 described
income-generating patterns
for sample
households.
Particular attention was paid to the relative importance of
net cropping income in total income.
In this chapter we look more
closely at crop production
activities,
addressing the following
questions:
What is the relationship between gross and net cropping income?
To what extent does cropping income cover minimum needs?
What are the land and input use patterns and costs?
What are the returns to labor?
How much does crop production contribute to women's personal
income?
To what extent is coarse grain production satisfying household
cereal needs?
In answering these questions, differences across zones, and crops
are discussed.
METHODOLOGY
Each field cultivated by sample households for the production years
1989/90 was enumerated'.
Data were collected on all purchased inputs
(usually fertilizer, insecticide, or herbicide), all inputs coming from
home production (usually seed and manure), and all labor inputs (family
labor and hired labor).
The quantity of inputs used and the price (if
purchased) were enumerated (except for manure).
Labor inputs were recorded for each of the major activity periods
(clearing,
land
preparation,
planting,
weeding,
harvesting,
transporting).
If labor was not household labor, amounts paid in cash
or in kind were recorded.
All labor inputs were divided into broad
gender and age categories.
' Although crop production data were calculated for HY90, there
is no corresponding income or expenditure data for this harvest year %
because the rural survey ended in December 1990, consequently, no
analysis of HY90 crop production data is presented in this report.
1
..;
.
:
*
8 6
The number of harvest units was enumerated for each field. Total
harvest was estimated by weighing two to three samples of the harvest
unit (a bundle for cereals , a sack for peanuts) and multiplying this by
the total number of units harvested. All fields were measured.
Gross income is calculated for each field by multiplying total
quantities harvested by the average producer price for each product and
zone.
Net income is calculated for each field by deducting the cost of
all purchased inputs and the imputed value of all home-provided inputs.
Returns to family labor are calculated by dividing the total number of
hours worked on a field (all activities combined) into the net revenue.
No allowance has been made for depreciation on animal traction
equipment.
An analysis of depreciation costs for the northern Peanut
Basin (Dakono 1991) showed that most equipment was so old that it had
already been fully depreciated.
For this reason, the depreciation
analysis has been dropped.
In the future, it may be appropriate to
examine depreciation in the southeastern Peanut Basin and Senegal
Oriental where traction equipment has been introduced more recently than
in the older part of the Peanut Basin.
Total household income per adult equivalent is obtained by summing
the income across fields within the household and dividing by the
average number of AEs present in the household for the harvest year.
Average crop income per AE for each zone is obtained by averaging across
all households and weighting by household size and a coefficient used to
adjust for nonproportional sampling of market villages.
Although data on total quantities harvest in HY88 are available, no
input data were collected because the survey began at harvest time
(October/November 1988).
Net cropping incomes for HY88 presented in
Chapter 4 were estimated by assuming that the HY88 ratio of input cost
to output value was the same as the one we calculated for HY89. This
provides only a rough approximation of net cropping income for HY88.
Given that all other income data are reported as net income, it was
deemed appropriate to present at least an approximation of net HY88
cropping income.
Because the analysis of input use and returns to labor presented in
this chapter relies entirely on HY 89 data, the results present more
rosy picture of economic returns than would be the case if both HY88 and
HY89 input data were available because the 1989 harvest was unusually
good while the 1988 one was relatively bad.
The reader is reminded that the income per AE data presented in
this chapter
will differ somewhat from that presented in Chapter 4
because we use, data for 177 households that had complete cropping
information rather than the more limited set of households that had
complete income information.
Table 5.1 summarizes data presented in other tables throughout the
chapter.
.
..;
:
.
‘:
,_
.’
Table 5.1 -
Cereal/pulse production sufficiency ratios by zone and harvest year
n?g
E;pBR
:F?!!?
SwEg
sEJ?J3
. . . ..,.,
w
Average Gross Incane (FCFA/AR)
16760
29550
59750
40870
55320
20805
Average A& Income (FCFA/AE)
13170
23800
48030
35010
46410
17440
Input Costs as Share of Gross Income
(2)
2 1
2 0
20
1 4
1 5
1 6
share of Ret Income by crop
Peanuts (2)
65
39
65
43
68
52
Cereals (X)
30
61
3 5
2 9
40
other (X)
5
0
0
0
3
8
Share of Uininnsn Reeds Covered by Ret Crop Income
(I)
50
82
249
173
244
78
Median Amber of Hectares cultivated
Per Household
4.3
4.5
1 1
7
14
5
Per AE
-49
-5
1.21
-75
1.1
.49
Madian Returns to Family Labor (FCFA Per 8 hr person-da7)
Peanuts
870
530
1460
1380
720
610
Cereals
160
1010
1100
900
640
510
Share of Croppin(; Income by Gender (X)
wale
9 0
85
86
88
84
68
Female
8
13
12
1 0
1 2
23
Cereal Production Sufficiency Ratios
1988/89
.14
-47
.50
-58
.83
.70
1989/90
-25
.76
.98
1.03
1.02
-69
. . .
Ave. kilos of Cereal Surplus par AI?
1988/89
(1
(1
1
2
1 3
56
198&90
(1
15
6 9
4a
58
31
88
GROSS AND NET CROPPING INCOMES
Table 5.2 illustrates the difference between gross and net income
and the relative importance of peanuts, cereals, and other products
(cowpeas in the north and cotton in the southeastern Peanut Basin) in
total cropping income.
Despite higher input costs (due primarily to
high seeding ratios), peanuts account for about 65 percent of net income
in three zones (the northern basin, the Colobane zone in the central
basin, and the southeastern basin).
Cereals are dominant in two zones -
- Niakhar in the central basin and in the southwestern basin. In
Senegal Oriental households boycotted cotton in 1989/90 giving both
peanuts and cereals a larger role than usual (about 50 percent of income
comes from each crop).
Table 5.2 -
Canparison of average gross and net cropping income by crop for 1989/W
(FCFA/AE)
ZCMES:
NPB
C P B N
CPBC
SWB
SEPB
so
Peanuts:
Gross income
11900
14890
42630
20210
39700
11900
Net income
8590
9320
31070
14960
31690
9111
Cereals:
Gross income
4150
14660
16750
20130
14060
7425
Net income
3940
14480
16600
19890
13440
7024
Other crops:
Gross income
710
0
370
320
1560
1478
Net income
630
0
360
170
1270
1305
Share of peanuts
in net income:
.65
.39
.65
.43
.68
.52
Source :
IFPRI/ISRA Survey
.i
‘:
.
‘.
.’
89
It is useful to examine the extent to which net cropping income
covers minimum needs.
Minimum needs are defined as the cost (using
zone-specific prices) of 1900 kcal/AE of coarse grains plus a margin of
20 percent to cover other food and non food essentials. The cost of
minimum needs, defined in this manner, provide us with some measure of
the relative well being of households across zones with different costs
of living.
Table 5.3 shows that following the above average 1989/90 harvest,
only half of the zones (Colobane in the central basin, and the two zones
in the southern basin) earned net cropping incomes that covered the
estimated cost of minimum needs. In the northern Peanut Basin and in
Senegal Oriental, cropping income at the 75th percentile was below the
IFPRI/ISRA estimate of minimum needs (see Table A5.1 for a full
breakdown by percentile).
Note:
Cereal price data for 1989/90 is not available for the northern Peanut Basin so
we usa the 1968/89 minimum needs for 1989/90.
Table 5.3 - Levels of net cropping income and the cost of minimm needs
ZONES:
(FcFA/AE)
NPB
CFBN
CPNC
SWPB
SEPB
So
cost of
minimum
26055
29310
19245
20135
18950
22230
needslaelyr
Average net
cropping
income/ae/yr
13170
23800
48030
35010
46410
17440
Share of minimum
needs covered (2) 50
82
249
173
244
78
Source :
IFPRI/ISRA survey data.
9 0
LAND CULTIVATION AND INPUT USE PATTERNS
The average amount of land cultivated per household in 1989/90
varies significantly across zones.
The smallest farms (median size of
4 to 5 hectares) are in the northern basin, central basin (Niakhar), and
Senegal Oriental.
The largest median farm sizes are in the central
basin (Colobane) and southeastern basin (11 and 14 hectares,
respectively).
In the southwestern basin the median size is about 7
hectares.'
Cultivated area per adult equivalent is less variable than total
farm size, averaging from .49 hectares in the northern basin to 1.21 in
the Colobane zone of the central basin.3
Results from village
reconnaissance surveys (see Chapter 2) suggest that the relatively small
farm size in Niakhar is due to a serious land constraint, while in
Senegal Oriental, the lack of equipment and heavier soils make farming
large areas more difficult than in the Peanut Basin.
Peanut seed is clearly the most important variable input in terms
of value.
In general, about 50 percent of its total value comes from
household stocks while the rest is purchased. The central basin zone of
Colobane exhibits the highest per adult equivalent level of purchased
peanut seed (over 5000 FCFA/AE vs. less than 2000 FCFA/AE for other
zones).
Very little cereal seed is purchasede4
Use of fertilizer is limited primarily to contract cotton and
confectionery peanut production where credit can be obtained; 80 percent
of cotton fields and 100 percent of confectionery peanut fields received
fertilizer.
No fertilizer was used by sample households in the
northern and the central Peanut Basin.
In the southern Peanut Basin
zones, 4 to 5 percent of cereal fields were fertilized while the share
increased to 9 percent in Senegal Oriental. Greater fertilizer use in
these zones is related to better rainfall and, in some cases, proximity
to the Gambia.'
* Median farm size is used in lieu of means because the
distributions are skewed due to a few exceptionally large farms in
each zone. Table A5.2 presents more detail on the distribution of farm
size across zones and households.
3 Mean and median values of area per AE are quite similar in all
zones but the southeastern basin.
4 The seed purchase data used to calculate the shares of seed
value purchased cover only cash purchases, seed purchased on credit
through the confectionery peanut and maize contracts is not factored
in.
5 These zones are close enough to the Gambia to be able to
.
purchase lower priced Gambian fertilizer. Price data suggest that
this was frequently, but not always, the case.
.
’
‘.
:
91
Use of manure is somewhat more common but still not very important
in terms of total cultivated surface areas.
Manure was used on peanuts
in the northern and central basin (1-2 percent of fields) and on
confectionery peanuts in the southwest (14 percent) of fields. No
manure was used on cereal fields in the north while 5-8 percent of
cereal fields received some manure in the central and southern zones.
In Senegal Oriental 31 percent of millet/sorghum fields and 48 percent
of maize fields were manured."
With the exception of cotton production (which demands substantial
f amounts of herbicide and insecticide), the only other chemical input
used in large quantities was fungicide for peanut seed.
Hired labor falls into two categories: temporary (hired on a daily
basis) and seasonal ("navetane" employed for an entire season).
"Navetanes"
are most common in the Southeastern basin and Senegal
Oriental where about 25 percent of households use them. Households that
employ "navetanes" in these zones tend to hire several at a time. In
other zones, fewer than 10 percent of households contracted a seasonal
laborer during the survey period.
Use of day laborers is not substantial in most zones. Most hired
labor was used at harvest time with smaller amounts for land clearing,
planting, and weeding.
Payments for hired labor are extremely variable
making it very difficult to calculate an average wage rate for each
zone.
The low levels of hired labor and the absence of a consistent
wage rate suggest that rural labor markets are not functioning very
well.
Table A5.3 in the appendix provides supplementary details on
average and median expenditures by crop for the three major input
categories (seeds, fertilizer and chemical products, and hired labor).
No attempt has been made to impute costs to seasonal laborers (usually
paid inking with room and board) and manure.
RETURNS TO FAMILY LABOR AND LAND
"Returns to labor" is defined simply as net cropping income per field
divided by the number of 8-hour labor days worked on a field. No
distinction is made for different categories of laborers (male, female,
child) and no allowance is made for returns to capital or management.
Only the principal crop has been valued. Because the value of crop by-
products (particularly peanut hay) has not been taken into account the
returns to peanut production are really higher than those shown.
6 Percentages are percent of fields and not percent of
cultivated areas.
As the smaller household plots tend to be the ones.
receiving manure, the share of cultivated area receiving manure is
most likely much lower.
: ;
1
.
92
Table 5.4 presents the median values per 8-hour work day while Table
A5.4 in the appendix prov ides details on returns at the 25th, 5Oth, and
75th percentiles7.
Table 5.4 -
Median returns t .o family labor in cropping activities
for 1989/90
(FCFAD-hour person day)
Cereals
160
1010
1100
900
640
510
Other crops
2.50
0
245
70
90
360
Source:
IFPRI/ISRA survey
Median returns to an 8-hour workday in household crop production
ranged from 150 to 1500 FCFA/day, depending on crop and zone, during the
relatively good 1989/90 season.
The higher end of these returns
compares favorably to the SMIG (official minimum wage in the formal
sector) of 1500 FCFA/day.
Analysis of returns to family labor shows that there is little
incentive for most farmers to increase cereal production given current
technology and price relationships.
Cereal production was significantly
more remunerative than peanut production (almost double) in only one
zone -- Niakhar in the central Peanut Basin.* In all other zones median
returns to labor in peanut cultivation were higher (12 to 500 percent)
than those for cereals.
In three zones (the northern Peanut Basin,
Niakhar in the central Peanut Basin, and the southwestern basin) the
difference in mean returns to labor for the two crops was statistically
significant at the .05 level.'
All zones but the southwestern basin
7 Numbers in this table supersede those presented in Kelly
(1993).
Some programming errors that resulted in miscalculation of
total household hours worked have been corrected.
a Numerous explanations for the higher returns to millet in this
zone can be hypothesized: climatic and pest patterns during the study
period that were more favorable to millet than to peanuts, poor
quality peanut seed, or farming techniques that favor millet over
peanuts.
Further analysis of the IFPRI/ISRA crop production data and
additional field work are necessary to fully understand why this zone
appears to be different.
' In the Colobane zone of the central basin, the difference was '
significant only at the .15 level.
93
appear to be giving more attention to the crop that provides the highest
returns to labor (see Table 5.2)".
Although we have not yet analyzed returns to land, evidence from
other sources suggests that peanut crops are more profitable per unit of
land than cereals (Martin 1988, Kelly 1988). This is due primarily to
much greater yields per hectare for peanuts grown without fertilizer
(800-1000 kilos) than for cereals (400-600 kilos).
DISTRIBUTION OF CROP INCOME BY GENDER
There is a great deal of interest in the distribution of income
by gender.
It is difficult to talk very precisely about distribution of
cropping revenue by gender in Senegal because there are fields that are
said to belong to the household head (usually male) but that are clearly
farmed by all family members and whose products are consumed by all
family members --
these are the cereal fields grown under the
supervision of the household head.
Ignoring this difficulty, Table 5.5
assumes that the cropping income for each field goes to the person
supervising the cultivation of that field. The table shows the share of
gross income earned by gender for each zone". Although the crop
Table 5.5 -- Distribution of crop income by gender
Northern
Central Basin
Southwest Southeast
Senegal
Basin
Niakhar
Colobane
Basin
Basin
Oriental
Year
1989/90
Percent of total cropping income
Male
90
a5
86
88
84
68
Female
a
13
12
10
12
23
Source:
IFPRI/ISRA project survey.
Note:
When percents do not sum to 100, it is because there are some cases of incorrect or
missing codes for the gender of the person supervising the field.
" We have not yet had the time to examine in-depth factors that
are causing the returns to labor to be different across zones. It
seems important to determine whether differences are a result of
factors such as soil quality and rainfall or because patterns of labor
allocation differ from zone to zone.
" Numbers in Table 5.5 supersede those in the draft report which
incorrectly showed the share of cereal income rather than share of
total cropping income.
_’
*
.
.’
.
94
breakdown is not presented, over 90 percent of women's fields in the
Peanut Basin are peanut fields; cowpeas are usually the second most
common crop.
In Senegal Oriental, women cultivate rice and fonio in
addition to peanuts and cowpeas.
Chapter 6 discusses the share of different crops marketed by
gender.
Although Table 5.5 shows women's share in total crop production
to be quite small, a fairly large share of what women produce is
marketed, giving them a more important role in crop sales than one would
suspect from the production data alone.
CEREAL PRODUCTION SUFFICIENCY LEVELS
As calculation of cereal production sufficiency levels requires
only harvest data, this section of Chapter 5 presents results for both
HY88 and HY89.
In evaluating cereal production sufficiency we use a
level of 1900 kcal/AE as the standard requirement'*. Given fairly high
levels of pulse consumption in Senegal and the not insignificant role
that vegetable oil, milk and vegetables play in rural diets, we use 1900
kcal rather than minimum caloric requirements (2400 kcal) because it
better reflects real demand for cereals,
Table 5.6 illustrates that in 1988/89 cereal production fell very
far short of the 1900 kcal standard in all zones.
Production
sufficiently ratios13 were less than .5 in three zones and between .6
and .8 in the remaining three.
Table 5.6 -- Average cereal product ion sufficiency rat ios for 1988/89
and 1989/90 by zone
zone
Northern
Central Basin
southwest
Southeast
S e n e g a l
Peanut
N i a k h a r Colcbane Peanut
Peanut
Oriental
Basin
Basin
Basin
1988189
.I4
.47
.50
.5a
.a3
.70
1989190
.25
.76
.9a
1.03
1.02
.69
Source:
IFPRI/ISRA survey.
Note:
The cereal production sufficiency ratio is the amount of cereal produced by the household
divided by the amount required to provide 1900 kcal/AE for the entire household.
'* The draft version of this report used a variety of assumptions
for calculation production sufficiency levels. More detailed analysis
has convinced us that the 1900 kcal level is the most appropriate one
to use so references to other levels have been eliminated.
l3 The cereal production sufficiency ration (CPSR) is the
quantity of cereals produced divided by the quantity of cereal
required to provide each adult equivalent in the household with 1900
kcal of cereal.
.
:
.,
.
:.
.
.’
:
9 5
The same table shows that 1989/90 cereal production improved
greatly.
Three of the six zones had an average cereal production
sufficiency ratio (CPSR) in the neighborhood of 1 (Colobane in the
center and the two zones in the southern basin) while two others were in
the .6 to .8 range.
The northern basin was the only one still
exhibiting a very large deficit.
Table A5.5 in the appendix provides
details of CPSRs at different percentiles.
Table 5.7 shows estimates of the cereal surplus per AE averaged
across all sample households for each of the two study years. In 1988/89
amounts were insignificant except in the southeast (13 kilos/AE) and
Senegal Oriental (56 kilos/AE).
The situation improved in 1989/90 where
amounts available were in the 50 to 70 kilo range for three zones
(Colobane in the central basin and the two zones in the southern basin)
and in the 15-30 kilo range for two (Niakhar in the central basin and
Senegal Oriental).
Table 5.7 -- Average kilos of surplus cereals per AE:
1988/89 and 1989/90
ZONE:
NPB
CPBN
CPBC
SYPB
SEPB SO
198ala9
MEAN
~1
<l
1
2
13
5 6
1989/90
MEAN
~1 15
69
48
5 8
31
Source : IFPRI/ISRA survey.
Information is provided in appendix Table A5.6 for cereal surplus
available at the 25th, 50th, and 75th percentile.
The surplus in both
years was zero up through the 50th percentile -- i.e., at least half of
sample households had no surplus, even following the better 1989/90
harvest.
The bulk of surplus production is coming from a small set of
households.
Using two by two contingency tables and chi-square tests of
significance,
we attempted to identify factors that distinguish
households with large surpluses from others.
In general, the 18
households with surpluses in 1988/89 were more likely than others to
exhibit the following characteristics:14
- total cultivated area > 6 hectares
- cultivated area per AE > 1 hectare
- in the highest income category (tercile)
- earn greater than 40 percent of income from cropping
l4 "More likely" means that there was a statistically significant
difference -- usually at the .Ol level of significance or better but
always better than .lO..
96
- reside in Southeastern basin or Senegal Oriental
- reside in a nonmarket village
- rice provides less than 10 percent of cereal calories
Factors that did not seem to explain the level of surplus in
1988/89 were level of caloric adequacy, dependency ratios, and ethnic
group.
Following the better 1989/90 harvest 52 of the sample households
had a surplus.
These households were more likely to have the following
characteristics:
total cultivated area > 6 hectares
cultivated area per AE > 1 hectare
*
be in the middle and highest income category (tercile)
earn greater than 40 percent of income from cropping
*
reside in the southeastern basin and to a less extent in
the southwestern basin and the Colobane zone of the
central Peanut Basin
reside in a nonmarket village
*
rice provides less than 25 percent of cereal calories
*
achieve caloric adequacy levels (2400 kcal/AE)
An li*' marks those characteristics that changed between the two
years.
Caloric adequacy is a factor that becomes significant with more
than 65 percent of those with surplus consuming more than 2400 kcal/ae
per day.
Despite the significance, there are still 18 households with
surplus that do not appear to have adequate levels of caloric intake.
Income, rice consumption, and zone are factors that appear in both
years, but following the better harvest those with surpluses are less
concentrated in the extreme categories.
For example, higher income is
associated with surplus; but unlike 1988/89 those with surplus are found
in larger than expected numbers in both the middle and upper income
groups while those with no surplus are concentrated in the poorest
income category.
SUMMARY OF MAIN POINTS:
The important points to retain about crop production are:
e
Net cropping income fails to cover the costs of minimum househol‘d
needs in 3 of the 6 zones studied, even in a better than average
year.
In the zones where minimum needs were not covered (northern
Peanut Basin, Niakhar in the central Peanut Basin, and Senegal
97
Oriental), from 50 to 75 percent of households had cropping income
below the minimum needs level.
In 4 zones (the northern basin, Colobane in the central basin, the
southeastern basin and Senegal Oriental) than half of cropping
income comes from peanut production while in the 2 other zones
(Niakhar in the central basin and the southwestern Peanut Basin)
cereal income was greater.
Use of modern inputs is extremely low, even in zones where
households have access to confectionery peanut and cotton
contracts.
Rural markets for agricultural labor do not appear to be very
active; most cropping labor is supplied by household members or
through reciprocal arrangements with neighbors.
Despite relatively low use of modern inputs such as fertilizer,
input costs represent from 15 to 20 percent of gross cropping
income, depending on zone.
The bulk of input costs are for peanut
seed.
Returns to labor in peanut production are higher than returns in
cereal production in 5 of 6 zones; the difference is statistically
significant in 2 zones.
All zones but one produce more of the crop providing higher
returns to labor;
Households producing more than adequate levels of cereals tend to
have larger farms and more land per AE, specialize more in crop
production than other households, and be in the upper income
categories;
The policy implications of these relatively low levels of cropping
income and productivity were discussed in the previous chapter in our
analysis of income diversification patterns so they will not be
discussed here.
The links between crop production patterns and
marketing behavior are examined in the next chapter.
98
REFERENCES
Dakono, A. 1991.
Comparaison des revenus au capital et an travail dans
les activites aqricoles et non aqricoles au nord du bassin
arachidier.
mimeo, Dakar.
Kelly, V.A.
1988.
Factors affectinq the demand for fertilizer in
Seneqal's Peanut Basin.
Ph.D. dissertation, Michigan State
University, E. Lansing.
Kelly, V.A. 1993.
Aspects Economiques de la Production et de la
Commercialisation des Produits Aqricoles au Niveau dese Menaqes du
Bassin Arachidier et du Centre du Seneqal Oriental.
Paper
presented at IFPRI/ISRA Project Workshop, Dakar, March 1993.
Martin, F. 1988.
Food security and comparative advantaqe in Seneqal:
A micro-macro approach.
Ph.D. dissertation, Michigan State
University, E. Lansing.
.,
:
:
,.
‘.
9 9
6. ANALYSIS OF HOUSEHOLD AGRICULTURAL TRANSACTIONS
INTRODUCTION
This chapter examines the nature of sales, purchases, and gifts of
the principal agricultural products produced and consumed by households
in the IFPRI/ISRA survey.
The value of transactions in FCFA per AE are
examined by zone, harvest year,
and product.
The discussion is
organized around the following questions:
0
How much (in FCFA and kilos) of each product are households selling
and purchasing?
How does this differ by study zone and by year?
What does this imply for aggregate demand and supply in the study
zones?
0
Who are the household members selling and buying? What is their
gender? What is their status in the household?
l
Why are the transactions taking place? need for food? need for
cash? normal postharvest marketing?
l
Where and with whom are the transactions taking place? What does
this imply about the role of the marketing infrastructure?
One of the main objectives of this analysis is to understand the
factors that influence the quantities of cereals marketed by rural
households and the quantities of local and imported cereals purchased.
Following
an overview of transactions,
the rural zones are
discussed first and then the urban.
METHODOLOGY
The analysis is based on data collected in fortnightly interviews
during the entire survey period -- October 1988 - September 1990 for
rural zones and July 1990 - June 1991 for urban zones.
A full recall
technique was used by the interviewer -- all purchases and sales made
between two sequential interview periods were enumerated even if the
interview period was longer than two weeks.' The household head as well
1
Due to interviewer illness or nonavailability of household
members,
it was not always possible to adhere to a strict fortnightly
schedule.
” 1’
)’
..
.
.
..
100
as all other members of the family who made transactions were
interviewed directly.
Respondents were asked about the products, forms (whole grain,
flour), quantities, units (kilos, calabashes, bowls), unit prices, total
amount paid, nature of payment (cash, credit, in-kind), source of
financing, place of transaction, type of buyer/seller, and reason for
making the transaction.
Fall et al. (1989) provide details of the
survey methods used.
All of these variables except prices are dealt
with in this chapter; prices are discussed in Chapter 9.
Estimates of average FCFA per AE are calculated by using actual
prices paid or received.
Gifts are valued using transaction-derived
producer or consumer prices drawn from the survey data.'
Kilos
transacted are obtained by converting all cereals to grain equivalent
form.
Most transactions were made in grain form so there is not a great
deal of measurement error introduced by making these conversions.
Products covered on the sales side of the questionnaire were
cereals, pulses,
home-produced fruits and vegetables, and crop by-
products.
On the purchases side, only cereals and pulses are covered by
this analysis.
Purchased fruits and vegetables were enumerated less
frequently and with less detail.3
Crop by-products are purchased
primarily for animal feed and are analyzed as part of the animal
maintenance costs in Chapter 7.
The discussion in this chapter is based on detailed information in
the appendixes to this report.
Tables A6.1-A6.55 in Appendix 6.1
present the value of sales, purchases, and gifts for each rural zone and
village in FCFA per AE.
Tables A6.80 and A6.81 cover the urban zones.
Tables A6.56-A6.72 list the characteristics of the transactions (gender
and status in household of person making the transaction, location of
transaction,
source of financing,
source of product, reasons for
transactions, and so on). Appendix 6.2 provides some additional
information on particularities of each zone (for example, weather or
interviewer errors) that influence the patterns observed,
' Because of extremely thin producer price data in zones which had
very poor harvests, we have not used the usual procedure of valuing
gifts given at the price prevailing for the month of the transaction.
We have instead used the annual price (averaged across the entire 12
months of the survey year).
3 These data are being analyzed in conjunction with the demand
modeling.
There are no descriptive results to report at present.
.
.’
. .
101
Preliminary results of analyses performed on data for 1988/89 were
reported in Kelly, Reardon and McNeilly (1991). Results reported here
will differ from earlier versions because weighting for market villages
has been incorporated and changes in the number of adult equivalents per
household during the survey period have been taken into account.
Data on purchases of cereals and pulses in this chapter will differ
from that presented in Chapter 8.
The differences are due to the types
of questionnaires providing the data.
In Chapter 6 we use a full recall
of all purchases and sales.
In Chapter 8 we estimate kilos of different
products purchased and consumed by extrapolating from a 24 hour recall
of foods consumed that was conducted only two times in each month.
OVERVIEW OF CROP SALES AND PURCHASES BY SAMPLE HOUSEHOLDS
Relationship Between Sales and Purchases
In 1988/89, half of the zones spent more on purchases than they
received from sales.
The northern Peanut Basin represents an extreme
case where purchases were almost five times greater than sales.
In the
central Peanut Basin, purchases were double sales for Niakhar and about
25 percent greater in Colobane.
The implication is that crop revenues
alone were unable to cover the full extent of cereal and pulse purchases
made by households in these zones.
Households in the northern Peanut
Basin needed almost 20,000 FCFA per adult equivalent from other sources
of income while those in the central Peanut Basin needed about 5,000.
The 1988/89 situation for the remaining three zones was more
favorable, with the value of sales minus purchases ranging from about
2,500 FCFA/AE in the southwestern Peanut Basin to 9,000 in Senegal
Oriental and about 12,000 in the southeastern Basin.
Total value of
1988/89 sales was highest in the southeastern Basin (about 22,000
FCFA/AE) and Senegal Oriental (about 17,000).
It was lowest in the
Niakhar zone of the central Peanut Basin (2,500) and the northern Peanut
Basin (4,500).
Of the five rural zones retained for 1989/90, both Niakhar in the
central Peanut Basin and Senegal Oriental had negative net transactions,
but the amount was less than 2,000 FCFA/AE. The total value of 1989/90
sales was three times the previous year's level in the Colobane zone of
the central Basin and more than double that in the southwestern Basin
(Table 6.1).
In the southeastern Peanut Basin, the increase in sales
was about 50 percent.
Only in Senegal Oriental did the value of sales
decline in the second year.
This occurred primarily because farmers in
the zone boycotted cotton.
The implication here is that in the Colobane
zone of the central basin and in the two southern basin zones, crop
sales minus cereal/pulse purchases left households with 15-25,000
102
FCFA/AE to be used for repaying credit, purchasing consumer goods,
savings,
or investing in other productive activities; households in
Niakhar of the central basin and in Senegal Oriental needed to find
other sources of income to simply cover food purchases.
Product Composition of Sales and Sources of Supply
Table 6.1 shows that peanut production provides the largest share
of cash income coming from crop production for most rural households.4
The only exception is Senegal Oriental in 1988/89, where 88 percent of
sales revenue came from cotton.
In a show of their discontent about
high input costs and low (frequently negative) net returns to cotton
production,
the farmers in the Senegal Oriental sample completely
boycotted the 1989/90 cotton program.
The boycott led to a sharp
increase in peanut production and sales (91 percent of sales revenue)
and increased marketing of cereals (8 percent in 1989/90 versus only 1
percent the previous year).
Table 6.1 --
Share of sales revenue from sales of home production by zone,
harvest year, and crop purchased
Part A -- Harvest year 1988/89
North
Central Basin
Southwest
Southeast
Senegal
crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Peanuts
91’
98
91
9ab
88
13
Cotton
0
0
0
0
5
86
Cereals
0
2
a
2
7
1 ’
Cowpeas
9
0
0
0
0
0
Total value per
AE of all sales
4041
1819
10645
a552
22210
15309
4 Readers should not confuse cash income with total real income
from agricultural production.
The latter includes production consumed
by households as well as cash sales (see Chapter 4 and 5).
103
Part 5 -- Harvest year 1989/90
North
Central Basin
Southwest
Southeast
Senegal
crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Peanuts
n.a.
91d
87
94
86
91
Cotton
n.a.
0
0
0
7
0
Cereals
n.a.
9
12
6
7
8
Cowpeas
n.a.
0
0
0
0
1
Total value per
AE of all sales
n.a.
3020
32338
15298
34772
5428
Source: ISRAIIFPRI
projeci survey.
Note: ".a. mmn~ not available.
’ Fourteen percent in shelled form.
’ Five percent are confectionery peanuts.
’ local rice.
’ Five percent in shelled form.
As a share of crop income, cereals were most important in the
Colobane zone of the central Peanut Basin in both years (8 and 12
percent of sales revenue, respectively).
In 1988/89 the southeastern
Peanut Basin followed Colobane, with 7 percent of cropping income coming
from cereals.
In the following year, the Niakhar zone of the central
Basin was in second place (9 percent), while the higher potential zones
in the southwest, southeast, and Senegal Oriental drew 6-8 percent of
sales revenue from cereals.
The share of cereal sales in total gross receipts from crop sales
illustrates the degree to which households depend on cereal production
for their cash incomes, but it does not improve understanding of coarse
grain supply.
Table 6.2 shows that per adult equivalent, the largest
supplier (29 kilos per AE sold) in 1988 was the southeastern Basin,
followed by Colobane in the central basin (15 kilos); quantities
marketed by other zones were insignificant.
In 1989/90, the Colobane
zone provided the market with the greatest quantity (89 kilos per AE),
followed by the southeastern Peanut Basin (60 kilos) and then the
southwestern Basin (15 kilos).
The fact that the Colobane zone was an
important supplier of cereals both years is worth noting because it is
the north-central part of the Peanut Basin in an agroclimatic area that
is not considered very conducive to production.
During the survey
period, the zone was considered a deficit area by the "Commissariat de
Securite Alimentaire" (CSA).
:
.’
:
”
I
:
104
Given the different population densities of the major cereal
marketing zones,
we extrapolated from survey households to the
production zones in general in Table 6.3 to show the total quantities of
marketed cereals that probably came from each study zone during the
survey period.
Table 6.2 -- Volume of cereal sales by zone and year
Harvest
North
Central Basin
Southwest
Southeast
Senegal
Year
Basin
Niakhar
Colobane
Basin
Basin
Oriental
(kilos per AE)
1988/89
0
1
15
3
29
1
1989/90
n.a.
5
a9
15
60
a
Source:
ISRA/IFPRI project survey.
Note:
n.a. means not available.
Table 6.3 -- Estimated volume of coarse grain sales by zone and year
Harvest
North
Central Basin
Southwest
Southeast
Senegal
Year
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(metric tons per zone)
1988/89
0
287
11362
886
3137
63
i989/90
n.a.
1437
67415
4432
6450
507
Source:
ISRA/IFPRI project survey.
Note:
n.a. means not available.
The Colobane zone of the central Peanut Basin is clearly the
biggest supplier of marketed cereals in the survey. Estimated marketing
from Colobane is almost four times greater than the next most important
zone in 1988/89 and ten times greater in 1989/90.
Product Composition of Purchases and Sources of Demand
Table 6.4 shows the share of purchases for home consumption by
zone, harvest year, and product.
Of greatest interest are the purchases
of rice versus coarse grains and the extent to which these differ by
zone and year.
.
;
:
:
:
s
‘.
105
Table 6.4 -- Share of purchases for home consumption by zone, harvest
year, and product
Part A -- Harvest vear 1988/89
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent of total FCFA/AE purchased)
Millet/sorghum
23
32
29
23
42
23
Maize
4
4
2
20
Imported rice
67
63
67
77
45
40
Cowpeas
3
Peanuts
3
1
3
13
1 1
Part B -- Harvest year 1989/90
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent of total FCFA/AE purchased)
Millet/sorghum
n.a.
23
29
7
46
22
Maize
n.a.
1
1
36
Imported rice
n.a.
71
6 4
91
47
3 4
Cowpeas
n.a.
1
Peanuts
n.a.
5
6
1
6
1
Source:
ISRA/IFPRI project survey.
Note: n.a. means not available.
In all zones but Senegal Oriental (1989/90), imported rice accounts
for the largest single share of expenditure on cereals and pulses,
representing 60 to 90 percent of purchases in the north, center, and
southwest and 30 to 45 percent in the southeast and Senegal Oriental.
In four zones (Niakhar in the center, both zones in the southern basin,
and Senegal
Oriental') the share of rice in purchases
increases
following the better harvest, suggesting that increased income and a
desire for diversity in the cereal diet are influencing demand. In the
Colobane zone of the central Basin, where the relative share of rice
declines following the better harvest, the drop is only 3 percent.
5 The better harvest for the first three zones was 1989/90; in
Senegal Oriental it was 1988/89.
.r
‘
.
.!
106
Purchases of pulses account for a relatively small share of
expenditures (less than 6 percent) except in 1988/89 when peanuts
accounted for 13 percent of purchases in the southeastern Basin and 11
percent in Senegal Oriental.
Cowpeas are purchased in such small
quantities that they are an insignificant share of expenditures in all
zones but the north (3 percent).
Table 6.5 presents data on cereal purchases in terms of kilos
purchased per adult equivalent.
With the exception of the northern
Peanut Basin in 1988/89, Colobane exhibits the largest cereal purchases
per adult equivalent for both years.
This is surprising given that the
zone also has one of the largest volumes of cereal sales in both years.
The production sufficiency ratio of 0.83 for Colobane in 1989/90
(Chapter 5) also suggests that the zone should not be purchasing 174
kilos of grain per adult equivalent in the same year. The possibility
of an error in data entry or the inclusion of commercial transactions
with household consumption transactions is being investigated. Assuming
there are no data errors, the implication is that households in the zone
are selling coarse grains and purchasing imported rice for their own
consumption.
Table 6.5 -- Volume of cereals purchased by zone, year, and product
North
Central Basin
Southwest
Southeast
Senegal
Year/Product
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(kilos per AE)
1988/89
Rice'
89
20
46
35
22
15
Coarse grains
52
17
41
16
38
28
Totalb
178
45
109
66
69
49
1989/90
Rice'
n.a.
22
60
43
23
16
Coarse grains
n.a.
10
87
7
64
52
Totalb
n.a.
41
173
68
97
75
Source: ISRA/IFPRI project survey.
Note:
n.a. means not available.
' Kilos of processed, imported rice.
' Rice transformed to whole grain equivalent using a coefficient of 1.43,
which implies a 70 percent processing loss.
Table 6.6 compares the level and composition of purchases in market
and nonmarket villages.
In both survey years, market villages spent
more per adult equivalent on cereal/pulse purchases than nonmarkqt
villages,
irrespective
of the zone.
The differences were more
,*
:
I
.’
.
107
pronounced
in the central Peanut Basin for both years,
in the
southeaste rn Basin for 1989/90, and in Senegal Oriental for 1988/89.
Table 6.6 -- Share of purchases for home consumption by zone, harvest
year, and product
Part A -- Market villages 1988/&X9
North
Central Basin
Southwest
Southeast
Senegal
crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
19
28
27
14
49
12b
Maize
1
7
1
22
Imported rice
70
63
67
86
48’
57
Cowpeas
5
Part B -- Nonmarket villages 1988/89
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
24
35
26
23
40
36
Maize
5
1
2
21
Imported rice
67
64
70
78
45
30
Cowpeas
3
Peanuts
3
2
15
13
:
,’
,.‘,
_.
‘;
.
I
.
,:
.
:
.
. .
.
108
Table 6.6 -- Continued
Part C -- Market villages
1989/90
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
n.a.
12 :
16
17
49
gb
Maize
n.a.
4
3
1
1
2 1
Imported rice
n.a.
a3
7 6
8 0
4 6
6 0
Cowpeas
n.a.
1
Peanuts
n.a.
2
4
2
4
9
Part D -- Nonmarket villages 1989/9o
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
n.a.
31
2 2
1 2
4 5
2 8
Maize
n.a.
3 3
Imported rice
n.a.
6 4
68
93
48
3 1
Cowpeas
n.a.
Peanuts
n.a.
6
9
1
5
9
Source: ISRAIIFPRI project survey.
Note: n.a. meene no1 available.
' One percent local rice included
b
All sorghum.
The only large differences in the share of rice purchased in
1988/89 occurred in the southwestern Peanut Basin and Senegal Oriental,
where the share of rice for market village households was 8 and 27
percent more than that in nonmarket villages.
In 1989/90, market
villages in three zones (Niakhar and Colobane in the central basin, and
Senegal Oriental) purchased a larger share of rice than the other
villages while the nonmarket villages in the two southern zones
purchased more rice than their market village neighbors.
Table 6.7 presents a picture of how total coarse grain demand for
the study zones looks when the population density of each zone is taken
into account.
Demand in Colobane alone is much greater than the sum of
demand across all other zones.
109
Table 6.7 -- Estimated volume of coarse grain purchases by zone and year
Harvest
North
Central Basin
Southwest
Southeast
Senegal
Year
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(metric tons per zone)
1988/89
10881
4887
31056
4727
4111
1775
1989/90
n.a.
2875
65900
2068
6923
3297
Source:
ISRA/IFPRI project survey.
Note:
n-a. means not available.
In comparing results from Table 6.3 with results from Table 6.7, it
is apparent that the Colobane zone is the greatest supplier and the
greatest demander of coarse grains.
Table 6.8 uses the estimated total
demand and supply data from Tables 6.3 and 6.7 to calculate net market
supply (supply minus demand) for each zone.
In harvest year 1988/89
demand exceeded supply in the zones covered by 41,703 tons; not a single
zone marketed more coarse grains than it purchased.
In harvest year
1989/90, the Colobane zone in the central basin and the southwestern
basin marketed 3,879 tons more than they purchased. In all other zones
purchases exceeded marketing.
The net effect across all survey zones
was a positive supply of 1,726 tons of coarse grains which would have
been available to other zones or urban areas.
Table 6.8 -- Estimated net volume of coarse grain transactions by zone
and year (quantity marketed minus quantity purchased)
Harvest
North
Central Basin
Southwest
Southeast
Senegal
Year
Basin
Naikhar
Colobane
Basin
Basin
Oriental
Total
(kilos)
1988189
-10881
-4600
-19694
-3841
-975
-1712
-41703
1989/90
n.a.
-1441
+1515
+2364
-433
-2790
+1726
Source:
ISRA/IFPRI project survey.
Note:
n.a. means not available.
Table 6.9 summarizes the value of seed purchases per adult
equivalent (175 to 5408 FCFA).
The Colobane zone of the central Peanut
Basin spends more per adult equivalent than other zones. Peanut seed is
clearly the dominant item, accounting for 91 to 300 percent of all seed
purchases.
It is informative to note that cowpeas are a larger share of
seed purchases than cereals despite the very small role cowpeas play in
”
.i
.
.
,’
110
total agricultural production.
This is most likely due to the
difficulty in storing cowpeas.
Table 6.9 -- Share and total value of seed purchases by zone, year, and
product
Part A -- Harvest year 1988/89
%
1
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Cowpeas
2
0
0.5
0
0
1
Cereals
1
1
0.5
0
0
0
Peanuts
97
99
99
100
100
99
Total value FCFA/AE
1590
400
5408
233
1486
621
Part B -- Harvest year 1989/90
North
Central Basin
Southwest
Southeast
Senegal
Crop
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Cowpeas
n.a.
1
1
0
0
3
Cereals
n.a.
0
0
0
0
6
Peanuts
n.a.
99
98
99
100
9 1
Total value FCFA/AE
n.a.
340
2547
571
1661
175
Source:
ISRA/IFPRI project survey.
Note:
n.a. means not available.
Role of Gifts in Total Value of Crop Transaction
It is sometimes argued that traditional mechanisms for transferring
cereals from surplus to deficit households provide rural areas with a
food security buffer that functions as well as any government program
might be expected to function.
Data on gifts given and received
presented in Tables 6.10 and 6.11 show that transfers do occur but that
they are a very small share of total transactions. The total value per
adult equivalent of gifts given ranges from 67 to 3446 FCFA across zones
. . . ,,
,,:
.
.’
.
.
:
;
111
and years.
For gifts received, the range is 28 to 2203 FCFA per adult
equivalent.6
Table 6.10 -- Share and total value of gifts out by zone, year, and
product
Part A -- Harvest vear 1988/89
North
Central Basin
Southwest
Southeast
Senegal
Product
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
27
88
68
99
86
51
Maize
0
0
0
0
1
26
Imported rice
4
2
2
0
2
18’
Cotton
0
0
0
0
0
Cowpeas
14
0
0
0
0
Peanuts
56
10
31
1
11
4
Total value FCFA/AE
of all gifts given
516
194
1237
1394
1887
1326
Part B -- Harvest year 1989/90
North
Central Basin
Southwest
Southeast
Senegal
Product
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
n.a.
100
56
98
91
63
Maize
n.a.
0
0
0
2
24
Imported rice
n.a.
0
0
1
4
6b
Cotton
n.a.
0
0
0
0
0
Cowpeas
n.a.
0
0
0
0
0
Peanuts
n.a.
0
44
1
3
a
Total value FCFA/AE
of all gifts given
n.a.
67
407
1749
3446
704
Source:
ISRA/IFPRI project survey.
Note: n.a. means not available.
' Includes 7 percent local rice.
b Includes 4 percent local rice.
6 The estimates of net transfers presented in Chapter 4 did not
consider the value of in-kind transfers.
Adding values reported in
Tables 6.7 and 6.8 to those in the "transfer" column of Table 4.5
provides a more complete picture of net transfer income.
-
1
I’.
.
‘
.’
.
,,
:
.
.
*.
:
112
Table 6.11 -- Share and total value of gifts received by zone, year, and
product
Part A -- Harvest year 1988/89
North
Central Basin
Southwest
Southeast
Senegal
Product
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
,,,
10
3 5
72
88
8 8
2 4
Maize
3
0
0
0
0
2 9
Imported rice
42
59
21
1 2
9
22'
Cotton
0
0
0
0
0
0
Cowpeas
26
0
1
0
0
1
Peanuts
19
5
6
0
3
23
Total value FCFA/AE
of all gifts given
9 9
28
174
1 4
756
716
Part 6 -- Harvest year 1989/90
North
Central Basin
Southwest
Southeast
Senegal
Product
Basin
Naikhar
Colobane
Basin
Basin
Oriental
(percent)
Millet/sorghum
n.a.
7 2
3 1
4 1
46
7
Maize
n.a.
0
0
0
0
3 1
Imported rice
n.a.
2 1
36
45
5
52’
Cotton
n.a.
0
0
0
0
0
Cowpeas
n.a.
1
2
0
0
4
Peanuts
n.a.
6
3 2
1 3
49
5
Total value FCFA/AE
of all gifts given
n.a.
174
208
241
2203
1188
Source:
ISRA/IFPRI project survey.
Note: n.a. means not available.
' Includes 1 percent local rice.
b The large share of peanuts may be due to a case where in-kind payments
for assisting neighbors with their peanut harvest were coded as gifts.
Gifts received by rural households are composed to a large extent
of imported rice and gifts given are primarily coarse grains.
These
results provide some basis for inferring that gifts are exchanged
between rural (giving coarse grains and receiving rice) and urban
(giving rice and receiving coarse grains) households.
.
.
I
.’
.(
.’
:
113
CHARACTERISTICS OF CROP SALES TRANSACTIONS
Identity of Household Members Sellinq
Figures 6.1A-E shows the share by gender of agricultural products
sold.
The shaded part of the bars shows the share of total sales made
by women.
The left axis shows the total value per AE of all sales of
the product concerned,
helping one keep in perspective the relative
monetary value of each crop.
Figure 6.1
Relative shares of sales by gender for (A) millet, (6) peanuts, (C) cotton, (0) cowpeas,
and (E) other products for harvest years 1988/89 and 1989/90. Key: SAG, Sagatta; NIA, Niakhar; COL,
Colobane: PAS, Passy; 010, Nganda-Dioly; MIS, Missirah.
Part A
Y
-2500-
F
\\
v-l 2000 -
w
2
IT
crl rsooj
-
11 II
t;
j lOOO-
2
II
500 -
rl n ! NA
i
I
I
I
I
I
I
1
3HU IYIh LUL I ,-Ta U,”
MIS
SAG NIA COLPAS DIO
1988/89
1989/90
0 MALE
m FEMALE:
114
Part B
35 (
SAG NIA COL PAS DIO iIS
SAG NIA COL PAS
1988/89
1989,'90
0 MALE m FEMALE
Part C
1988/89
1989,'90
0 MALE @FEMALE
115
Part D
400
NA
I I
I
I
I
I
I
I
m
I
1
SAG NIA COI PAS DIO MIS
I
I
SAG NIA COL PAS DIO MIS
1988/89
1989/90
0 MALE @ FEMALE
Part E
0.5
20.45
g 0.4
E 0.35
\\
z 0.3
-J
2 0.25
c
SAG NIA COL PAS DIO MIS
SAG NIA COL PAS DIO MIS
1988/89
1989/90
0 MALE m FEMALE
,’
.
::.
._
.,
;
_‘.
.
.’
.’
.
::’
.
.
116
Crop sales are dominated by males.
Women are more active in peanut
sales than in other crops.
They accounted for 14-17 percent of peanut
sales in the central, southwestern and southeastern Peanut Basin in
1988/89 and 14 to 29 percent in 1989/90. Women's share in the northern
Peanut Basin was unusually low (4 percent) and unusually high in Senegal
Oriental (33 percent in 1988/89 and 29 percent in 1989/90). Women's
share of peanut sales may have increased in 1989/90 because household
heads were obljged to concentrate on cereal production after the very
poor 1988/89 harvest.
These results suggest that the tendency of women to shift to cereal
crops immediately after the government seed distribution program was
dropped in the mid-1980s (see Ndoye, unpublished) has not been
sustained.
As time has passed, women have found the means to acquire
peanut seed and are now back in the peanut business.
A comparison of coarse grain sales and production data (Chapter 5)
for women suggests that many are selling small amounts of cereal
received from the household head for daily meal preparation and using
the receipts to purchase condiments.
This is particularly true in the
southwestern Peanut Basin where the women felt free enough to reveal
their "theft" of household cereals to interviewers. It probably occurs
in other zones to an extent greater than that revealed in the data.7
Women dominated cowpea sales in the central basin -- Niakhar (100
percent) and Colobane (50 percent) in 1989/90. In the northern Basin,
where cowpeas have been promoted as a cash crop alternative to peanuts,
men dominated but women had a 13 percent share.
Women had an important share of sales revenue from "other' products
in the central basin zone of Colobane, in the southeastern basin and in
Senegal Oriental in 1988/89 and in both zones of the central basin in
1989/90.
Only in Colobane in 1989/90, however, was the value of sales
of any importance.
Products sold were sorrel and gumbo.
In most zones and years, the household head sells 50 percent or
more of the principal cereal and cash crops.
Reasons for Sellinq
Because 50-95 percent of peanuts are usually sold during the
official marketing campaign, respondents did not usually elaborate on
reasons for selling.
In 1988/90 the Colobane zone in the central Peanut
Basin appears to have sold only 20 percent during the official campaign
(these data are being checked for entry errors because the table
concerning partners shows 71 percent of sales made to institutional
purchasers).
Since Colobane is a zone where it is common to process
7
The zone of Colobane in the central basin presents an exception
because women did produce most of the cereals they marketed in this
zone.
‘.
.:
.
.
.
i
I
:
.
:
%.
‘.
117
peanuts to produce oil, it would not be surprising to find a large share
of peanut production marketed unofficially, but 80 percent seems too
high.
The second most important reason for selling peanuts is to
purchase food.
Cereals are sold primarily to purchase other foods, finance
ceremonies,
or finance other economic activities.
Each zone and year
has its own particularities and there are no clear trends.
Location of Sales
Most peanuts appear to be sold to official marketing agents (71 to
100 percent,
with Colobane at the lower end).
Senegal Oriental
presented an exception to this rule in 1989/90 when a project by a
nongovernmental organization appears to have purchased about 90 percent
of the sales.
Local private traders are important as a secondary
purchaser in the central Peanut Basin, as are neighbors.
In the
northern Peanut Basin, private traders from outside the immediate area
were the most important secondary purchasers.
The type of persons to whom cereals were sold changed in some zones
between the two survey years, probably due to the increase in cereal
production in 1989/90.
In the Niakhar zone of the central basin, local
traders became more important in the second year, diminishing the share
of interhousehold transactions.
In the Colobane zone of the central
basin, traders from outside the village became more important as the
share of village traders declined.
In the southwest the share of
village traders declined as the share of local ambulant traders
increased.
Traders from outside the village increased in importance in
the southeastern basin while interhousehold transactions continued to
account for 25 percent of transactions.
In Senegal Oriental, most
cereal transactions in both years were interhousehold, with about 75
percent taking place in the interviewer's own village and the rest in
neighboring villages.
The market village was the most important location for sales of
cereals in the central Peanut Basin (Niakhar), the southwestern basin,
and the southeastern basin.
In Colobane, the arrondissement is most
important, but the arrondissement is also the site of the weekly market
for Colobane.
In Senegal Oriental most cereal sales are made in the
village of the seller, but cowpea sales are concentrated in the market
village.
Given the importance of market villages in marketing cereals,
programs to increase the number of market villages in zones which are
not well endowed (the southeast and Senegal Oriental) should encourage
greater cereal sales.
The locations of peanut sales are varied and determined by the
location of the official marketing services. In most cases, these are
in nearby villages or rural communities.
118
CHARACTERISTICS OF PURCHASE TRANSACTIONS
Identitv of Household Purchasers
Figure 6.2 A-E presents a breakdown of purchases by gender. Males
dominate cereal purchases, as is the tradition throughout Senegal.
Dependent household heads and wives of the household head are the other
major categories of cereal purchasers.
Women's participation was
greatest
in the northern Peanut Basin, where the harvest was
particularly bad i% 1988/89.
There appears to be a somewhat greater
female contribution to rice purchases than to coarse grains.
These
could be cases of women too busy to prepare the usual coarse grain meals
resorting to buying rice to shorten preparation time.
Figure 6.2
Relative shares of purchases by gender for (A) millet, (6) peanuts, (C) cotton, (D)
cowpeas,
and (E) other products for harvest years 1988/89 and 1989/90.Key:
SAG, Sagatta; NIA,
Niakhar; COL, Colobane; PAS, Passy; DIO. Nganda-Dioly; MIS, Missirah.
Part A
4500
2. 4000
2 3500
L
F 3000
\\
g 2500
LA
; 2000
5
& 1500
L3 1000
5 500
0 T
,
I
4
8
SAG MIA 'COL PAS DIO MIS
SAG NIA COL PAS DIO MIS
1988/89
1989/90
n MALE m FEMALE
119
Part B
NA
I
I
I
n
-
-
I
1
I
I
I
I
I
I
SAG NIA COL PAS DIO MIS
SAG NIA COL PAS DIO MIS
1988/89
989/90
0 MALE m FEMALE
Part C
14
SAN NIA COL PAS DIO MIS
SAG NIA COL PAS DIO MIS
1988/89
1989/90
n MALE m FEMALE
.‘_
I
:
.’
.
.
.
120
Part D
rfY
SAG NIA COI PAS DI( MIS
SAG NIA COL PAS DIO MIS
i988/89
i989/90
u MALE m FEMALE
Part E
600CI--
2
b
b5OOC )-
5
>4000 I -
t;f
IQ0 3000
5
a
r
g 2000
ii
2 1000
n
s
k!
NA
0 t-
J
l-l
I
I
1
I
1
SAG NIA COL PAS DIO MIS
SAG NIA CDL PAS DIO MIS
1988/89
1989/90
(MALE m FEMALE
1
‘.
.
+
,
.
:.
.
:’
:
.
..
.
*
.
:
121
Peanuts are purchased primarily by household heads and their wives
and dependent household heads.
Peanuts for consumption are the only
product where women's share is noticeable, although this is not true in
all zones.
Women in both zones of the central basin and in Senegal
Oriental generally purchased more peanuts than women in the southern
Peanut Basin.
This behavior reflects the traditional obligation that
women provide ingredients for the sauce.
Reasons Products Are Purchased
Most cereals are purchased for home consumption. In the northern
Peanut Basin and Senegal Oriental, the pattern of buying and stocking
rather than purchasing for immediate consumption is more common than in
other zones.
During the difficult 1988/89 harvest year, households in
the central basin did not purchase and stock but they did following the
better harvest in 1989/90.
Peanut purchases are primarily for seed (more than 50 percent of
all peanut purchases except in Senegal Oriental, where seed accounts for
30 to 50 percent).
Sources of Financinq for Purchases
Money used for cereal purchases comes from a wide variety of
sources.
It is clear that income from the sale of cash crops
contributes only a very small share to cereal purchases across all zones
(less than 20 percent). Wages, migration, commerce, and livestock are
the primary sources of income used to purchase cereals.
Livestock is
more important (26 percent) in the southeastern basin and Senegal
Oriental while wages and commerce dominate in other zones. Migration
income is the biggest single source of cereal money in Senegal Oriental.
Location and Sources of Purchases
Cereals are usually bought from traders; Senegal Oriental is the
exception where interhousehold transactions are more common. The market
village is the most common place to purchase cereals; Senegal Oriental
is again the exception, with nearby villages playing an important role.
AGRICULTURAL PRODUCT TRANSACTIONS IN URBAN ZONES
Town of Kaolack
Tables A6.55a-c show the average amount (in FCFA) of purchases,
sales
and transfers by adult equivalent of agricultural products
undertaken in Kaolack by households in the sample during the period from
July 1990 to June 1991.
Level of transactions.
The global level of agricultural products
purchased for all uses is 18,383 FCFA/AE.
This is above the level
registered in the rural zones.
The value of the sales, as anticipated,
is very small:
188 FCFA/AE.
.
.
;.
‘_
.’
122
As predicted,
transactions for home consumption are far more
important than purchases for trade or seed.
In fact, purchases of
coarse grain, for home consumption,
represent fourteen times the value
of those undertaken for trade reasons:
17,206 and 1,178 FCFA. The
outflows are primarily gifts (90 percent) with some sales (10 percent).
Composition'bv product.
Most of the purchases are imported rice. This
product represents 71 percent of the total value of cereal/pulse and 73
percent of cereal purchases and gifts received.
Millet, the second cereal consumed in Kaolack represents 19 percent
of cereal/pulse purchases and almost half of gifts received or made by
households.
Originating from harvests of relatives living in the
countryside, millet gifts reflect the links between the rural and urban
zones.
Locally grown products make up two thirds of gifts received;
the rest is imported rice.
The importance of household purchases of cowpeas is relatively
small compared to peanuts.
Of all purchases and gifts received, peanuts
follow rice and millet in importance.
In general, peanuts are bought
shelled but are given unshelled.
Finally,
maize makes up
a marginal
share of agricultural
transactions, especially in the form of purchases. The other products
such as sorghum, fonio and cotton are not found at all.
Commercial transactions are made up in large part of purchases of
millet for processing and resale.
The sale of millet cereal, especially
for breakfast, is a common activity in Kaolack.
Rice, vegetables,
cowpea and peanuts are also purchased for resale.
It is important to
note that these products purchased for trade are transformed (prepared
for meals) before they are resold.
Hence, the importance of their
resale doesn't appear in the sales which include only unprepared
products.
Seed transactions are non existant at the level of households being
studied because urban households in the Kaolack sample do not farm.
Town of Tambacounda
The average value (in FCFA/AE) of agricultural purchases, sales,
and gifts received and given by households in the Tamba sample from July
1990 to June 1991 is presented in tables A6.56a-c.
Level of transactions.
The total value of all agricultural products
purchased or received as gifts is 19,692 FCFA/AE.
It is higher than
that observed for Kaolack.
The very low value of sales and gifts given
(275 FCFA/AE) exceeds that of Kaolack.
Here, as in Kaolack, transactions for home consumption dominate the
transactions for trade and seed.
The inflows of coarse grains, for
consumption,
account for 19,626 FCFA/AE while only 41 FCFA are
..<I
.*
.J
.
.
:
.’
123
attributed to trade and 28 FCFA to seed. The outflows are made up, as
in Kaolack, of gifts.
There are no other reasons for grain outflows.
Composition b.y product. The product composition of transactions is very
diversified in Tamba. Almost all products, with the exception of fonio,
cotton, and local rice, play a part in household transactions.
Imported rice dominates transactions:
75 percent of the total
value of inflows of agricultural products for Tamba households. In
addition, rice makes up 78 percent of household cereal purchases+
Rice
also accounts for a few gifts received by households (only 6 percent)
but constitutes 24 percent of gifts made by these same households.
Maize, sorghum, and then millet are much less important than rice
in household purchases.
Their individual share is about 6 to 7 percent.
On the other hand, more than half of gifts received by households are
maize,
followed by sorghum.
Gifts made are exclusively cereals,
especially traditional local cereals.
This larger diversification of
gifts received or given can be explained by the fact that there are
relatively more households active in agriculture in Tamba than in
Kaolack.
Peanuts come in first place among pulses, but far behind
cereals as a share of purchases.
Commercial transactions are essentially millet purchases (94
percent).
As in Kaolack, the sale of millet cereal for breakfast can
explain the predominant place of millet.
Seed purchases (primarily
peanuts) also take place, even though at a low level.
SUMMARY OF FACTORS INFLUENCING THE QUANTITY OF CEREALS MARKETED
One of the major objectives of the IFPRI/ISRA study is to
understand the factors that influence the quantity of cereals marketed
by rural households in Senegal.
Economic theory tells us that producer
price should be one of the most important variables. This chapter deals
with factors other than price, leaving the discussion of price for
Chapter 9.
At this stage of the analysis we have simply noted correlations
between large quantities marketed and particular zones, types of
households (market or nonmarket village location), and types of
individuals within households (gender and status). The discussion ends
with presentation of results of chi square tests used to identify
characteristics distinguishing households selling large quantitities of
cereals from other households.
Crop production levels and the production sufficiency ratios
explain a great deal about cereal marketing behavior.
In general, zones
with the better production sufficiency ratios marketed a larger share of
their cereals.
124
Cereal marketinq ratios
Tables 6.12 and 6.13 present the ratio of cereals and peanuts
marketed to total production by zone and year.
Data on marketing are
available for harvest years 1988/89 and 1989/90 only. A comparison of
the two tables makes the point that peanuts play a much more important
commercial role than cereals.
Table 6.12 -- Share of total cereal production marketed by zone and year
Northern
Central Basin
Southwest Southeast Senegal
Basin
Niakhar Colobane Basin
Basin
Oriental
Year
1988189
0.03
0.08
0.14
0.16
0.21
0.18
1989/90
n. av.
0.04
0.12
0.23
0.28
0.15
Table 6.13 -- Share of total peanut production marketed by zone and year
Northern
Central Basin
Southwest Southeast Senegal
Basin
Niakhar Colobane Basin
Basin
Oriental
Year
1988/89
0.62
0.30
0.62
0.34
0.72
0.21
1989/90
n. av.
0.16
0.72
0.49
0.68
0.39
In general, one notes a correspondence between the propensity to
market and the higher CPSRs.
The only notable exception is Senegal
Oriental, which exhibited the lowest CPSR in 1989/90 (except for the
northern basin where deficits are chronic) but a relatively large share
of cereals sold (15 percent).
This behavior, coupled with the low
levels of caloric intake reported in Chapter 8, prompts one to ask if
consumption levels are not being pushed lower by households selling
cereals from stocks that were inadequate from the beginning.
The correlation between CPSRs and marketing behavior underscores
that importance of increasing cereal productivity as a means of
increasing quantities marketed.
Productivity was extremely low in 1988/89 when rural households in
the zones surveyed purchased almost 42,000 tons more coarse grains than
they sold.
Following the better-than-average harvest of 1989/90, the
net supply of cereals (sales minus purchases) in the zones covered is
estimated to be only 1726 tons.
125
Peanuts remain the major source of cash income in rural areas.
When peanut income is adequate, households are less likely to sell
cereals. The higher 1989/90 cereal sales in Senegal Oriental -- despite
a relatively low production sufficiency ratio -- are partly due to a
drop in cash crop revenue following the cotton boycott.
The Colobane zone of the central Peanut Basin marketed the greatest
quantity of cereal in both years (Table 6.3). Much of this zone falls
in areas of unreliable rainfall and rapidly deteriorating soils,
therefore it,is surprising that the production sufficiency ratio was so
favorable.
The zone is characterized by a very active commercial sector with
several weekly markets in close proximity (Colobane, Mbar, Gnibi) and
the Mecca of Peanut Basin commerce--Touba--not very far away.
Many
households in this zone paid attention to price information coming from
Touba, and sold their products in Touba when prices there were more
favorable than locally.
Two of the three villages in the zone are located on a paved road;
the third village is about 7 kilometers from the paved road via a very
sandy road not conducive to truck traffic.
Dispite the disparity in
road infrastructure, the village with more difficult access sold a
larger share of its cereal production and about the same share of its
peanut production as the other nonmarket village in the zone. We have
not analyzed sales patterns (to whom and where sold) on a village-by-
village basis. For the zone as a whole, 72 percent of cereals were sold
to a village trader.
It would be useful to know if households in the
more isolated village were transporting their production to market or if
the commercial network in the zone was present at the farmgate.
Household heads are clearly the most important marketers of cereals
in all zones.
Women in the Colobane zone, however, played an unusually
large role in the marketing (37 and 27 percent of sales, depending on
the year).
A quick review of crop production data for Colobane reveals
that the wives of household heads produced about 3600 kilos of cereals
during 1989/90--this is approximately the amount marketed by them. The
implication here is that more attention to women in extension and
marketing interventions might bring about a greater marketed surplus
than one gets when dealing with men who have the obligation of providing
household cereals.
The extent to which this would be true in other zones remains to be
examined.
Women in the southwestern Peanut Basin also sold a large
share of cereals (45 percent) in 1989/90; however the crop production
analysis shows only 103 kilos of cereals produced by women (less than 3
percent of total marketings).
This is a case where they were clearly
selling millet from the daily food allotment given to them by the
household head.
126
Table 6.14 shows the distribution of households by quantities of
cereals marketed in both 1988/89 and 1989/90. Households selling more
than 500 kilos almost doubled between the two years. The tendency was
for those in the 50 to 200 kilo range in 1988/89 to move into the higher
up categories.
We find very little change in households that were
selling less than 50 kilos; approximately the same number of households
are found in this category both years.
In general, they tend to be the
same households,
suggesting that something is preventing these
households from reaping the commercial benefits of the better rainfall
in the same way that the bigger sellers do.
We suspect that these
households are ones that do not significantly increase levels of cereal
production during the year of better rain; more work needs to be done
here to identify the exact nature of the problem.
Table 6.14 -- Distribution of households by kilos of cereals marketed in
1988/89 and 1989/90
Category
Number of households
Number of households
1988/89
1989/90
no sales
12
10
<50
25
2 7
<200
5 7
36
<500
3 1
37
>50
1 6
3 1
Source:
IFPRI/ISRA survey data.
The entire sample had 47 households that were selling more than 200
kilos of cereals per adult equivalent in 1988/89 and 68 in 1989/90.
Again using two by two contingency tables and chi-square significance
tests we looked for characteristics that differentiated these big
sellers from other households.
We found that big sellers were more
likely to:
- be in the southeastern basin in 1988/89 and in both southern
basin zones in 1989/90
- be living in nonmarket villages
- be Wolof or Fulani
- be in the upper income categories
- earn more than 40 % of total household income from
cropping activities
- have CPSRs >.75
- cultivate > 6 hectares and > 1 hectar/ae
- derive ~10 percent of cereal calories from imported rice
,
:
.
.
‘.
127
Although big sellers are more likely to exhibit the above set of
characteristics,
it is important to note that there are numerous
exceptions to the general tendency.
It is of interest that some large
sellers are in the low income categories
(13 percent in 1988/89 but
only 4 percent in 1989/90) and also ranked as food insecure households
(57 percent in 1988/89 and 40 percent in 1989/90).* These food insecure
households warrant further investigation to determine why they are
selling large quantities of cereals rather than retaining production for
their own consumption.
Somewhat disconcerting is the fact that Senegal Oriental had the
largest marketable surplus in 1988/89 but was not among the zones where
we found a larger than expected share of big cereal sales.
Is this due
to lack of demand for their production (primarily maize and sorghum) by
other zones? Transport costs that make their production too expensive
to market in other zones?
Access to land appears to be an important factor in encouraging
cereal sales and production.
The effect of farm size being greater than
6 hectares appears to be more important than farm size per AE.
Proximity to market infrastructure is normally expected to
encourage
increased crop production and marketing as
it reduces
transportation and handling costs for obtaining inputs and selling
production.
In the rural areas surveyed, however, proximity to
marketing infrastructure is not serving as an impetus to greater crop
production and marketing.
This behavior is in sharp contrast to the
very important influence that proximity to markets appears to have on
the consumption side.
Information presented in Chapters 5 and 6 makes it quite clear that
most farmers are not producing enough cereal for themselves -- even in
an above average year.
This provides a relatively pessimisitic
appraisal
of the potential to achieve national
cereal production
objectives.
Although the level of analysis conducted to date does not provide
precise guidance on how to increase cereal production, a number of
related issues warrant further consideration.
The fact that returns to labor in peanut cultivation are greater
than those in cereal cultivation in four of the six zones, clearly
diminishes the incentive for farmers to produce and market cereals.
Unequal progress in liberalizing cereal and peanut markets may also
be discouraging cereal production and marketing. The peanut marketing
a The food insecure classification is based on analysis of food
consumption data and a determination that average caloric intake for the
household was less than 2400 kcal/ae.
128
structure (guaranteed purchases at guaranteed prices) provides added
incentive for farmers to favor peanut production over cereal
cultivation.
Cereal producers market their produce in a coarse grain
market that is liberalized so producers do not benefit from guaranteed
purchases and prices.
Furthermore, they are faced with an imposed
"ceiling" price because imported rice prices are controled at about 135
FCFA/kilo.
SUMMARY OF FACTORS INFLUENCING THE QUANTITIES OF CEREALS PURCHASED
Chapter 8 presents a more detailed analysis of food consumption
patterns and the link to cereal purchasing.
The major factor
influencing purchases of cereals to be identified to date is living in
a market village--the quantity of cereal purchases per adult equivalent
for market village residents is higher across all zones.
There is no clear pattern suggesting that living in a market
village increases the share of rice purchased compared with the share of
coarse grains.
In 1989/90, three zones had a larger share of rice in
purchases made by market village residents; in two zones the share is
lower.
Purchases appear to be lower in zones where nonagricultural incomes
are lower (Niakhar,
in the central basin, the southeastern basin, and
Senegal Oriental).
This suggests that households lacking access to
nonagricultural
incomes have difficulty finding the liquidity to
purchase cereals.
This hypothesis is supported by relatively low levels
of caloric intake for these zones (see Chapter 8).
Men, particularly household heads, are the primary purchasers of
cereals.
Women purchase a greater share of rice than of other cereals,
but it is still a small share,
..;
.
.j
:
.
.
:
..
129
REFERENCES
Fall, A.A., V. Kelly, and T. Reardon.
1989.
Household-level survey
methods used in the IFPRI/ISRA study of consumption and supply
impacts of agricultural price policies in Senegal.
Document 2.
IFPRI, Washington, D.C.
Kelly, V., T. Reardon, and L. McNeilly, 1991. Sales, purchases, and
gifts of cereals and pulses in the Peanut Basin and Senegal
Oriental.
Document 6.
IFPRI, Washington, D.C.
Ndiaye, 0.
Unpublished Government regulations and cereal marketing in
Senegal,
in Delgado, Reardon, Matlon (eds.) Commodity priorities
and changing food patterns in West Africa, IFPRI, Washington, D.C.
,,
.
.
1
.I
.
:
I
130
7.
LIVESTOCK TRANSACTIONS
INTRODUCTION
This chapter examines livestock sales, purchases, and consumption
from household herds.
It addresses five questions:
a
Which animal groups are most important in sales, consumption,
and purchases?
0
Why do people sell?
0
Is there any seasonality in livestock sales?
0
What does animal husbandry contribute to the total level of
income?
0
How is income from livestock distributed?
The data used for the analyses in this chapter come from monthly
interviews of household members owning livestock.
The interviews
enumerated all sales, purchases, gifts, births, and deaths (including
slaughtering for home consumption).
They also enumerated all cash
expenditures for livestock maintenance (vaccinations,
herder fees,
supplementary feed).
This analysis does not take into consideration income (gains or
losses) due to changes in stocks during the year.'
Gifts are not
analyzed either.
Maintenance costs do not cover inputs from family
resources, such as unpaid family labor and home-produced peanut hay or
cereal straw.
The analysis concerns only rural households because livestock
transactions in urban households were extremely minor.
Results are
presented for harvest years 1988/89 and 1989/90. Data collection began
well after the beginning of the 1988 harvest year (January or February
of 1989, for most zones).
In a few cases, the interviewees were able to
do a full recall to the beginning of the harvest year; most of the time
' Ndiaye (1993) reports results of a more in-depth analysis of the
IFPRI/ISRA livestock data for three zones (Niakhar and Colobane in the
central basin and Passy in the southwest) in his M.S. paper Micro-level
Small Ruminant Transaction Analysis and Livestock Policy in Seneqal. He
estimated beginning and ending stocks for the period October 1989
through September 1990 and used these estimates to develop income
statements for small ruminant enterprises.
Appendix 7 contains the
summary, conclusions,
and policy recommendations reported in the final
chapter of his paper and several pertinent tables.
131
this was not possible.
As a result, data presented for harvest year
1988 underestimate total transactions.
The period missing is a period
of both sales and purchases, hence,
it is difficult to draw any
conclusions about the direction of any bias in net transactions income
that has been introduced.
Net transactions income is calculated as follows:
sales - purchases t home consumption - maintenance.
Table 7.1 presents a summary of net livestock income across zones
and shows the share of sales receipts by animal category. More detailed
tables are in Appendix 7.2.
Table 7.1 -- Summary of net income from animal transactions
NPB
CPBN
CPBC
SUP6
SEPB
cso
Net income
FCFA/AE
7015
749
6473
1457
6119
3778
Share of sales receipts
from
Cattle (%)
4 4
41
28
50
5 6
a 4
Small ruminants (%)
5 5
5 4
58
47
4 3
1 5
Poultry (X)
1
5
14
3
1
1
Maintenance Expenditure
(FCFA/AE)
a01
908
490
160
18
1340
Pal-t: 3:' r9&9/9o
NPB
CPBN
CPBC
SUPB
SEPB
cso
Net Income
FCFA/AE
N.AV.
988
3032
1248
7878
4904
Share of sales receipts
from
Cattle (X)
N.AV.
54
28
4 6
53
91
Small ruminants (%)
N.AV.
4 2
61
52
4 6
a
Poultry (X)
N.AV.
4
11
2
1
1
Maintenance Expenditure
(FCFA/AE)
N.AV.
866
1120
1074
726
720
:
.I
.
.
I,
‘_
.
132
IMPORTANCE OF ANIMAL GROUPS
Sales
In the periods and zones surveyed, the value of sheep and goats
sold roughly equalled that for cattle and traction animals.
The
exceptions are in the central Peanut Basin (Colobane), where sales of
small ruminants were double those of larger animals in harvest year 1989
and in Senegal Oriental, where sales of cattle in both harvest years
were more than 80 percent of total sales.
The contribution of poultry to total value of livestock sales was
generally very low (O-Z percent) except in the central basin (Colobane)
where poultry accounted for lo-15 percent of sales depending on the
year.
Consumption
The contribution of home consumption to net transactions income was
less than 15 percent in most zones and years.
The most striking
exception is the southwestern Peanut Basin where home consumption
accounted for 62 percent of livestock income in harvest year 1989. The
central Peanut Basin (Colobane) also had relatively high home
consumption income (31 percent of total livestock income in 1988 and 33
percent in 1989),
as did the southeastern Peanut Basin in 1988 (24
percent).
Sheep and goats provided the bulk of animal protein consumed from
owned stocks, supplying more than 90 percent of the livestock income
derived from home consumption in all zones but two. This is undoubtedly
the "Tabaski" effect from households slaughtering their own sheep to
fulfil1 religious obligations.
A somewhat lower share of home
consumption came from small ruminants in the northern Peanut Basin in
1988 (79 percent) and Senegal Oriental in both harvest years (84 and 42
percent).
Consumption of cattle contributed to net transactions income in
only two zones.
In the northern Peanut Basin, cattle accounted for 20
percent of home consumption value and 4 percent of net transactions
income.
In harvest year 1989, 55 percent of home consumption and 6
percent of net transactions came from cattle in Senegal Oriental.
Where household stocks were consumed, poultry contributed from 1 to
16 percent of income derived from home consumption.
Although home
consumption of poultry was a small share of total livestock income,
home-consumed poultry accounted for a relatively large share of net
poultry income in the southwestern basin (53 percent), southeastern
basin (81 percent),
and Senegal Oriental (46 percent) in harvest year
1989.
No home consumption of owned poultry stocks was reported in the
Niakhar zone of the central basin in either period and none in the
southwest Peanut Basin for harvest year 1988/89.
133
The major conclusion to be drawn here is that animal husbandry
activities do not make a very large direct contribution to food security
because the amount of protein provided by home consumption (valued at
less than 2,000 FCFA per adult equivalent per year) is quite small. In
very rough terms, a single chicken costs about 750 FCFA and a kilo of
beef from 500 to 700 FCFA, depending on the zone. This translates to a
maximum of 4 kilos of home-produced meat per adult equivalent per year.
Purchases
The relative importance of different types of livestock as a
percent of total livestock expenditures varies with region. In 1988/89,
nearly all of the livestock purchased in the northern and southwestern
Peanut Basin and in Senegal Oriental consisted of sheep and goats.
During the same year, expenditures on cattle and small ruminants were
about equal in the southeast.
In the Colobane zone of the central
basin,
expenditures on small ruminants were almost double those on
cattle, while the reverse was true in the Niakhar zone of the same
region.
Livestock purchases in all zones increased 2.5 to 3-fold between
harvest years.
Part of this increase is probably due to the incomplete
enumeration of transactions in the postharvest period of harvest year
1988.
The relatively good crop harvest in 1989, however, did provide
many households with a big increase in cropping income (see Chapter 5),
thereby encouraging a higher level of livestock purchases than in
harvest year 1988.
In the central basin, the composition of purchases in the second
year was similar to the first.
Niakhar continued its preference for
cattle and traction animals while Colobane gave priority to small
ruminants.
In the southwest and Senegal Oriental, the
1988/89
preference for small ruminants was shifted to cattle in 1989/90.
Poultry purchases to increase stocks were minor (O-6 percent of
purchases) except in the Colobane zone in the Central Basin, where they
were 15 percent of purchases in 1988.'
REASONS FOR SELLING
In all regions, livestock was sold largely to obtain food. The
portion of annual sales that went for food is generally in the 30-50
percent range.
The exceptions are Colobane in 1989/90 and Dioly in
1988/89 - - 19 and 24 percent respectively.
The next most important
reasons households sold their livestock, both as a portion of the budget
and frequency cited, was to improve their herd by reinvesting in
different animals.
Only in two regions during the survey period
* Note that purchases for immediate consumption are not considered
in calculating net livestock income.
,.
._
t
.
.
.
.,
;
*
!
‘.
,’
:
134
(central and southwest basin) was livestock sold to obtain agricultural
inputs.
It should be noted, however, that only the principal reason for
selling an animal was enumerated.
Frequently the money obtained from
the animal sale was used to purchase food as well as other goods. In
transactions data for purchases of agricultural inputs (particularly
seeds), "animal sale" was frequently cited as the source of financing
for the purchase.
There were a number of other reasons for selling livestock, such as
paying debts, paying for ceremonies, buying gifts, financing trips, and
assorted other activities.
Except in Senegal Oriental where 19 percent
of monthly sales in 1988/89 went for ceremonies, and the northern Peanut
Basin where 17 percent of sales went for gifts in 1988/89, the other
reasons comprised less than 15 percent of annual sales and there is no
pattern or consistency among the regions. The reasons households cited
for selling livestock during peak periods are discussed in the section
on seasonality.
The major conclusion to be drawn from this section is that animal
husbandry plays an important role in ensuring food security by providing
a relatively liquid form of savings that can be quickly transformed into
cash when food supplies run low.
Animal husbandry's direct contribution
to meat consumption is relatively low, with small ruminants slaughtered
at Tabaski accounting for a large share of this consumption.
DISTRIBUTION OF NET INCOME FROM LIVESTOCK TRANSACTIONS
Livestock transactions are one of the most difficult activities to
treat in a general context because the level of holdings as well as the
importance of transaction income can be extremely variable both within
and across zones.
Appendix Figures A7.1-A7.11 are histograms showing the distribution
of household income from livestock transactions for each zone and each
year of the study.
The following conclusions can be drawn from these
histograms:
Incomes are most skewed in the zones with the highest average and
median
livestock incomes.
The northern Peanut Basin and the
southeastern Peanut Basin have average livestock incomes in the 6,000-
8,000 FCFA range and median incomes in the 2,000-4,000 FCFA range. The
gap between lowest and highest incomes in both zones is extremely large
(-3,000 to 72,000 in the north and -500 to 300,000 in the southeast for
the 1988/89 harvest year).
While average incomes in other zones are consistently higher than
median incomes, the gaps tend to be smaller than those in the north and
southeast, suggesting that average values are less influenced by a few
households with very high livestock incomes.
.I’
:
135
The range from lowest to highest incomes is also much smaller in
these zones.
With the exception of Senegal Oriental in 1989/90, the gap
between high and low incomes ranges from 12,000 to 37,000 FCFA per year.
One household in Senegal Oriental had a particularly high income in
1989, pushing the gap for this zone up to 66,000 FCFA.
Ndiaye (1993) estimated Gini coefficients of livestock holdings
using IFPRI/ISRA data for three of the six rural zones.
His results
(see Appendix 7.3) confirm that the distribution per capita is quite
skewed with coefficients greater than .65 for cattle and .42 for small
ruminants.3
The implication of this extremely variable level of income from
livestock activities is that programs to help households increase
livestock production will have important equity effects. Care must be
taken to ensure that any policy interventions to improve livestock
production do not accrue primarily to those households already at the
top of the pyramid for livestock income.
Another policy implication can be drawn from the fact that one of
the zones with the biggest herds of livestock (Niakhar in the central
Peanut Basin) has the smallest livestock income.
There are clearly
cultural explanations for this behavior (see comments in Chapter 4);
however, analysts should be asking if there is not some way to encourage
greater economic exploitation of livestock holdings in this zone, which
exhibits the lowest levels of per capita income in the entire survey.
CONTRIBUTION TO TOTAL LEVEL OF RURAL INCOME
Chapter 4 discussed the sectoral composition of total household
income in detail.
The analyses presented (Table 4.5) show that net
transactions income from livestock contributes less than 25 percent of
total household income across all zones and years.
Zones where livestock was most important were the northern Peanut
Basin (17 percent of total income) and Colobane in the Central Basin
with 23 percent in 1988/89 (but only 5 percent in 1989/90).
In both
cases, the large share occurred in 1988 after the very bad harvest.
Livestock was more than 15 percent of income in the southeastern
Peanut Basin and Senegal Oriental during both years. These zones are
not traditionally identified with livestock production. It appears that
pasture and land constraints further north are pushing livestock
production toward the south.
In these zones, livestock is the second
most important single source of income after crop production.
3
The Gini coefficient is an indication of how skewed the
distribution of livestock is across households; a coefficient close to
zero implies a relatively equal distribution while a coefficient
approaching 1 means the distribution is unequal.
136
Zones with relatively small shares of livestock income (5 percent
or less) are the central Peanut Basin (Niakhar in both years and
Colobane in 1989) and the southwestern Peanut Basin (Passy).
Low
livestock incomes in the central basin appear to be due to investments
(purchase of new animals for the herd) that exceed sales severalfold.
This is a pattern that is typical of the Serer ethnic group, which
predominates
in Niakhar and is important in Colobane.
In the
southwestern Peanut Basin, the overall level of transactions is quite
low, with the value of purchases 2-3 times sales.
Tables in Appendix 7.2 present information about changes in stocks
during the two study periods.
Although no value is placed on changes in
stocks resulting from births, deaths, and gifts, the data do help one to
draw some tentative conclusions about what role animal husbandry
activities play in changing a household's total assets. For example, in
the Niakhar zone, which has a very low level of livestock income, stocks
increased substantially during 1988/89 (38 new goats).
I n 1989/90,
stocks of goats declined due to deaths (down 7 animals net) while sheep
increased by 6.
This suggests that livestock activities are probably
more profitable to households in this zone than the net transactions
analysis presented above suggests.
A quick look at the tables for other
zones suggests, that in general, livestock herds increased during the
two-year survey period.
Additional analyses have been done by Ndiaye (1993) to combine
transactions income with capital gains and losses to obtain a full
picture of net income from raising small ruminants in harvest year
1989/90.
A summary table of his results for the central and
southwestern basins is presented in Appendix 7.4.
Ndiaye's analysis shows that poorer households (those with average
income per AE less than the median income for the zone) lost money on
small ruminants in the Niakhar zone of the central basin and in the
southwest.
The primary cause was a substantial drop in the value of
ending inventories due to high death rates.
Poorer households in the
Colobane zone realized a net income of 12,500 FCFA (6,500 per small
ruminant head).
The wealthier half of the sample in each zone (income per AE
greater than the median) realized a net profit. Colobane had extremely
good returns (21,600 FCFA per small ruminant head), while the profits in
the southwest and Niakhar were more modest (3,800 and 500 FCFA per
head).
These results suggest that poorer households may not have adequate
resources for animal feed and medical care, thereby limiting the extent
to which they can supplement cropping income with livestock income.
137
SEASONALITY OF SALES
Appendix 7.2 contains tables illustrating the seasonality of sales
for each zone and year.
The 1988/89 harvest year was very bad in the northern Peanut Basin
and 41 percent of livestock sales for the year took place in October and
December.
These sales were primarily for food, ceremonies, and gifts,
and consisted mostly (82 percent) of cattle, with sheep and goats making
up the rest.
In the Niakhar zone of the central basin, which also suffered in
1988/89, 69 percent of the year's livestock sales occurred in June. A
quarter of these sales were for food, another quarter for reinvesting in
livestock, and the remaining share (42 percent) for unspecified reasons.
Sales were mainly cattle (42 percent -- this was the only month cattle
were sold all year) and hogs (28 percent).
In the 1989/90 y e a r , sales peaked in May and July, with the
livestock sales consisting of cattle (42 and 46 percent) and goats (40
and 32 percent respectively).
Sales in May were to finance trips, to
buy food, and for ceremonies and other uses, while all of the sales in
June were for livestock reinvestment.
In the Colobane zone in the central basin, peak sales in the
1988/89 harvest year occurred during June and July, with the bulk of
sales being sheep sold for food, agricultural inputs, and other reasons.
The following year also saw peak sales in June and July composed of
sheep (44 percent), cattle (28 percent), and horses (26 percent). These
sales were mostly to finance another activity or to pay off debts.
In the southwestern Peanut Basin during the 1988/89 season, the
heaviest livestock sales were distributed fairly evenly from February
through May.
Horses made up 58 percent of sales in February, but
farmers shifted to cattle after February (46, 72, and 73 percent
respectively for March, April, and May).
The horses sold went primarily
for reinvestment in livestock and some for food. The later sales were
for food, to pay off debts, taxes,
and for livestock reinvestment. In
1989/90, sales peaked in January, March, and May, with horses making up
70 percent of sales in January, and cattle in March and May (90 percent
and 58 percent respectively).
Debts absorbed most of the income from
sales in January while livestock was sold almost exclusively for food in
March and for food and agricultural inputs in May.
In the southeastern Peanut Basin, sales peaked in March of 1988,
with cattle supplying 64 percent of the sales, more than half of them
for livestock reinvestment. In the next harvest year, sales peaked in
March and May and consisted mostly of cattle and horses.
In March
nearly all the sales were for food while in May about a quarter were for
agricultural inputs.
,
’
)’
;
.
.
.
:
1
.
.
138
In Senegal Oriental, peak sales in the 1988/89 harvest year took
place in January and February and consisted largely of cattle (75 and 97
percent respectively).
Some of the January revenue went for food but
also to finance trips and for other uses.
In February almost half of it
went for ceremonies.
In the 1989/90 season, peak sales occurred in
October and December and June, and again were largely cattle (92 and 81
percent).
Some of the sales in January were for food, but a number of
the animals were sold because they were sick.
In June more than 60
percent of sales were for food.
Only in the northern Peanut Basin during 1988/89 and in Senegal
Oriental during 1989/90 did households sell animals and purchase food
early when prices were relatively low.
Most sales for food occurred in
June/July when cereal prices were high. The question is, does the value
of an animal increase enough between harvest and June/July to warrant
this type of behavior? The experience in Senegal Oriental suggests that
it does not because prices of goats and sheep were very low in June and
July in 1989/90 when many households were in distress and selling
animals.
Households in Colobane in the central basin and those in the
southwestern basin are the only ones that seem to sell animals in large
numbers to pay for agricultural inputs or to pay off agricultural
credit.
In the southwest this is related to credit from a confectionery
peanut program.
Farmers with poor harvests will liquidate other
resources
to reimburse
suppliers for credit to maintain their
eligibility for the program in the following year.
In Colobane, peanut
production is very important and farmers sell animals to purchase seed.
The links between livestock and household food security have
already been mentioned in Chapter 4.
Information presented here on the
seasonality and reasons for livestock sales only serves to underscore
the important contribution that animal husbandry makes to smoothing
income across years and seasons.
This income smoothing helps households
cover cereal production shortfalls with purchases and, in some cases,
provides liquidity for purchasing agricultural inputs, particularly
peanut seed.
SUMMARY OF MAIN POINTS
Animal husbandry income is used by most households to smooth total
household income and compensate for crop shortfalls.
As a result,
livestock income tends to fluctuate across years (between 5 and 20
percent of total household income,
in general) and is negatively
correlated with cropping income.
Interannual fluctuations are lower,
however, in the southeastern Peanut Basin and Senegal Oriental where
livestock income was consistently at least 15 percent of total household
income in both years.
139
Income from livestock sales is used, above all, to purchase food.
The next most common reason for sales is to reinvest in different types
of animals.
Distribution of livestock assests across households is quite skewed
with Gini coefficients in the .4 to .5 range for small ruminants and in
the .7 to -9 range for cattle (Ndiaye, 1993).
Distribution of livestock income across households is also quite
skewed with many households having zero or negative incomes and a few
households having very high incomes.
Size of herds is not always positively correlated with the level of
net income from livestock transactions nor with the share of livestock
income in total income.
The Colobane zone of the central basin has one
of the smallest herd sizes but one of the higher levels of livestock
income per AE.
The Niakhar zone in the central basin has one of the
largest herd sizes but very low levels of transactions income.
The
former zone appears to raise animals as a means of supplementing income
while households in the latter zone appear to use livestock as a form of
wealth accumlation
and savings.
Policies to improve livestock
productivity will need to develop programs that respond to these
different types of strategies.
Using IFPRI/ISRA data, Ndiaye (1993) shows that poorer households
tend to have much higher loses than the rich due to animal deaths. As
a result, the poor realize negative net livestock income when changes in
the value of animal stocks are considered. If the livestock sector is
to serve as a complement to the cropping sector, enabling both poor and
rich households to better cope with cropping shortfalls, greater
attention will have to be paid to improving the health of livestock
raised in poorer households.
:
.
.
:
.
140
REFERENCES
Ndiaye, 0. 1993.
Micro-level Small Ruminant Transaction Analysis and
Livestock Po1ic.y in Seneqal.
Masters Paper,
Michigan State
University, E. Lansing.
,
.:
.
:
:
s
:
‘.
141
8.
FOOD CONSUMPTION PATTERNS
INTRODUCTION
The primary objective in examining food consumption behavior in
this study is to understand how different types of households will
respond to changes in consumer and producer prices for basic food crops.
Such knowledge will make it possible to predict the probable outcome of
proposed policy initiatives and to evaluate the impact of such policies
on households that differ by income level, location, and sectoral
composition of income, for example.
This chapter presents brief summaries of food consumption for each
of the agroclimatic regions by zone and compares factors that influence
the level, composition,
and source of food consumed by the sample
households.
More detailed analyses for each zone and village are given
in the appendixes for this chapter, which also contain the tables and
charts upon which the analyses are based.
While hypotheses are offered about the role of prices, income, and
market village infrastructure in shaping food consumption behavior; the
primary objective of this chapter is to describe current patterns. More
rigorous,
quantitative analysis of how prices, incomes, and other
relevant variables influence food consumption will be conducted in the
future using econometric techniques.
Quality of Harvest
One expects food consumption levels and composition to vary in
rural Senegal with the quality and composition of the harvest.
The
quality of the harvest depends to a large extent on the climatic
conditions,
amount of crop disease, and level of pests.
These factors
are highly variable in Senegal; one village may have terrible rains
while a neighboring village no more than 5 kilometers away might have
good rainfall.
The next few paragraphs provide a general description of
the quality of the harvest for the two years of the study.
Due to
substantial
intrazone variation
in the quality of the harvest,
situations that depart from these generalizations will be mentioned when
necessary.
The 1988/89 crop production year ended with a very poor harvest in
the northern Peanut Basin and in a large part of the central Peanut
Basin.
Poorly distributed rains were responsible to some extent, but
142
the main culprit was a locust attack that occurred just before the
harvest.
Cereals suffered the most from this attack.
The southern parts of the Peanut Basin and Senegal Oriental were
not particularly affected by locusts in 1988 and the harvest was
considered average in most parts of these zones.
The 1989/90 cropping season was one of the best in the past ten
years for most of Senegal.
Although pockets of irregular rainfall
persisted in some zones,
the 1989/90 gross value of agricultural
production per adult equivalent (AE) across all survey areas was 1.5 to
2.4 times greater than that for 1988/89 (see Chapter 5 for a full
discussion of crop production).
Nature of Data and Analyses Performed
Data on food consumption in each of the six rural zones were
collected over the two-year study period (HY88 and HY89); data on the
two urban zones were collected for a one-year period (July 1990~June
1991) .'
The surveys were designed to provide information on:
Caloric adequacy per adult equivalent (AE);'
Composition by product;
Sources (home production, purchase, gift);
Seasonal variation in levels, composition, and sources;
Influence of market villages on consumption patterns;
Differences in consumption between the poorest and the
richest income quartiles.
The data were collected fortnightly. The interviewer asked the
respondent to recall, for the previous day, household consumption of
cereals, pulses, tubers, animal products, oil, and bread -- at home and
outside the household.
1
Several preliminary versions of this chapter have been
distributed
in both French and English.
Among them are "Note
d/information sur la consommation des cereales et des legumineuses en
milieu rural senegalais,"
issued in July 1991, and project document 7,
Food Consumption in the Peanut Basin and Seneqal Oriental: October 1988
- September 1989 issued in February 1992.
Results presented here
supersede those presented in all previous reports where weighting for
disproportionate sampling of market villages was not taken into account.
* In calculating adult equivalents present at each meal we use the
following rules adapted from FA0 1975 standards:
adult male 15
2 years = 1
adult female 2 15 years = .73
boy 5-14 years =
.83
girl 5-14 years =
.70
boy < 5 years =
.44
girl < 5 years =
.44
.;
;
143
The number of household members and guests present for each meal
was also recorded, as were dishes sent to others or received as gifts.
The woman who cooked the day before was the principal respondent; those
having purchased food away from home were interviewed separately about
their consumption; data on food consumed away from home are not analyzed
in this report.3
The "Office de recherches sur l'alimentation et la nutrition
africaine" (ORANA) has established 3000 kilocalories (kcals) per day as
the intake norm for a moderately active Senegalese male.4
Populations
consuming less than 80 percent of this recommended level (fewer than
2400 kcal) are considered to be "at risk." The IFPRI/ISRA analysis
evaluates the extent to which the sample households have consumed at
least 80 percent of the recommended norm and, therefore, are no longer
"at risk".
Throughout the discussion, we are interested in the extent
to which the 2400-kcal level of "minimum requirements" is met by sample
households.
Calorie calculations for cereals took into account processing
losses
and nonconsumption of bran for millet, sorghum, and maize.
Using secondary data as well as a mini survey of processing losses for
a subset of sample households in Senegal Oriental, the following caloric
availability was established for coarse grains:
Calories Available per 100 grams of Whole Grain
Product
Before Processinq
After Processinq
Millet
353
281
Sorghum
354
281
Maize
357
285
Calories available for other products are taken from ORANA and are as
follows:
Product
Calories per 100 qrams
Peanuts (shelled)
550
Cowpeas (shelled)
342
Milk
79
Bread (wheat)
252
Vegetable Oil (peanut)
884
Cassava
146
Rice (processed)
363
3
See Fall, Kelly, and Reardon (1989) for details of survey
methods.
4 ORANA, "Cartes des calories et des vitamins pour 150 aliments
africains," no date.
' See Appendix 8.2 - for a fuller discussion of this issue.
.
,-,
,
:
.:
‘
.
”
‘.
,
144
The results reported are also weighted to adjust for non-
proportional sampling of households in market villages (see Chapter 2).6
A few weaknesses in the data need to be pointed out. The results for
the 1988/89 harvest year are based on data for just the period December
1988 through September 1989.
Given that the harvest season (October and
November) is missing, these results understate the overall consumption
level and the share of home production, while overstating the share of
rice.
The biases are probably minor in the northern and central Peanut
Basin because the harvest was relatively poor in these zones.
The analysis examined calories obtained from cereals and pulses
plus those obtained from vegetable oil, milk, bread, and tubers. The
term "total calories" means total calories obtained from the products
analyzed.
In discussions of composition and source, "percent of total
calories" means percent of total calories for the food qroups analvzed
to date.
The major food groups not addressed in this report are sugar,
vegetables, meat, and fish.
Given that the primary objective of the research was to provide
data for an examination of agricultural price policies with respect to
cereals and pulses, a great deal of time and effort was devoted to
ensuring high-quality data for these products.
The "other food"
products examined are more difficult to measure because of the small
quantities consumed and irregular patterns of consumption. Estimates of
average calories consumed per AE for these products exhibit very high
coefficients of variation and must be used with caution.
Given the
dearth of information on consumption of these products, it is important
to present what is available, fully acknowledging that it is a very
rough estimate of calories obtained from these products.
In other chapters (particularly Chapter 4), we divided sample
households into income terciles to examine the relationship between
income and different types of production or consumption behavior. In
this chapter, we compare the consumption patterns of the richest
quartile with that of the poorest quartile to examine the income effect.
This was done because the use of terciles failed to provide us with
useful information on how consumption varies by income group. The range
of caloric intake levels in the sample is not extremely large.
This
made it difficult to find significant differences when the sample was
divided into only 3 income groups.
Even with the use of quartiles,
differences in cereal and pulse consumption between the wealthiest and
poorest quartiles are not statistically significant. We believe that
the greatest differences in consumption between rich and poor relate to
the amount of expenditure on products not included in the caloric
analysis presented in this chapter (fish, meat, vegetables, and sugar).
This hypothesis
needs to be tested, however, with econometric
techniques.
6 This weighting produces results that are occasionally different
from preliminary,
unweighted results reported in Project Document 7.,
Food Consumption in the Peanut Basin and Senegal Oriental: October 1988
- September 1989 by Kelly et al., 1992.
145
SUMMARIES OF FOOD CONSUMPTION BY ZONE
This chapter is a summary of a more detailed zone by zone
discussion of results that is contained in the Appendix along with a set
of detailed tables presenting analysis of results for each zone and
village.
Tables 8.1 and 8.2 present summary statistics on the levels and
composition of calorie intake by zone and year.
The summary statistics
illustrate the extent to which access to market infrastructure and level
of income influence consumption levels, composition and sources.
Northern Peanut Basin: Zone of Saqatta
The typical AE in the northern Peanut Basin consumes about 2220
calories daily from products analyzed to date. Ninety-three percent of
minimum requirements are covered by these products, suggesting that - on
average - the typical AE will not fall in the "at risk" category when
unanalyzed products that are regularly consumed in the zone (sugar,
meat, fish, and vegetables) are added.
The typical northern Peanut Basin AE relied on purchases for 82
percent of cereal/pulse and 90 percent of cereal calories in HY88.
Purchasers in the zone favor imported rice, which accounts for 66
percent of cereal calories consumed.
Although the Northern Peanut Basin
is not very far from rice producing zones of the Senegal River Valley,
there was no evidence that any rice from the river valley was being
consumed by sample households.
Sixty-eight percent of households in the sample averaged less than
the 2400 minimum requirements during HY88. The poorest income quartile
of the sample consumes an average of 1925 cereal/pulse calories per AE.
This is 200 calories less than average consumption by the richest
quartile.
It is unlikely that households in the poorer group reached
the 2400 minimum requirements by purchasing foods to supplement cereal
and pulse consumption.
Although seasonality in consumption is not extreme (ranging from
93 to 105 percent of the annual average, depending on season),
consumption drops during the March-July period and rises as soon as the
cowpea harvest becomes available in August.
Recent Government of Senegal policies to increase cowpea production
in this zone appear to have had a strong positive effect on ensuring
food security during the August-October period when the new cowpea
harvest is available but millet and peanuts have not yet arrived at
maturity.
Surprisingly, households and villages with better harvests (higher
value of production per AE) tend to have lower levels of caloric intake.
The households with better harvests also tend to have less diversified
incomes with greater than 30 percent of total income coming from
cropping while households with better caloric intake levels realize only
146
lo-15 percent of total income from cropping. These results suggest that
food security in this zone is better ensured by concentrating on non-
cropping activities.
The market village has more diversified consumption patterns, with
less reliance on cereals and pulses than the nonmarket villages. This
is as expected given that "other" foods are more readily available in
market villages and that opportunities to earn cash to purchase "other"
foods also tends to be greater.
Looking at only cereals and pulses,
average calories per AE differ across villages by as much as 555
calories per day.
When "other foods" are added, the maximum difference
drops to 264 calories.
More work is needed in analyzing the links between income
diversification and food security in Senegal. Our impression is that
food security in this zone could be enhanced by rural development
policies that help farmers to diversify income sources while using crop
production as a means of supplementing non-cropping income (see Chapter
4 for more discussion of this issue).
Central Peanut Basin: Zone of Niakhar
During HY88, the typical household in the central Peanut Basin zone
of Niakhar consumed only 2162 calories per day per AE from the product
groups analyzed to date.
It is unlikely that products not yet analyzed
(sugar, meat, fish, and vegetables) would have pushed this level to the
minimum requirement of 2400 calories.
HY89 was a better year, with
average consumption in the zone reaching 2415 calories per day per AE.
In HY88, 68 percent of households failed to achieve an average
level of consumption (from products analyzed) of 2400 kcal/AE.
The
share below the minimum requirements dropped to 32 percent following the
better 1989 harvest.
Households in the lowest income quartile consumed 154 fewer
cereal/pulse calories per AE than the highest quartile in HY88 and 279
fewer in HY89.
In both years it is unlikely that households in the
lowest quartile consumed the minimum 2400 calories per day, even after
consumption of all other foods is taken into account.
The top income
quartile passed the 2400 level in HY89 with cereal/pulse products alone.
It is not clear that the wealthy quartile passed the 2400 line in HY88
because cereals and pulses covered only 2065 calories, leaving about 350
calories to be obtained from other products in this zone,
which exhibits
fairly low consumption of noncereal pulse products.
In sum, the zone appears to have some problem ensuring food
security at the minimum required level of 2400 calories.
Following poor
harvests, both rich and poor have problems; following good harvests, the
poor continue to have problems while the wealthy pass the "at risk"
level and move toward normal consumption.
The typical central Peanut Basin household in Niakhar is much more
self-reliant in terms of cereal and pulse consumption than other parts
.”
*
‘.
:
:
..;
.,
;
147
of the Peanut Basin.
Following the poor 1988 harvest, 21 percent of
total cereals and pulses were purchased while just 8 percent of m
produced cereals and pulses were purchased; this dropped to 5 percent in
HY89.
Although this appears commendable given the government's
objectives of encouraging self-sufficiency in cereal production, one
must not lose sight of the fact that consumption levels were not
adequate in HY88.
The degree of self-reliance would diminish if all
households in the zone had adequate resources to bring their consumption
up to the recommended minimum level through purchases.
Farmers in this zone tend to give more attention to growing their
own millet than to expanding their peanut production, despite the land
shortage that. plagues the zone.
Farmers in other areas of the Peanut
Basin (see the discussion of the Colobane zone) have been turning
increasingly to peanuts and relying on the market for cereals.
Imported rice is not as large a share of the overall diet (15
percent of cereals in HY88 and 14 percent in HY89) as it is further
north.
Although prices and incomes play some role in determining the
mix of rice and coarse grains, the Serere people living in this zone are
known to have stronger preferences for millet than other ethnic groups
in the Peanut Basin.
Rice is a relatively small part of total cereal/pulse consumption,
yet it is a very important part of purchases (65 percent of all
cereal/pulse purchases in HY88 and 73 percent in HY89). Interestingly,
the share of rice in purchases increased after the better harvest. This
was most likely due to (1) higher incomes that permitted purchases of
rice and (2) a desire to introduce some variety into the essentially
millet-based diet in the zone.
In looking at seasonality, we find that households in the zone tend
to consume their own production first, increasing purchases of both rice
and millet when stocks of millet run out.
This results in the bulk of
cereal products being purchased when the price of millet is at its
highest point relative to the stable (government-controlled) price of
rice.
This leads one to ask whether rice consumption would be lower if
cereal purchases were made when millet prices were still low.
The phenomenon of an early and abundant 1989 cowpea harvest that
helped families supplement their calories during the end of the rainy
season occurred in this zone, but on a smaller scale than in the
northern Peanut Basin where more cowpeas are cultivated.
Unlike the
northern Peanut Basin, where increased cowpea consumption was
concentrated in the lowest income quartile, both rich and poor exhibit
approximately the same share of cowpeas in total cereal/pulse
consumption.
Although the share of imported rice in total cereal/pulse
consumption is not strongly affected by income following a good harvest,
the share of imported rice in the diets of market village households is
about double that of households in other rural villages both years'.
_’
.
‘.
:. . .
'i
.-
:
:
I
.
'.
148
Households in market villages also purchase a larger share of their
cereals (about 35 percent versus 16 percent for the nonmarket villages).
Central Peanut Basin: Zone of Colobane
The typical household in this zone consumed on average 2624
kilocalories per adult equivalent in harvest year 1988/89 and 2462 kcal
in HY1989/90.
This average consumption met the minimal calorie
requirements for both harvest years.
Despite the acceptable average, 50
percent of households in HY88 and 32 percent in HY89 exhibited average
intake lower than the 2400 kcal/AE "at risk" level.
Consumption was
based primarily on cereals (74-77 percent),
particularly millet, with rice as a supplement. Although cereals remain
import, Colobane has the lowest share of cereals in total calories of
all rural zones but those in the northern Peanut Basin.
This zone also
exhibits the second highest level of imported rice consumption (about 30
percent of cereals).
Rice consumption in the market village is two
times higher than in the nonmarket villages.
The zone also has an
extremely high level of vegetable oil consumption, which resulted in
more calories coming from the "other food" category than from pulses in
1988/89.7
This zone has the second highest dependence on purchases (following
the northern Peanut Basin), with more than 40 percent of cereal/pulse
calories and 48 percent of cereal calories being purchased.
This
behavior is especially marked in the market villages and in both the
highest and lowest income quartiles.
In poor households, there was no
substantial increase in the quantity of home-produced millet consumed
following the better 1989 harvest.
Wealthy households exhibited much
greater consumption of home-produced millet in 1989-90, suggesting that
they were more able to profit from better rains than the poor.
Seasonality is not a problem, given that average annual consumption
levels are adequate, and the seasonal variation never goes below 93
percent of the annual average.
Demand for rice does vary by season and
harvest year.
The pattern differs from that in the Niakhar zone of the
Central Basin.
Southwest Peanut Basin: Zone of Pass-v
On the whole, the typical household in the southwest basin consumed
an average of 2451 kcal/AE/day, which supplied the minimal calorie
requirement for the two harvest years.
In both years, 40 percent of
households in this zone failed to average 2400 kcal/AE. This was the
lowest rate of inadequate consumption in HY88 and one of the lowest in
HY89.
' Transforming peanuts into oil is an important income-generating
activity in this zone.
149
The share of calories coming from cereals is highest in this zone
(83 percent) while the contribution of "other foods" is quite small (4-5
percent).
Overall consumption was based on millet (65-66 percent of the
cereals and pulses consumed), rice (more than 20 percent), and peanuts
as a supplement (a little more than 10 percent). Production remains the
primary source of calories but purchases are common (39 percent of
cereal calories in HY88 and 27 percent in HY89).
Rice is the most frequently purchased product -- 68 and 83 percent
of cereal/pulse purchases in 1988 and 1989, respectively. The share of
rice purchases increases following the better harvest, probably to add
variety to the diet.
As in the central Peanut Basin zones, seasonal variation in caloric
intake is not high (92-106 percent of annual average).
Surprisingly,
cereal purchases and imported rice account for about the same share of
cereal calories in one of the non-market villages as they do in the
market village.
When both non-market villages are averaged together,
however, the market village rice consumption and purchases are slightly
higher.
In 1988 there was little difference in consumption levels of the
rich and poor.
In 1989, however, the rich consumed about 400 calories
more.
The poor in this zone ate a smaller share of rice (35 percent of
cereal calories) than the rich (43 percent) in 1988. In 1989, rice was
a smaller share of the cereal diet for both rich and poor, and rich
continued to consume a larger share than the poor (36 versus 29
percent).
Southeast Peanut Basin: Zone of Nsanda-Dioly
The typical household in the southeastern basin consumed on average
2232 kcal/AE/day in 1988/89 and 2304 in 1989/90.'
This represents on
average, 95 percent of minimal requirements over the two harvest years.
Although average
levels approached minimum needs, 70 percent of
households in HY88 and 60 percent in HY89 fell below the 2400 kcal/AE
minimum needs.
This suggests the existence of a serious food problem
that needs to be clarified.
In the present analysis, nothing suggests
that including other food groups would make up the calorie deficit.
There is a big difference between overall consumption levels in the
market village (92-97 percent of minimum needs covered) and the
nonmarket villages (78-90 percent).
On average, the nonmarket villages
had a higher production sufficiency ratio than the market village in
' The draft report showed much lower consumption levels for this
zone (2024 and 2006 kcal/AE for HY88 and HY89, respectively). A review
of the data files and analysis resulted in a number of corrections.
Time did not permit us to rerun all the analysis tables in the Appendix
nor to revise all the graphs.
The text of the final report, however,
has been modified by removing comments concerning the extremely low
average consumption levels for this zone.
;,
I
.
‘.
150
both years (see Chapter 5).
About one quarter of households in the
nonmarket villages, however, had an extremely low (0.24) production
sufficiency ratio in 1989/90 and one equal to that in the market village
(0.44) in 1988.
This suggests that the overall low consumption levels
in the zone are probably being strongly influenced by consumption levels
for this set of about six households.
The results also suggest that
nonmarket villages in the zone have much more trouble purchasing
cereals.
The extent to which this is due to inadequate supply (the
nonmarket villages are quite isolated) versus lack of purchasing power
needs to be examined more thoroughly.
Grains supply the essential food needs, in particular millet (67-69
percent of grains consumed), sorghum (9-12 percent), peanuts (11-14
percent), and finally imported rice (6-8 percent). Less than 5 percent
of calories came from other foods analyzed (mostly vegetable oil in 1988
and milk in 1989).
About three-quarters of the calories consumed came
from production and a quarter from purchases. Market villages purchase
a much larger share of cereals than nonmarket villages.
Seasonality is evident in rice consumption, which is high at peanut
marketing time when money is available and low during the rest of the
year.
Peanut consumption is also seasonal -- high at harvest and
declining regularly until the next harvest.
Millet and sorghum consumption patterns differed during the two
years.
The extent to which this was a price phenomenon is examined in
Chapter 9.
Finally, consumption of cereals and pulses by the poor was only 71
percent of that of the rich.
In no other part of the basin was one able
to observe such a large difference.
Following the 1989 harvest, the
rich were able to reduce purchases from 25 to 14 percent while the poor
increased purchases from 24 to 28 percent,
Clearly, there are serious
constraints preventing the poor from realizing the full benefits of a
better rainy season.
Seneqal Oriental: Zone of Missirah
The typical household in the Senegal Oriental consumed on average
2113 kcal/AE/day in 1988/89 and 2047 in 1989/90.
Given the very low
level of "other food" consumption enumerated in the zone, it is unlikely
that analysis of the remaining products (sugar, meat, fish, vegetables)
will bring average consumption to a satisfactory level.
On average,
only 85-88 percent of the minimal level of calories was met during the
two harvest years; this was among the lowest in the zones surveyed. The
share of households not meeting minimum requirements was also high --63
percent in HY88 and 59 percent in HY89.
These low consumption levels are surprising because the higher
rainfall in this region should provide better yields and therefore
better food security.
Although the zone had a higher cereal sufficiency
ratio in 1988/89 than all zones but the southwestern Peanut Basin,
cereal production still only provided 55 percent of minimum caloric
151
needs (45 percent if sales are netted out).
It appears that the
nonagricultural
income
stream,
with a large
share
coming from
intermittent migration remittances, failed to cover the gap, leaving
many households with inadequate consumption levels.
A peculiarity of this zone is that cereal consumption is very
diversified.
No grain dominates the others in the overwhelming fashion
that millet does in the most part of the Peanut Basin or imported rice
does in the northern basin or in the urban zones. Cereals and pulses
supply essentially all the calories consumed, especially maize (39-43
percent of cereals and pulses consumed), sorghum (15-20 percent), millet
(16-20 percent) peanuts (16-18 percent), and finally imported rice (4-6
percent).
The very large share of calories derived from peanuts (about
equal to those of millet and sorghum) is surprising in Senegal Oriental.
Does this suggest that the Peanut Basin is moving toward the southeast?
About three-quarters of the calories consumed come from production and
a quarter from purchases.
Gifts are more important in this zone than
elsewhere (3-5 percent) but still supply a very small share of total
calories.
The market village households had lower average caloric intake in
HY88 for cereals, pulses, and "other" products.
The difference
remained, but was less pronounced in HY89. Share of cereals purchased
and share of rice were consistently higher in the market village. The
difference in share of consumption coming from rice in market versus
non-market villages was much more pronounced in this zone where the
ratio was 1 to 4 in HY88 and 1 to 10 in HY89 (the ratio for other zones
was about 1 to 2).
Differences between the richest and poorest quartiles were not very
large, and not consistent across years.
Total caloric intake was about
the same.
Share of purchases and share of rice was larger for the poor
in HY88 but the pattern reversed itself in HY89.
Inter-annual variation in the composition of consumption is more
important in this zone than elsewhere.
Purchases of maize and sorghum
were more important relative to rice in 1989/90. This was probably due
to an excellent maize harvest which reduced the maize price
substantially during February-June 1990 (see Chapter 9).
Intra-annual seasonality is also more pronounced in the zone than
elsewhere.
The variation around the mean level of caloric intake ranged
from 91 to 111 percent.
This is not very different than the range in
other zones.
It has much more serious consequences in this zone,
however, due to the low average annual consumption level. Consumption
levels were highest during the harvest and lowest during the rainy
season.
Maize plays a very important role as a hungry season crop
because it can be consumed "green" toward the end of the rainy season
before other cereals become available.
Caloric intake may well be
higher during the "green" maize season than reflected in our data
because it was extremely difficult to enumerate the consumption of this
product, which is eaten in large quantities as a snack food.
152
Urban Zone:
Town of Kaolack
Approximately 50 percent of Kaolack households surveyed failed to
attain an average intake level of at least 2400 kcal/AE. The average
daily consumption per adult equivalent in this zone is about 2100 kg
from the products analyzed to date.
This represents a level covering 87
percent of the 2400 kcal/AE minimal caloric requirements. Given that
products not yet analyzed are likely to be more important in urban than
in rural zones, it is logical to think that average intake levels will
approach 2400 kcal/AE when all products are considered.
The typical Kaolack household purchases nearly all the products
that it consumes.
Imported broken rice constitutes the principal
element of the food regime, followed by millet, vegetable oil, and
peanuts.
Consumption of cereals and pulses appears to be most important
during the period April to September relative to the other months of the
year (October to March).
Urban Zone:
Town of Tambacoumba
During 1990, a typical household in the urban zone of Tambacoumba
consumed 2328 kcal/AE/day.
This attains nearly total coverage of the
minimal level of calories needed with only the products already
analyzed.
Despite the generally acceptable average level, 53 percent of
households in the sample were
below the 2400 kcal/AE minimum
requirements - about the same share as found in Kaolack.
Cereal consumption is more diversified in Tamba than in Kaolack.
Imported rice remains the principal source of calories (45 percent of
all calories analyzed and 70 percent of cereal calories).
Maize and
sorghum are the next cereals of importance, with millet in last place.
This preference for sorghum and maize is linked to proximity to the
zones of production.
The importance of peanuts is surprising (14
percent of total calories) in this zone, which lies outside the Peanut
Basin.
The majority of calories consumed by households in Tamba are
purchased (more than 90 percent of total calories). In contrast to the
urban zone of Kaolack, consumption of home-produced grains does exist,
even if the share remains low.
In October-December the portion of home-
produced cereals and pulses consumed by the households climbed to nearly
17 percent of total cereal/pulse calories.
It is during the harvest (October-December) when the highest
consumption of cereals and pulses was recorded in this zone and from
April through September that lowest levels were recorded.
Purchases
were highest from July through September and lowest during October,
November,
and December.
Maize is most important from October to
December and sorghum in January, March, and June.
.
.
.
.,
:
.
:
‘.
153
SUMMARY OF FACTORS INFLUENCING THE LEVEL, COMPOSITION, AND SOURCE OF
FOOD CONSUMED BY SAMPLE HOUSEHOLDS
Appendix 8.1 contains detailed analyses of the level, product
composition,
and source of the food basket for each of six rural and two
urban study zones, looking at how such factors as income, seasonality,
commercial infrastructure, and harvest quality influence consumption
patterns within each zone.
The zone-level findings were summarized
briefly in the preceding pages.
In this section, the analyses are used
to make comparisons across zones and to draw generalizations about how
these various factors might influence
aggregate consumption for all
zones covered by the ISRA/IFPRI study.
The discussion is divided into four major sections based on
effects:
(1)
agroclimatic
and
study
zones,
(2)
commercial
infrastructure, (3) income, and (4) harvest quality and seasonality. In
addition to drawing broad generalizations across zones, the discussion
is used to develop hypotheses about the types of variables that should
be included in a demand model and the anticipated size and signs of
model coefficients.
Differences in Consumption Due to Asroclimatic and Study Zone Effects
level of Ca'loric Intake
A number of study zones fall in the "at risk" category, with fewer
than 2400 calories per AE per day.
This becomes apparent in looking at
Figure 8.1 and Table 8.1, which show levels of caloric intake across
years and zones.
Surprisingly, the rural zones considered to have the
greatest agricultural potential due to higher rainfall, lower population
density,
and better soils are the ones that have consistently lower
levels of caloric intake.
Average consumption per AE in the
southeastern Peanut Basin and Senegal Oriental is 100 to 300 calories
below minimum needs both years of the survey. Given the extremely low
level of expenditures on other foods in these zones, it is unlikely that
the average will change significantly when foods (sugar, meat, fish, and
vegetables) not yet analyzed are added.
The Niakhar zone in the central Peanut Basin experienced relatively
low consumption following the poor harvest of 1988; average levels were
acceptable, however, in HY89.
Overall consumption is low in both the urban zones but more so in
Kaolack (2099 calories/AE) than in Tamba (2328 calories/AE).
The
caloric value of products in Tamba not yet analyzed should exceed the 75
calories necessary to bring average consumption levels up to the 2400
minimum.
Even though urban diets have a much larger "other food"
component than rural zones (suggesting that products not yet analyzed
account for an important share of total calories) and snacking is
154
figure 8. --
Consumption by product group, zone and year. SA =
Sagatta; NI, Niakhar; CO, Colobane; PA, Passy; DI,
Dioly; MI, Missirah; KA, Kaolack, and TA, Tambacounda
SA NI CO PA DI MI
SA NI CO PA DI MI
KA TA
1988/89
1989/90
1990/91
0 CEREALS m LEGUMES
OTHER
Note: The correspondence between the names of zones used in Figures 8.1 - 8.12 and the names of
zones used in the text is as follows:
SA = Sagatta = northern Peanut Basin
NI = Niakhar = central basin (Niakhar)
CO = Colobane = central basin (Colobane)
PA = Passy = southwestern Peanut Basin
DI = Dioly = southeastern Peanut Basin
MI = Missirah = central Senegal Oriental
KA = Kaolack = urban zone
TA = Tambacounda = urban zone
Corrections to data files for Dioly (southeastern Peanut Basin) were made after these
graphs were prepared. The caloric intake appears to be about 2300 Kcal/AE for this zone rather
than the 2000 Kcal/AE shown on the graphs.
See discussion in text concerning these changes.
‘,
:
:
.
155
Table 8.1 --
Summary of Finding on Caloric Intake Levels and Shares.
Z 0 N E S
NPB
CPB.N
CPB.C
SYPB
SEPB
cso
LEVELS OF CALORIE INTAKE:
TOTAL SAMPLE
CEREAL/PULSE KCAL/AE
1820
2008
2199
2350
1955
2033
SHARE CEREALS PURCHASED (X)
9 0
2 1
4 8
3 9
2 4
17
SHARE OF RICE IN CEREALS (%)
7 0
15
33
28
9
9
"OTHER" KCAL/AE
488
154
425
131
6 9
a0
MARKET VILLAGE
CEREAL/PULSE KCAL/AE
1545
1922
1818
1761
2062
1856
SHARE CEREALS PURCHASED (X)
100
38
aa
51
4 4
37
SHARE OF RICE IN CEREALS (X)
8 0
2 6
57
3 4
19
2 2
"OTHER" KCAL/AE
587
230
634
251
143
4 6
NON-MARKET VILLAGE
CEREAL/PULSE KCAL/AE
1943
2046
2300
2410
1954
2083
SHARE CEREALS PURCHASED (%)
8 4
18
39
3 6
2 2
13
SHARE OF RICE IN CEREALS (%)
4 7
13
28
2 6
a
5
"OTHER" KCAL/AE
N/A
121
341
110
60
9 6
POOREST QUARTILE
CEREAL/PULSE KCALIAE
1925
1911
2126
2411
1586
1964
SHARE CEREALS PURCHASED (X)
66
20
55
3 3
2 7
2 1
SHARE OF RICE IN CEREAS (%)
50
10
3 5
22
9
1 2
RICHEST QUARTILE
CEREAL/PULSE KCAL/AE
2124
2065
2162
2326
2237
1979
SHARE CEREALS PURCHASED (%)
9 4
2 5
57
5 0
2 8
11
57
22
43
3 6
1 9
8
i:-s~~~~~F-::::~~,“:E
::I N: CEREACL~
~‘4 :::::
.
.
,.., :: .,, :::.:::::.j.j::.::j::::.j:.:~:
. . . . .:::I’
7:::::.., .,,: .:,,. .,.,. ....::.,x..:~
.c.;::+. .: . . . :.: :I..,. ..:.,,::......: .: :;::::,.,
,, ,.:::.;,
:,‘PA’RT
...3- .:: j:i,:::.;,
.
. . . . .._. .
LEVELS OF CALORIE INTAKE:
TOTAL SAMPLE
CEREAL/PULSE KCAL/AE
2274
2224
2320
1981
1952
SHARE CEREALS PURCHASE0 (X)
17
4 7
27
2 2
1 9
SHARE OF RICE IN CEREALS (X)
14
31
2 5
6
5
"OTHER" KCALIAE
141
238
119
2 5
9 5
HARKET VILLAGE
CEREAL/PULSE KCAL/AE
2035
1868
1829
2260
1925
SHARE CEREALS PURCHASE0 (%)
3 3
7 6
4 1
3 0
2 9
SHARE OF RICE IN CEREALS (X)
28
52
3 4
16
2 0
"OTHER" KCAL/AE
211
405
179
7 8
3 2
NON-MARKET VILLAGE
CEREAL/PULSE KCAL/AE
2324
2310
2409
1952
1989
SHARE CEREALS PURCHASED (%)
15
4 2
2 5
1 9
1 5
SHARE OF RICE IN CEREALS (X)
11
27
23
5
2
"OTHER" KCAL/AE
1 2 1
167
07
20
120
POOREST QUARTILE
CEREAL/PULSE KCAL/AE
2135
2176
2305
1597
2069
SHARE CEREALS PURCHASED (%)
2 8
72
27
17
12
SHARE OF RICE IN CEREALS (%)
1 6
29
21
5
4
RICHEST QUARTILE
CEREAL/PULSE KCAL/AE
2414
2337
2733
2207
’ 1911
SHARE CEREALS PURCHASED (%)
2 1
5 5
3 9
1 9
20
SHARE OF RICE IN CEREALS (%)
1 7
3 6
2 8
5
6
156
Table 8.1 -- continued.
LEVELS OF CALORIC INTAKE:
TOTAL SAMPLE
CEREAL/PtlLSE/KCAL/AE
1530
1823
SHARE CEREALS PURCHASED (X)
9 5
82
SHARE OF RICE IN CEREALS (%)
62
7 0
"OTHER" KCAL/AE
569
505
source :
IFPRI/ISRA Survey.
common and extremely difficult to capture in urban survey data, it is
likely that products still needing analysis would approach but not
surpass the 300 calories needed to achieve an acceptable average level
for Kaolack.
Looking at average consumption levels does not address the question
of what share of households within each zone fall below or above the
2400 calorie minimum. Data presented in Table 8.1 show that in HY88, 50
to 70 percent of sample households, depending on the zone, had average
levels of caloric intake below 2400 kcal/AE from cereals, pulses, milk,
bread, vegetable oil, and tubers.
In HY89, 30 to 40 percent were below
the "at risk" level in the central and southwestern Peanut Basin while
60 percent were in the "at risk" category in the southeastern basin and
in Senegal Oriental, the urban zones had about 50 percent of households
with inadequate consumption.
In summary, two of the zones thought to have high agricultural
potential (Senegal Oriental and the southeastern basin) tend to have
lower average consumption levels and/or a higher share of households in
the "at risk" category than other zones.
The analysis in Chapter 5 revealed that in 1988/89, the cereal
sufficiency ratios for these zones were higher than for the lower
potential zones, suggesting that production is not the problem in the
first year,
In 1989/90,
the cereal sufficiency ratio for Senegal
Oriental was lower than all other zones but the northern Peanut Basin
while the ratio was highest in the southeast.
Households in the southeastern Peanut Basin had the highest average
income per AE in both years (Chapter 4), suggesting that households
should have been able to ensure food security through purchases.
The
causes of this apparent contradiction are being investigated.
Two
possible explanations are an income distribution that is highly skewed
(this might give a misleadingly high average income for the zone) or
poorly distributed over time (causing periods of low consumption).
.
I
:
.
1
.
<’
157
Another possible explanation could be poor supply due to the remote
nature of many villages in these zones.
The contribution to total caloric intake made by each of the major
product groups is summarized in Figure 8.2 and Table 8.2.
In Senegal
one would clearly expect cereals to contribute the largest share of
calories, followed by pulses, and then other foods..
Figure 8.2 and
Table 8.2 show that the assumption about cereals dominating is correct
across all rural and urban zones, but that the picture for pulses and
'other food' is mixed.
Cereals accounted for a greater share of total calories analyzed
(more than 80 percent) in the southwest and southeast Peanut Basin in
both survey years.
Cereals were also more than 80 percent of the diet
in the central Peanut Basin (Niakhar) following the poor peanut harvest
of 1988/89, which reduced households' ability to supplement cereals with
pulses and other foods.
The urban zones and the northern Peanut Basin
consume the smallest shares of cereals (63-65 percent of total
calories).
Cereals provide from 1364 to 2058 calories per AE, depending
on the zone and year -- there are no zones where cereals alone cover the
minimum needs of 2400 calories.
Pulses are usually the second most important source of calories.
"Other foods" are more important than pulses only in the urban zones and
in two of the rural zones (northern Peanut Basin and central Peanut
Basin - Colobane) following the poor 1988/89 harvest. There is no zone
where total cereal/pulse calories alone attain the minimum 2400 calorie
requirement.
The southwestern Peanut Basin zone of Passy is the only
one that comes close with slightly over 2300 calories per AE coming from
cereals and pulses in both years.
P r o d u c t C o m p o s i t i o n
In this section the product composition of each of the major groups
is discussed.
Cereals.
Millet dominates the cereals group in the four study
zones representing the central and southern Peanut Basin, providing
between 1000 and 1500 calories per AE per day. Imported rice provides
the remaining cereal calories in the central and southwestern Peanut
Basin while sorghum, rice, and maize (in order of importance) complete
the cereal diet in the southeast Peanut Basin.
Maize provides slightly more than half the cereal calories in
Senegal Oriental, with sorghum and millet contributing another 45
percent and imported rice about 5 percent.
The urban zones and the northern Peanut Basin rely most heavily on
imported rice, which accounts for 62 to 70 percent of cereal calories.
Millet and very small amounts of maize complete the cereal picture for
the northern Peanut Basin and Kaolack, while Tamba residents complete
their cereals with maize (12 percent), sorghum (12 percent), and millet
(6 percent).
.
.,;
:
.
158
Figure 8.2--
Cereal and pulse consumption by product, zone, and
year.
SA = Sagatta; NI, Niakhar; CO, Colobane; PA
Passy; DI, Dioly; MI, Missirah; KA, Kaolack, and Ti,
Tambacounda.
> 2000 -
OQ\\ 1500-
2\\
; lOOO-
0
-I
;:
5 0 0 -
SA Nl CO PA DI Ml
‘SA NI ‘CO PA DI MI
KA TA
19as/a9
1989/90
1990/w
0 MILLET
@@$j SORGHUM m CORN
@j&j IMP RICE
NIEEE
m PEANUTS
..:i
:
.
(
:
159
Table 8.2 --
Share of minimum calorie requirements covered by type of product,
zone, and year
Peanut Basin Rural Zones
Urban Zones
Central
Central
Senegal
Share
North
(Niakhar) (Colobane) Southwest
Southeast
Oriental
Kaolack
Tambacounda
Share of 2400 cal/AE
covered by cereals
& pulses(%)
HY88/89'
76
84
92
98
91
85
n.a.
n.a.
HY89/90
n.a.
95
93
97
95
81
n.a.
n.a.
7-90/6-91
64
65
Share of 2400 cal/AE
covered by all foods
analyzed(%)
HY88/89'
96
90
109
103
93
88
n.a.
n.a.
HY&?9/90
n.a.
101
103
102
96
85
n.a.
n.a.
7-90/6-91
87
86
Source: ISRA/IFPRI project survey.
Notes: Minimum caloric requirement is 2400 calories per AE per day.
This is 80% of ORANA recommended
standards of 3000 kcallday for a moderately active adult male.
Those below this level are considered
"at risk." Foods analyzed are cereals, pulses, milk, tubers, bread, and vegetable oil. n.a. means not
available.
' Data for harvest year 1988/89 cover December 1988 through September 1989.
The above points are all well illustrated in Table 8.3 and Figure
8.2.
This review suggests three distinct cereal consumption patterns:
(1)
Millet-based consumption complemented with imported rice
(central and southern Peanut Basin)
(2) Rice based consumption complemented with coarse grains (urban
zones and northern Peanut Basin)
(3) Mixed coarse grain consumption (Senegal Oriental)
Although differences in consumption patterns begin to appear among
the various zones, it is not yet clear to what extent these differences
are due to agroclimatic and socioeconomic factors that distinguish the
zones, or to price, income, and infrastructure variables, which may
differ across zones.
Pulses.
Peanuts provide 80-99 percent of all pulse calories and
cowpeas the remaining l-20 percent in all zones but the northern Peanut
Basin where
’
:
.
.’
160
Table 8,3-- Product composition of cereals consumed by zone and year
Peanut Basin Rural Zones
Urban Zones
Survey
Central
Central
Senegal
Product Period
North (Hi a k h a r ) (Co1 obane) Southwest.
Southeast
Oriental
Kaol ack
Tambacounda
Rice
wia8/89’
70
15
33
28
16
9
HY89/90
-
14
31
25
10
5
7-90/6-91
-
62
70
Millet and
sorghum
tiYaa/a9’ 2
5
a4
66
72
a3
38
HY89/90
-
86
68
75
88
48
7-90/6-91
-
36
18
Maize
nY8aia9’
5
Cl
cl
<l
1
53
HY89/90
-
Cl
cl
<l
2
47
7-90/6-91
-
1
12
Source: ISRA/IFPRI project survey.
Note: Numbers are percent of total cereal calories per adult equivalent provided by the cereal products
listed.
' Data for harvest year 1988/1989 cover December 1988 through September 1989.
the pulse diet is split equally between cowpeas and peanuts. Table 8.4
provides the exact shares of each product by zone and year.
Peanuts provide additional calories in the form of oil in zones
where home processing of peanuts is common (parts of the central Peanut
Basin).
These calories are accounted for in the "other food" section
because data are not precise enough to allow peanut oil to be separated
from other vegetable 0i1s.~
Cowpeas are a crop that received substantial donor and government
support in the mid-f980s.
Cowpeas are traditionally consumed in Senegal
as a condiment, being added in small quantities to the sauce.
The
ISRA/IFPRI data suggest that in the north and central parts of the
Peanut Basin, where the support programs were introduced, cowpeas
account for a much larger share of pulse consumption (20-50 percent)
than in the southwest Peanut Basin and Senegal Oriental (1-4 percent of
' In recent years vegetable oil processed industrially and sold in
Senegal has been a mix of various vegetable oils and peanut oil,.
Consumers tend to be unaware of what type of oil they have purchased,
therefore we are unable to distinguish among the different types.
161
pulse consumption).
This raises the question of whether consumers in
these low-consumption zones could be encouraged to increase the share of
cowpeas in their diets if the surplus production from the producing
zones was made available.
Table 8.4-- Product composition of pulses consumed by zone and year
Peanut Basin Rural Zones
Urban Zones
Central
Central
Senegal
Product
North (Niakhar) (Colobane] Southwest
Southeast
Oriental
Kaolack
Tambacounda
Peanuts
tma/a9’
50
al
83
98
95
93
HY89/90
78
79
99
07
96
7-90/6-91
a7
99
Cowpeas
HY88/89'
50
19
17
2
5
7
HY89/90
22
21
1
13
4
7-90/6-91
13
1
Source: ISRA/IFPRI project survey.
Note:
Numbers are percent of total pulse calories per adult equivalent provided by peanuts and cowpeas,
respectively.
' Data for harvest year 1988/89 cover December 1988 through September 1989 only.
"Other Foods."
"Other foods" examined to date are vegetable oil,
bread, milk, and tubers.
Figure 8.3 clearly shows that vegetable oil
makes the biggest contribution to the "other food" category in all zones
but Senegal Oriental, where milk predominates.
Bread is the second
biggest contributor in the urban zones, the Northern Peanut Basin, and
the Colobane zone of the Central Peanut Basin.
In all other zones,
bread is inconsequential and milk is the second most important product.
If a larger share of "other foods" in the diet is associated with
a more urban type of consumption, one concludes from Figure 8.3 that the
"rural" zone of the northern Peanut Basin and the Colobane zone of the
central Peanut Basin are clearly moving in the direction of an urban
consumption basket.
The oil consumption in these two zones provides
more calories per AE than the sum of all
"other food" calories for each
of the remaining rural zones.
Tubers account for a very small share of "other food" calories in
all zones.
This is generally true for the rural areas.
Tuber
consumption in urban areas is probably somewhat underestimated here due
to difficulties in obtaining accurate unit measures for tubers consumed.
162
Figure 8.3--
Composition of "other food" by product, zone, and year.
SA = Saaatta; NI, Niakhar; CO, Colobane; PA, Passy; DI,
Dioly; f?I, Missirah;
KA, Kaolack, and TA, Tambacounda.
6007
1
500 -
2'400-
\\
T
\\300-
07
!A
g 200 -
i
loo-
0
SA NI CO PA DI Ml
KA TA
1988/89
1989/90
1990/91
m VEG OIL
m WH BREAD m MILK
0 TUBERS
,..
.
.i
:
163
T y p i c a l P a t t e r n s
Modifying the three major types of cereal consumption patterns
identified above by adding considerations concerning the caloric
importance of "other foods" and relative shares of cowpeas in the pulse
diet produces the six groups shown in Table 8.5.
Cowpeas are a very small and generally insignificant part of the
diet.
Using the level of cowpea consumption to classify consumption
patterns would not normally be justified.
Given the interest in
promoting cowpea production in the northern regions of Senegal and
problems encountered in marketing surplus production, we have paid more
attention to the crop than would normally be the case.
Eliminating
cowpeas from the classification system leaves four consumption types:
(1) millet base/low "other," (2) millet base/high "other," (3) rice
base/high "other," and (4) mixed coarse grains/low "other."
It becomes apparent from this system of classification that there
is not a sharp distinction between rural and urban patterns.
Rice,
typically associated with urban consumption, predominates in one rural
zone and accounts for 25-33 percent of cereal calories in two other
rural zones (central Peanut Basin: Colobane and southwestern Peanut
Basin).
"Other foods," also associated with urban consumption, are
important sources of calories in the rural zone of the northern Peanut
Basin and the Colobane zone of the central basin.
Cowpeas, sorghum,
and maize consumption do,
however, follow
expected patterns of being much more heavily consumed in their zones of
production than elsewhere.
Given the relatively low production
sufficiency ratios reported in Chapter 5 -- even in the high potential
zones -- it is not surprising to find cereals being consumed in their
zones of production.
There was no zone where sample households produced
enough cereal to provide an average consumption of 2400 kcal/AE during
the two study years.
The three best cases were Colobane in the central
basin and the two zones in the southern basin where approximately 80
percent of the 2400 kcal/AE could have been covered by home produced
cereals in 1989/90.
Because these zones obtain about 75 percent of
their calories from cereals, there is a margin of only 5-10 percent of
production which could have been "exported" from these zones to the less
productive ones.
The only serious maize-producing area could cover
about 40 percent of their 2400 kcal/AE needs in HY88 and 60 percent in
HY89 with home produced cereals.
S o u r c e s o f P r o d u c t s C o n s u m e d
Table 8.6 and Figure 8.4 show that cereal/pulse purchases account
for the smallest amount of calories in Senegal Oriental (about 350 per
AE per day or 17 percent of cereal/pulse consumption).
In the
southeastern Peanut Basin and the Niakhar zone of the central Peanut
Basin, purchases are also low (both approximately 450 calories per AE
per day -- about 20 percent of cereal/pulse consumption).
.
.
.
.
164
Table 8.5-- Classification of consumption patterns
C o n s u m p t i o n
P a t t e r n
Cereal
“Other”
Pulse
Zone
Millet based with
Low share
Low to medium
Southeast Peanut Basin
rice supplement
share of cowpeas
Southwest Peanut Basin
Millet based with
Low share
High share of
Central Peanut Basin: Niakhar
rice supplement
cowpeas
Millet based with
High share
High share of
Central Peanut Basin: Colobane
rice supplement
cowpeas
Rice base with
High share
High share of
Northern Peanut Basin:
mixed coarse
cowoeas
Kaolack
grain supplement'
Rice base with
High share
Low share of
Tambacounda
mixed coarse
cowpeas
grain supplement
Mixed coarse
Low share
Low share of
Senegal Oriental
grain base
cowpeas
Source: ISRA/IFPRI project survey.
' Coarse grain supplement is predominantly millet in these zones with a small amount of maize.
Table 8.6 --Share of cereals and pulses purchased by zone and year
Peanut Basin Rural Zones
Urban Zones
S u r v e y
Central
Central
Senegal
Product Period
North (Ni a k h a r ) (Co1 o b a n e ) S o u t h w e s t
Southeast
Oriental
Kaol ack Tambacounda
Cereals
HY88/89' 9 0
21
48
2 4
1 7
HY89/90
-
17
4 7
2 2
1 9
7-90/6-91
-
95
9 1
Pulses
HY88/89' 4 9
1 8
1 8
1 6
1 0
1 8
HY89/90
-
4
15
10
19
1 0
7-90/6-91
-
95
82
Source: ISRA/IFPRI project survey.
Note: Numbers are percent of total calories purchased in each product group.
In rural zones, most
unpurchased cereal and pulses are home produced.
In Kaolack 1 percent of cereals are home produced
- all other nonpurchased products come from gifts.
In Tamba 1 percent of cereals and 13 percent of
pulses are home produced, the rest come from gifts.
I
' Data for harvest year 1988/89 cover December 1988 through September 1989 only.
165
Figure 8.4--
Sources of cereals and pulses by zone and year. SA =
Sagatta; NI, Niakhar; CO, Colobane; PA, Passy; DI,
Dioly; MI, Missirah; KA, Kaolack, and TA, Tambacounda.
3000
E
SA NI CO PA DI Ml
SA NI CO PA Dl Iv41
KA TA
1988/89
1989/90
1990/91
0 PURCHASES m PRODUCTION
GIFTS
166
Households in urban zones practice very little agriculture; hence,
more than 90 percent of their cereals and pulses are purchased.
Rural
zones that are described above as having more urbanized consumption
habits (northern Peanut Basin and Colobane zone of the central basin)
exhibit the largest share of cereal/pulse purchases among the rural
zones (80 and 46 percent, respectively).
The southwestern Peanut Basin
falls between the two extremes
with about 30 percent of cereal/pulse
calories being purchased.
Given that average consumption in most zones just meets minimum
requirements or falls somewhat short, this relatively high dependence on
purchased cereals and pulses implies that rural households must have
substantial cash incomes -- from sales of crop produce or other
activities -- to ensure minimum acceptable levels of consumption.
This is true following the poor 1988/89 harvest as well as the good
1989/90 harvest.
Rural zones where purchases provide the fewest
calories are also the zones that have calorie deficits.
A low share of
purchases appears to be correlated with a low level of consumption
rather than with adequate levels of production that make it unnecessary
to resort to the market.
There is also evidence that when households enter the market to
purchase cereals, there is a greater tendency to purchase rice than
coarse grains.
Assuming that households want some diversity in their
diets, it is not surprising that one would purchase cereals that differ
from those produced, hence the large share of rice in total purchases
shown in Table 8.7.
Table 8.7-- Share of rice in cereal purchases by zone and year
Peanut Basin Rural Zones
Urban Zones
S u r v e y
Central
Central
Senegal
Period
N o r t h (Niakhar) ( C o l o b a n e ) S o u t h w e s t
Southeast
Oriental
Kacl ack
Tambacounda
HY88/89'
7 4
7 3
7 0
71
35
4 1
HY89/90
7 7
66
88
2 8
1 9
7-90/6-91
6 5
7 7
Source: ISRA/IFPRI project survey.
Note: Numbers are percent of purchased cereal calories per adult equivalent provided by imported
rice.
' Data for harvest year 1988/89 cover December 1988 through September 1989 only.
Differences in Consumption Due to the Level of Commercial Infrastructure
167
One-third of the sample in each rural zone is located in a village
that is the site of a weekly market.
Many of these market villages have
other infrastructure such as permanent shops, schools, health services,
government offices (arrondissement level), peanut and cotton purchasing
agents,
and so on.
It is our hypothesis that the presence of this
infrastructure encourages a more urban food consumption pattern than
that prevailing in villages lacking such infrastructure.
The urban
consumption pattern is encouraged for three reasons: (1) supplies of
rice and "other foods" are more readily available in market villages,
(2) prices of these foods tend to be lower in market villages than in
nonmarket villages, and (3) the higher level of economic activity in the
market village makes it easier for one to earn the cash needed for
purchasing these "urban" food products.
In this section the differences in level, composition, and source
of food products consumed by households in market and nonmarket villages
is examined.
level and Composition of Caloric Intake from Cereals and Pulses
Figure 8.5 presents two sets of bar graphs showing the level and
composition of consumption by zone and year.
In general, there is a lower level of caloric intake from cereals
and pulses in the market villages.
The difference between market and
nonmarket cereal/pulse consumption ranges from about 100 calories to as
much as 500 per AE per day.
The only exception here is the southeastern
Peanut Basin where the market village exhibits about the same level of
consumption as non-market villages.
There is consistently a larger share of rice in the cereal/pulse
diet of households in market villages.
In zones where rice consumption
is generally low, market villages consume as much as 8 times more rice
per AE than nonmarket villages.
In zones where rice is more common, the
market villages consume 1.5 to 2 times more than their neighbors.
Market villages with the lowest rice consumption levels still cover 15-
20 percent of cereal/pulse calories with this imported product. At the
higher end, as much as 68 percent (northern Peanut Basin) and 50 percent
(central Peanut Basin: Colobane) of cereal/pulse calories come from
rice.
Market villages generally obtain fewer calories from pulses, but
there is no apparent difference with respect to the relative shares of
cowpeas and peanuts.
" This observation is based on partial analysis of data files that
were revised subsequent to the draft report.
Further analysis of the
revised files is required to be sure about what is happening in
individual villages.
:
_“,
I’.
.,
:
i
#
.,
I
:
*
:
.’
168
Figure 8.5--
Product composition of cereal and pulses consumption
compared for (A) nonmarket and (B) market villages.
A
NON - MARKET VILLAGES
1
C
-F ‘ROOUCT COMf’OSlTiON OF C O N S U M P T I O N
I-
/-
NOT
AVAL
IT
SA .
Y
.
m
I.4
II
Y
'*'m'fu'a'u
HY 1988/89
HY 1989/90
B
:;-j/
~300, PROOUCT COMPOSrrKxl
O F C O N S U M P T I O N
:
I
:
.
‘_
I
.
169
S o u r c e o f P r o d u c t s C o n s u m e d
Households in nonmarket villages consume N-30 percent more
calories from home production than do their market village counterparts.
As Figure 8.6 shows, this leads to a greater share as well as a greater
absolute quantity of cereal and pulse calories being purchased by market
village households.
As mentioned previously, households purchasing
more cereals and pulses appear to favor rice over coarse grain
purchases.
This is illustrated in Figure 8.7, which shows that shares
and absolute quantities of rice calories purchased are greater in market
villages.
Policy Relevance of Urbanized Consumption Patterns in Market Villages
The 1988 census made it clear that the rate of urbanization in
Senegal is quite high (3.9 percent growth rate for Dakar compared with
2.8 percent for the entire country).
Although it is unlikely that
market villages are growing as fast as Dakar, a quick comparison of 1976
and 1988 census data for rural villages suggests that market villages
are growing moge rapidly than other rural villages.
In most of the
rural study zones, market villages account for 15-20 percent of the
rural populatidn.
As it becomes increasingly difficult to make a living
with crop production and animal husbandry activities alone,
rural
populations are likely to migrate toward market villages as well as the
urban centers.
This type of migration will increase demand for rice
substantially through both demand and supply factors because market
villagers.eat more rice and produce less coarse grain cereal (see
Chapter 5) than their counterparts in nonmarket villages.
Income Effects on Consumption Patterns
In order to examine differences in consumption that might be
associated with different levels of income, we divided the sample for
each zone into quartiles and compared the cereal/pulse consumption
basket of the lowest quartile with that of the wealthiest.
In the
following paragraphs,
the differences in cereal/pulse consumption of
rich and Door households are examined in terms of levels of caloric
intake,
product composition,
and sources of products consumed for the
rural sample."
L e v e l o f C a l o r i c Intake
Most rural zones exhibited surprisingly little income effect on
total cereal/pulse calories durinq HY88. The difference between the two
groups is only about 100 calories per AE. Given the high coefficients
of variation on these data, such a small difference is not likely to be
significant.
The only exception to this tendency is the southeast
Peanut Basin where there is a difference between rich and poor of about
600 calories per AE!
This is one of the zones where a large share of
" Time did not permit us to do a comparable analysis for urban
zones.
‘(%
;
,
j
Figure 8.7--
Composition of cereals and pulses compared for (A)
nonmarket and (6) market villages.
A
-
NON - MARKi VILLAGES
D
H-f 1988/89
H Y 1989/90
RI MARKET VlLLAGES
I
I
CXMPOSITION OF
PURCHASES
.“.
;:
,,
i
*’
:
.’
.
.
”
172
households (60-70 percent) failed to cover minimum needs of 2400
kcal/AE.
Following the much better 1989/90 harvest, there is a larger
difference between rich and poor, with the rich consuming from 150
(Colobane in the central basin) to 700 (southeastern Peanut Basin) more
calories per AE than the poor in all zones but Senegal Oriental.
Figure 8.8 illustrates these points by comparing levels and
composition of cereals and pulses consumed by the lowest and highest
income quartiles in each rural zone.
The top half of the figure
illustrates consumption patterns following the poor 1988/89 harvest and
the lower half represents consumption following the good 1989/90
harvest.
Source of Ca7ories by Income Group
The fact that there was very little improvement in caloric intake
for the poor following the good rainfall of 1989/90 suggests that (1)
production constraints prevent them from reaping the benefits of better
climatic conditions and (2) lack of alternatives to crop income keep
them from supplementing cereal/pulse production with cash purchases of
food.
The rich,
on the other hand,
appear
to have produced
substantially more food in HY89 than they did in HY88. The increase in
production translated into an increase of from 300 to 700 calories per
AE per day more coming from home production in HY89.
Figure 8.9 illustrates this point by comparing the source
composition of cereal and pulse consumption between the two harvest
years.
Senegal Oriental is the only zone where the wealthy were worse
off in HY89 and the poor had a better harvest."
Figure 8.9 also shows that both the rich and the poor relied
heavily on purchases of cereals and pulses in the northern Peanut Basin
(56 percent for the poor and 90 percent for the rich in HY88) and in the
Colobane zone of the Central Peanut Basin (53 percent in HY88 and 45
percent in HY89 for the rich;
50 and 62 percent, respectively, for the
poor).
In other words, both the rich and the poor in these zones share
a tendency to purchase a large part of their food.
Even in the zones relying the least on purchases (Niakhar in the
central Peanut Basin,
the southeastern Peanut Basin and Senegal
Oriental),
a minimum of 15-20 percent of cereal/pulse calories is
purchased.
This leads us to the question of whether these zones are all
deficient in cereals or if they are selling coarse grains and purchasing
rice to introduce variety into the diet.
The discussion of crop
production in Chapter 5 confirms that the average household
" This may be due to households who received large amounts of
migration remittance toward the end of the year being classified as
wealthy when they were really poor for most of the year.
.I
:
:
:
:
.
.
.
‘.
.
i
Figure 8.8--
Income effects on level and composition of cereal and
pulse consumption for two harvest years. NPB, Northern
173
Peanut Basin; CPBN, Central Peanut Basin - Niakhar-
CpBC, Central Peanut Basin, Colobane; SWPB, South&t
peanut Basin; SO, Senegal, Oriental.
A
JO00
2750
2500
2 2250
E 2000
a
w 1750
Q:
E 1500
2 1250
?I!
ig 1000
5
750
500
250
0
LOM
Lo M
Lo H
NPE
CPB.C &P;
so
mw=wJn~m mpwcm LYEiflm
B
2750
NPB
CPB.N
CPB.C SWPB
SEPB
s o
uzw~wo~RRxT
m-s
EfEamwPw
Figure 8.9--
Income effects on sources of cereals and pulses
consumed for two harvest years.
NPB, Northern Peanut
Basin; CPBN, Central Peanut Basin - Niakhar; CPBC,
174
Central Peanut Basin, Colobane; SWPB, Southwest Peanut
Basin; SO, Senegal, Oriental.
A
btu
IAH
Lola
LDHI
Lota
Lo H
NPB
CPB.N
CPB.C SWPB
SEPB
so
E3l-m-M~
1
B
Lo HI
Lo H
Lo HI
Lo Hl
Lo HI
Lo H
NPB
CP8.N
CPB.C SWPB
SEPB
SO
Ex-rsa- nlx-rs
3
. .
!..
,’
175
across all zones fails to produce 2400 kcal/AE of cereals.
Even when
one assumes that the caloric demand for coarse grains in rural Senegal
is only 1900 kcal/AE, home production only covers these needs in the two
zones of the southern Peanut Basin and only after the better than
average 1989/90 rains.
P r o d u c t C o m p o s i t i o n o f C o n s u m p t i o n
In HY88 the wealthier households consumed more rice than the poor
(Figure 8.8).
This is true for absolute quantities per AE as well as
share of rice in the diet.
Senegal Oriental was the only zone where the
poor ate more rice than the rich (see footnote 15).
Following the good 1989/90 harvest, the absolute amount of rice
consumed went down for both income groups in all zones with the
exception of the poor in the Niakhar zone of the central Peanut Basin
where rice consumption increased by 69 calories per AE.13
In HY89 wealthy households in all rural zones consumed more rice
per AE than the poor; the share of rice in the diet of the rich was,
however, substantially greater in only two zones (central Peanut Basin -
Colobane,
and southwestern Peanut Basin).
A large increase in
consumption of home grown millet was largely responsible for the reduced
share of rice eaten by wealthier households in HY89.
In HY88 the poor obtained 8 to 39 percent of their cereal/pulse
calories from rice, depending on the zone.
In HY89 the range was 4 to
24 percent.
Zones with the higher shares in rice are the most densely
populated; this means that a large part of the rural poor depend on rice
for about 20 percent of their cereal/pulse calories.
The fact that the rich eat more rice than the poor does not
necessarily mean that they eat less millet.
In HY88 the poor consumed
mire calories per AE from millet than the rich in every zone.
In HY89
this was true only in Senegal Oriental. In other zones, the poor ate
an equal or smaller absolute quantity per AE than the rich. Millet did,
however, account for a slightly larger share of the cereal/pulse diet in
poor households than in rich ones due to the lower overall level of
consumption by the poor in HY89.
In terms of pulse consumption, the biggest distinction between rich
and poor is found in the northern Peanut Basin where the poor consumed
substantially more cowpeas late in HY88 than the rich.
This came about
because the 1989 cowpea harvest was early (August) and abundant,
permitting households with inadequate food supplies to supplement millet
l3 Analysis of agricultural product transactions data in Chapter
6 suggests that increased rice consumption following a good harvest may
be more prevalent than a partial analysis of the lowest and highest
income quartiles suggests.
The effect of increased income due to a good
peanut harvest appears to encourage households to supplement their
millet consumption with purchased rice.
;
.’
:
176
and rice dishes with very inexpensive cowpeas.
On an annual basis,
cowpeas provided 14 percent of the poor's cereal pulse calories but only
5 percent of those for the rich.
There is no clear pattern linking income class and absolute
quantity or share of total calories obtained from peanuts -- sometimes
the poor eat more and sometimes the rich, depending on the year and the
zone.
Composition of Purchases
Figure 8.10 illustrates the product composition of purchases. Rice
is generally a more important component (in absolute quantity and share)
of purchases made by the rich than those made by the poor. An exception
to this rule is found in the northern Peanut Basin in HY88 where poor
households purchased less rice per AE but the share of their purchases
devoted to rice was quite high (70 percent vs. 57 percent for the rich).
Millet is generally a more important component (in absolute
quantity and share) in the purchases of the poor. Exceptions to this
were the northern Peanut Basin, where the rich purchased three times
more millet than the poor in HY88, and the southwestern Peanut Basin,
where the share of millet in purchases made by the rich was 6 percentage
points higher than the share for the poor.
There are no striking differences in purchasing patterns for pulses
between the rich and the poor that appear in both harvest years.
Absolute quantities and shares of purchased calories coming from pulses
differ little by income group.
The only exceptions occurred in HY89:
cowpeas accounted for 10 percent of purchased calories for the rich and
only 5 percent for the poor in the southeastern Peanut Basin; and
peanuts accounted for 23 percent of purchases by the poor, but only 6
percent of purchases by the wealthy.
Seasonal Chanqes in Consumption Patterns14
It is difficult to compare seasonality across zones in the
ISRA/IFPRI sample because the onset of the different seasons does not
fall in the same calendar month in all zones.
To simplify the
presentation,
four basic seasons are used in all zones:
l4 This section has not been revised to incorporate the higher
average levels of consumption calculated for the southeastern Peanut
Basin since the draft report was issued. It is unlikely that the higher
levels will substantially change the pattern of seasonality.
.
.
.’
‘.
:
Figure 8.10--
Income effects on composition of cereals and pulses
177
purchased for two harvest years. NPB, Northern Peanut
Basin; CPBN, Central Peanut Basin - Niakhar; CPBC,
Central Peanut Basin, Colobane; SWPB, Southwest Peanut
Basin; SO, Senegal, Oriental.
1750
PRODUCT COKPOSll7ON
OF PMXXASES
2 1500
E
a 1250
2
E 1000
E
c
750
f5
ij
500
250
NPB
CPB.N
CPB.C SWPB
SEPB
so
palucr tEEi= Emu
o=
IxxImm-
2000
1750
HARVEST YEAR 1989/90
>
PROCKT CoMposmoH OF PLIXtiASEs
Q 1500
n
(II
2 1250
W
Q
[r
1 0 0 0
t
z
750
z
d
500
0
0
Lo HI
Lo H
La Hl
Lo
M'
'Lo'HI'
'Lo'H
'
NPE
CPB.N
CPB.C SWPB
SEPB
so
mu msa=w ez4~
Elm
Ixxl-EEEI-
178
SEASON
CALENDAR MONTHS
Harvest period
October - December
Peak period for non-
agricultural
activities
January - March
Period of
clearing fields
and planting
April - June
Period of
weeding
July - September
The major problem encountered with this classification is that the
early rains in 1989 led to early harvests, particularly of cowpeas and
millet in the northern and central Peanut Basin.
The early harvest had
the unusual effect of increasing consumption of home produced crops
during the weeding season -- a period that one would expect to exhibit
serious calorie deficiencies following a harvest as poor as that for
1988/89.
Another shortcoming of this seasonal classification is that data in
the weeding season are an average over three months that can exhibit
very different calorie levels and composition.
An earlier report
presented data for August, September, and October for several zones.
Readers interested is the details of month-to-month changes in level of
caloric intake should consult this document (Kelly et al., February
1991).
The following paragraphs discuss seasonal changes in level,
composition,
and source of cereal and pulse consumption.
Leve7 of Ca7oric Intake
The level of caloric intake is examined from two perspectives: (1)
changes in absolute level and (2) percent fluctuation about the average
annual level.
Part A of Figure 8.11 shows average calories consumed per AE per
day during the harvest season for each zone and year covered by the
rural survey.
Each bar in the graph shows the average level per AE of
cereal and pulse calories consumed for each zone/year combination.
Parts B-C of the same figure illustrate consumption patterns for the
other three seasons.
Given that only cereal and pulse calories are
examined in this section, 2200 calories is used as a rough indicator of
minimum acceptable consumption levels (with this level of cereal/pulse
consumption,
most will attain the 2400 mark with "other foods"; 1900
represents the level below which we believe populations are clearly "at
risk" because it is unlikely that they would obtain 500 calories per AE
of "other food."
:
.
,.;
.
.
:
.
c
INTER-ZONE COUPARISON
HARVCST
PEnlOO CONSUMPTION
.
a
HY 19aa/a9
H Y 1989/90
i
J
.
.
B
r-
.
.
,cm
,I%0
INTER-ZONE COUPARISON
,%a I-----
MN - t<G PERIOD CONSUUPTDN
-
.
_
: .
180
At harvest time, 6 of the 11 bars are equal to or greater than 2200
calories;
this declines to 5 bars in the nonagricultural season and 3
bars in the planting and weeding season. Hence, at the upper end of the
scale, almost half the zone/year averages are below the 2200 level even
at harvest.
The number of zones falling below this cutoff point
increase throughout the year,
leaving only three cases above the 2200
level during both the planting and the weeding season.
Of the six cases maintaining this relatively high level of cereal
pulse consumption,
the southwest Peanut Basin appears three times
(planting seasons 1988 and 1989, and weeding season 1989). Niakhar in
the central Peanut Basin -- appears two times (planting and weeding
1989) and Colobane, also in the central basin, once (weeding 1988).
This result reflects the relatively strong preference for cereals vs
"other foods" in two of the zones (Niakhar in the central Peanut Basin -
and the southwestern Peanut Basin) as well as the ability of households
in these zones to manage stocks and finances to avoid periods of low
caloric intake.
Looking at the lower end of the scale, there is no zone/year
combination with average consumption below 1900 calories per AE at
harvest time,
one case during the HY88 nonagricultural season, four
cases during the planting season, and just two during the weeding
season.
Zones falling into this "at risk" group are the northern Peanut
Basin (nonagricultural, planting, and weeding HY88), Senegal Oriental
(planting HY88 and HY89 and weeding HY89), and the southeastern Peanut
Basin (planting HY88).
The fact that the northern Peanut Basin falls into this at risk
grouping is of less concern than the presence of the two other zones due
to the very high levels of "other foods" consumed in the north (about
512 calories per AE per day).
The low levels of cereal/pulse
consumption,
coupled with extremely low levels of "other food"
consumption suggest that one would find evidence of nutrition related
health problems in both the Southeastern Peanut Basin and Senegal
Oriental.
In looking at percent of variation around the mean level of
consumption for each zone/year combination, we find that seasonal
differences are usually not more than plus or minus 5 percent of the
annual average.
During the harvest period, consumption ranges from
0.98 to 1.15 of the annual average with 8 of the 11 observations falling
between 100 and 106 percent of the average.
During the nonagricultural season, the range is 0.93 to 1.07, with
6 observations equal to or below the annual average and 5 above. Moving
into the planting season, the range is 0.93 to 1.05 with only 2 cases
where consumption is equal to or greater than the annual average.
During the weeding season, the picture for HY88 differs from that
of HY89.
This difference is due to the early 1989/90 harvest, which
increased consumption during the HY88 weeding season, and the late
1990/91 harvest which was not available until late October 1990. IN
.
._
’
:
.
:
.
:
181
HY88 only two of the 6 zones fell below the annual average; in HY89 all
5 zones fell below, with the range being from 91 to 98 percent of the
average.
By way of summary,
we conclude that seasonal variation in
consumption levels exists but stays within a band of plus or minus 5
percent of the annual average in most cases. The exception to this rule
is Senegal Oriental, which has quite large fluctuations (plus 11 percent
in harvest period 1989/90 and minus 9 percent in weeding period
1989/90).
This wide range in seasonality is of concern because the
overall zone average is quite low.
It suggests that households in the
more drought-prone areas of the Peanut Basin have learned to better
manage food stocks and cash reserves,
thereby
ensuring less
interseasonal fluctuation in consumption levels.
In modeling demand for cereals and pulses one would normally expect
there to be a seasonal effect on the levels of demand (declining levels
of consumption as one moves from harvest to harvest) as well as a
harvest effect (better harvest, higher level of caloric intake across
all seasons).
The discussion above suggests that the seasonal effect
exists but is relatively small.
Earlier discussion concerning average
levels across years suggested that the harvest effect also exists; it is
also generally small (l-5 percent for all zones but Niakhar in the
central Peanut Basin, which had a 12 percent increase in average
consumption between HY88 and HY89).
We have, however,
found some tendencies that disturb these
apparently general patterns.
For example, even though the harvest
effect was relatively small in most of the zones, there was a very
significant harvest effect across
all zones for the wealthiest
households whose millet production rose significantly between HY88 and
HY89.
The general seasonal trend is also rendered more complex by an
increase in consumption levels between the planting and weeding season
in HY88 but a decline during the same period in HY89 due to interannual
differences in the timing of the rains and, subsequently, the harvest,
This discussion suggests that in modeling demand it may be
necessary
to introduce variables that will capture (1) harvest
quantity/quality, (2) harvest timing, (3) seasonal effects, and (4) some
type of interaction effect between income and harvest quantity/quality.
Composition of Consumption
The most striking aspect of seasonal changes in composition is the
role played by cowpeas during the weeding season of HY88. Without the
increase in calories obtained from cowpeas during this season, the
average cereal/pulse consumption in the north would have been about 1525
calories per AE.
Although less important in the central basin zones of
Niakhar (4 percent of total cereal/pulse calories) and Colobane (6
percent) than in the north (19 percent), it clearly helped households in
these zones get though a difficult period. These findings suggest that
expansion of cowpea production in the northern and central Peanut Basin
.,;
*
,:
.
.,;
.
.
:
;-
..
182
can be justified on nutritional grounds in addition to the more common
argument that cowpeas offer farmers in the zone a cash crop alternative
to peanuts.
Seasonal and interannual differences in rice consumption are also
noteworthy.
Part A of Figure 8.11 shows that rice consumption does not
disappear in the central and southwestern Peanut Basin during the
harvest season when new coarse grains are readily available. We would
expect relatively high rice consumption during the 1988 harvest period
in the northern and central Peanut Basin because their cereal harvests
were generally poor.
Rice remained important during the harvest season
of HY89 in all these zones.
In the Niakhar zone of the central Basin
and in the southwestern Basin, the HY89 level is about the same as that
in HY88;
rice consumption drops to 67 percent of the HY88 level in the
Colobane zone of the central Peanut Basin but remains high at 577
calories per AE per day.
This relatively high level of rice consumption during a period when
coarse grains are readily available and inexpensive is probably due to
the effect of large amounts of cash income from peanut and cotton sales
coming in during this same period.
The presence of the income effect
would have to be verified for each zone by looking at the exact timing
of the cash crop sales, but the official cash crop marketing period
usually opens in December and continues until March.
Tastes must also play a part because rice consumption virtually
disappears in the southeast Peanut Basin and Senegal Oriental during the
1989 harvest, suggesting that when coarse grains are available in these
zones, households show little interest in rice.
Tracking the ups and downs of rice consumption across the four
seasons, we find three patterns.
(1) Small, incremental increases of the share of rice in the diet from
harvest to harvest (Niakhar zone in the central Peanut Basin and
Senegal Oriental).
(2)
Large and relatively constant shares of rice throughout the entire
year (northern Peanut Basin, Colobane zone of central Basin,
southwestern Peanut Basin).
(3) A relatively small share of rice in total cereal consumption rising
from the harvest through the nonagricultural season and then
falling during the planting and weeding season (southeastern Peanut
Basin).
The first pattern is probably a reflection of three factors working
together:
(1) households in these two zones have strong taste
preferences for coarse grains,
(2) they do not have liquidity to
purchase coarse grain stocks during the harvest period when prices are
low (see Chapter 4 on income), and (3) as the seasons progress, the
price of rice relative to coarse grains is declining (rice price i's
constant but the coarse grain prices are rising).
1 8 3
The second pattern
is probably a reflection of tastes and
commercial
infrastructure
which
makes
rice
readily
available,
facilitates sales of agricultural produce, and provides opportunities
for nonagricultural income.
Although every zone has a market village,
these three zones have bigger, more important markets than the other
zones.
One of the three zones has a chronic cereal deficit. Cereal
production per AE for Colobane
in the central basin and in the
southwestern basin are among the highest in the study, yet purchased
rice remains important in these zones, suggesting that factors other
than production are driving the demand.
The quantity of rice consumed in the third case (southeastern
basin) is so small that it is hard to know if this is a real pattern.
One frequently hears farmers saying that only coarse grains provide the
energy required for the heavy field work during the planting and weeding
seasons.
This could possibly be the explanation for the higher rice
consumption during the periods of less intense agricultural work.
Seasonality o f P u r c h a s e s
Figure 8.12 (A-D) illustrates how the level of calories and the
product composition of purchases vary across zones in each of the four
seasons.
The level of calories obtained from purchased cereals and
pulses increases constantly from harvest to harvest in almost all zones.
Purchases start out relatively low during the harvest period. In most
zones levels purchased during the nonagricultural and planting season
are 1.5 to 2 times that purchased during the harvest period. During the
weeding season there tends to be another big jump, with amounts
purchased increasing another 1.5 to 2 times.
The composition of cereal purchases is mixed. Rice is an extremely
important share of purchases at harvest time because most households
have their home-produced millet to consume. During the nonagricultural
season the quantity of rice rises but the share remains steady or
declines as households in some zones begin purchasing millet.
This
behavior continues with both millet and rice purchases increasing in
tandem across the seasons until we reach the weeding period, when rice
purchases decline and millet increases.
Without examining the price data here, it is difficult to speculate
about the role of prices versus preferences in determining the mix of
cereals purchased.
Since the government-controlled price of rice does
not vary during the year, price response to rice is a function of the
relative price of rice and coarse grains.
When coarse grains are
inexpensive,
rice is most expensive; when the price of coarse grains
rises, rice becomes relatively less expensive.
:
:
_:
.
. 1
.‘.
Ava Y3d 3V Y3d S3lYOlV3
IV0 tr3d 37 Y3d S3tYOlY3
i
:
.
,#’
.,
I’
.
:.
.
’
.
.
185
The high processing losses associated with coarse grains means that
a simple comparison of per kilo prices for millet and imported rice is
inappropriate.
It is informative to compare the price of rice and
coarse grains per calorie rather than per kilo; most Senegalese cooks
probably do this unconsciously because they know that they need more
kilos of millet than rice to feed a given number of people. Using the
government price of 135 francs CFA (FCFA)/kilo of imported rice, the per
calorie price is 0.379 FCFA.
When the price of a coarse grain reaches
107 FCFA per kilo, this is equivalent to 0.379 FCFA per calorie. In
other words, a consumer interested in purchasing the most calories per
FCFA should be indifferent between rice at 135 FCFA per kilo and millet
at 107 FCFA per kilo.15
If price were the major factor in determining cereal purchases, we
would expect coarse grains rather than rice to be purchased during the
harvest period because the millet price was never above 100 FCFA at this
time.
We would also anticipate rice purchases to grow faster than
coarse grain purchases as the seasons progress and coarse grain prices
rise. This does not happen in most zones; millet purchases increase
significantly during the weeding period and rice purchases decline
somewhat.
Factors other than price are also at work here. Households
appear to have a preference for a particular mix of rice and coarse
grains.
As stocks of coarse grains decline, it is not a matter of
switching over entirely to rice but of purchasing a mix of the two types
of cereal.
Purchases of pulses account for a relatively small share (less than
10 percent) of total cereal/pulse purchases.
Peanuts are rarely
purchased during the harvest and nonagricultural season, but quantities
increase during the planting and weeding seasons. With the exception of
the Northern Peanut Basin, cowpeas are rarely purchased. This suggests
that there is room for developing markets for cowpeas outside the
principal zone of production.
Statistical tests of factors influencinq consumption patterns
Two-by-two tables and chi square tests (.05 level of significance)
were used to assess the statistical significance of factors that appear
to influence the level of caloric intake and the share of rice in cereal
consumption.
Differences in the level of caloric intake were found to be
statistically different across agroclimatic zones.
Households in the
higher potential zones of Senegal Oriental and the southeastern basin
I5 We are ignoring here the fact that the cost of the sauce that
goes with the millet or rice might differ and the decision to purchase
one cereal or the other involves an evaluation of the cost of the entire
meal, not just the cereal.
This is a more important consideration in
urban areas and market villages where "other foods" are a substantial
part of the diet, and less important in villages where sauces are made
from gathered products or home produced condiments.
186
had a greater probability of falling below the 2400 kcal/ae minimum
needs level while households in the Colobane zone of the central basin
and in the southwestern basin were more likely to exceed minimum needs.
Members of the Serer ethnic group are more likely to exceed the
2400 kcal/ae level while the Wolof, Manding, and Diakhanke are more
likely to experience inadequate consumption levels. Ethnic group and
agroclimatic zones are highly correlated, making it difficult to assess
the individual importance of each factor.
Households in market villages were shown to be statistically more
likely to purchase larger shares of their cereals, to consume larger
shares of rice, and to have a lower level of caloric intake from goods
analyzed.
Households that cultivated less than one hectare per AE (all zones
combined) were more likely than others to exhibit inadequate levels of
caloric intake.
We were unable to show any statistically significant link between
levels of caloric intake and dependency ratios, income categories,
degree of income diversification, cereal sufficiency ratios, and the
share of rice in household cereal consumption.
In other words, the
analyses conducted to date have not provided statistically significant
support for the hypotheses that these factors influence the level of
caloric intake.
More thorough forms of multivariate analysis will be
able to shed further light on these relationships.
Households living in a market village, earning more than 30 percent
of total household income in off-farm activities, and having lower
cereal production sufficiency ratios have a statistically significant
tendency to consume more imported rice than their neighbors.
Other
factors that one might expect to influence the share of rice in the diet
(dependency ratio, income category, and cultivated area per AE) were not
found to be statistically significant.
SUMMARY OF MAIN POINTS
By way of summary it is interesting to group the findings in three
categories:
conventional wisdoms confirmed, conventional wisdoms not
confirmed, and unanticipated results.
In the category of "conventional wisdoms confirmed" we include the
following points:
Millet is the most commonly consumed cereal in the Peanut Basin
(two-thirds of calories consumed) except in the north where
imported rice dominates.
In urban areas, rice provides 100 percent of cereal calories at
noon and 50 percent in the evening.
:
187
Rice consumption has made important inroads in rural areas,
particularly in the north, in the Colobane zone of the central
basin, and in the southwest (70, 32, and 26 percent of cereal
calories, respectively).
Maize consumption is important only in its zone of production,
Senegal Oriental (50 percent of rural and 12 percent of urban
consumption).
The northern Peanut Basin exhibits an essentially urban consumption
pattern due to its low levels of agricultural productivity and high
share of off-farm income in total income.
Purchases are an important share (at least 20 percent) of all
cereals consumed in rural areas; greater purchases are encouraged
by low levels of agricultural productivity and the monetization of
the rural economy (both agricultural and nonagricultural sectors)
which facilitates purchases of basic needs.
The consumption patterns in market villages are becoming
increasingly urbanized (more rice and more noncereal, nonpulse
products).
Overall, an important share (30 percent or more depending on year
and zone) of rural households do not meet the minimum levels of
caloric intake (2400 kcal/AE).
The most satisfactory levels of
consumption are found in the southwestern Peanut Basin and the
Colobane zone of the central basin.
In the category of "conventional wisdoms not confirmed" we have the
following finding:
The role of factors such as dependency ratio and level of income do
not seem to have a statistically significant effect on the levels
of caloric intake in the analyses conducted to date.
Given that
these are factors that normally exhibit an important effect on food
consumption patterns,
the final word on this issue will have to
await more thorough statistical analysis of the data.
The most surprising finding is that there appears to be a more
serious calorie deficit in the zones considered to have better
agricultural potential -- Senegal Oriental and the southeastern Peanut
Basin.
Limited access to off-farm incomes (particularly in the
southeast)
and
relatively less developed commercial
and
road
infrastructure (particularly in Senegal Oriental) are two factors that
might be keeping households in these zones from earning adequate
agricultural incomes and smoothing total household income throughout the
year.
The most important policy implication coming from the analysis of
consumption patterns is that purchased cereals play an important role in
assuring food security for both urban and rural households.
Imported
rice constitutes a substantial share of these purchases in urban areas
188
(70 percent) and throughout most of the rural zones covered in the
Peanut Basin (at least 20 percent). For this reason, it is extremely
important to examine the potential effect that a change in cereal prices
(either rice or millet) might have on the real income of Senegalese
households.
This issue is examined in greater depth in the next
chapter.
.’
:
‘.
189
REFERENCES
Fall, A.A., V. Kelly, and T. Reardon.
1989. Household-Level Survey Methods
Used in the IFPRI/ISRA Study of Consumption and Supply Impacts of
Agricultural Price Policies, IFPRI/ISRA Project Document 2, September,
IFPRI, Washington, D.C.
Kelly, V., T. Reardon, A.A. Fall, 6. Diagana, and L. McNeilly. 1992. Food
Consumption in the Peanut Basin and Senegal Oriental: October 1988 -
September 1989.
IFPRI/ISRA Project Document 7, September IFPRI,
Washington, D.C.
Hellegouarch, R., R. Giorgi, L. Monjour, and J. Toury. 1988. Enquete de
Consommation alimentaire dans une Zone pilote du Seneqal (1968).
Document du Travail, Organisme de Recherches sur 1'Alimentation et la
Nutrition Africaines (O.R.A.N.A.), Dakar.
Thirion, M-C. and J-L. Garnault. 1991.
"Le sous-espace ouest tourne vers
la mer," La lettre de Solaqral - Strateqies Alimentaires, no. 37,
April-May.
190
9.
PRICE RELATIONSHIPS AND THE IMPLICATIONS FOR CHANGES IN
RELATIVE PRICES
INTRODUCTION
The objectives of this chapter are two-fold:
(1)
to describe the methods used to develop price series for cereals
and pulses and make some observations about the price trends and
relationships during the survey period; and
(2)
to summarize results of price simulations conducted to examine the
potential household level impact of changes in relative prices of
principal crops.
PRICE TRENDS AND RELATIONSHIPS
A complete and accurate set of prices for the commodities of
interest is crucial to understanding the demand and supply behavior of
rural and urban households.
The IFPRI/ISRA survey was designed to
collect unit prices for all the agricultural products bought and sold by
sample households.
The method of collecting the data is described
briefly in Chapter 6 and in more detail in Fall, Kelly, and Reardon
(1989).
Two sets of consumer and producer prices were developed for each
study zone.
The first is a set of transaction-derived prices based
entirely on household-level survey data (Tables A9.1-A9.14, Appendix
9.1).
The second series uses the household data as a base but
supplements it with information from the ISRA/MSU market surveys
conducted from 1985 through 1988 and the CSA market surveys conducted
from the mid-1980s to the present (Tables A9.15-A9.20).
At present, the consumer price series is considered acceptable for
modeling demand and has been used in a number of preliminary runs. The
producer price series is still not complete. The very poor harvest in
1988 made it extremely difficult to obtain household-level observations
for producer prices
because there were few sales by the sample
households.
Because very few peanuts are sold after the end of the
official peanut marketing program,
it was also difficult to get a full
price series for the April through December period in both harvest
years.
Efforts to supplement the household data with market surveys are
still under way.
Appendix 9.2 discusses some potential problems with
the quality of the price data for particular products and zones.
,
191
Transaction-Derived Prices
A transaction-derived price (TDP) was calculated for each purchase
and
sale
recorded on
the
agricultural
product
transactions
questionnaire.
When unit prices were provided by the respondent, the
unit price was converted to a price per kilo using unit weights. If
unit price or unit weights were not available but total expenditure for
a transaction and total kilogram weight were available, this information
was used to calculate a price per kilo.
Monthly TDPs were then
calculated as the averages of all the prices recorded in each month
(weighted by the kilograms transacted).
Before the monthly averages were taken, one modification was made
to correctly match the prices with monthly time periods.
Initially,
each transaction price was allocated to the month of the interview, with
no regard for the recall period.
However, transactions with a recall
period that straddled more than one month were assumed to have been
distributed evenly over the recall period for the interview.
Consequently, these transactions were split into parts proportional to
the number of days falling in each month.
The final step was the
calculation of a TDP for each month.
For example, if 2 kilograms of
millet were purchased for 140 FCFA in the two-week period prior to June
7, then 1 kilogram would be allocated to June and one to May.
Both
months would then have a line of data for a millet transaction at a TDP
of 70 FCFA per kilogram.
The TDP series runs from October 1988 to October 1990, except for
the zone of Sagatta, where the series stops with October 1989. Separate
series were compiled for millet and/or sorghum, maize, unshelled
peanuts, shelled peanuts, and shelled cowpeas. Unfortunately, there are
few cases of a complete series for a good. In addition, several of the
prices are based on only a small number of transactions, raising the
question of their accuracy in reflecting market conditions. The next
section discusses how these missing and "thin" prices were estimated for
the full price series.
Full Price Series
In moving from a TDP series to the full price series, three
problems had to be addressed:
l
Measurement error from unit weights
l
Thin prices (few observations per month)
0
Missing prices
Measurement errors entered the analysis when the quantity per unit
was converted into kilograms
for all transactions that were not
.
.
192
originally made in kilograms.
A decision was made to use only
transactions enumerated in kilos or very common measures (for example,
tomato paste cans) for the price series.
Since most of the transactions
in the study zones were purchased in kilograms, this decision meant
eliminating fewer than 10 percent of observations.
Missing prices for the major crops (millet, peanuts, and so on)
were filled in using extrapolation techniques and secondary data, In
addition, "thin" prices based on only one or two transactions which did
not seem in line with seasonal trends were eliminated and then
estimated.
A number of procedures were used to estimate missing or
eliminated values; among the most common were:
Extrapolation by using the average price from adjoining
months;
Direct borrowing of a data point from the market survey data
if overlapping months of TDP and the market data were
comparable;
Borrowing with adjustment from market survey data if
overlapping months of TDP and market data followed a similar
trend but were not at identical levels;
Replacing the TDP with market survey data for months where
TDP data were very thin and contradicted other price
information.
Observations on Consumer Price Trends
This section discusses consumer price trends briefly to ascertain
whether they conform to predetermined expectations.
The discussion
refers exclusively to the full price series (Tables A9.15-A9.20).
Imported rice prices are government controlled and should be
consistently in the 130 to 140 FCFA/kilo range for the entire survey
period.
Prices in Senegal Oriental and the Southern Peanut Basin should
be somewhat greater than those closer to Dakar because of transportation
costs (the amount of the transport cost is also controlled).
Tables A9.1-A9.8 confirm these expectations. The price of rice
throughout the Peanut Basin is in the 130-135 FCFA/kilo range.
It is
lowest in the Colobane zone where households frequently purchase'large
sacks and receive slightly lower prices.
It is highest in Senegal
Oriental where it reaches a maximum of 140 FCFA/kilo.
Unlike rice, coarse grains are not subject to government price
controls.
Prices for coarse grains should be lowest at harvest
(October-November) and highest during the rainy season (July-September)'.
Prices should also be higher in the poorer harvest year (1988/89), and
,
.
.
.
;
Figure 9.1 -- Consumer price trends for millet/sorghum. (Source: ISRA/IFPRI
survey data, adjusted.
Note:
Prices in Missirah are for sorghum, while those
in other zones are for millet.)
121I-
IIC)-
IOC l-
80
70
60
50.
40.
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
Oct88
I
I
Jan89
I
Apr89
Jul89
I
Oct89
Jan90
Apr90
Jut90
Oct90
+ SAGATTA
- + - , NIAKHAR
~~*~~ COLOBANE
--El- MISSIRAH --X- PASSY
-& DIOLY
194
in the zones with greater production deficits (the Northern and Central
Peanut Basin in 1988/89 and Senegal Oriental in 1989/90).
Figure 9.1 shows millet and sorghum price trends for all rural
zones.'
In general the prices follow the pattern described in the
preceding paragraph. In Senegal Oriental, however, there are a number
of unexplained ups and downs:
these are more pronounced in 1988/89 than
in 1989/90.
In Passy and Colobane, prices are a bit more volatile than
in Niakhar and Sagatta.
The prices tend to rise in January and then
fall again until April/May when they begin to rise again and do so
continuously until the harvest.
The Southeastern Peanut Basin also
exhibits this tendency, but with much smaller movements. These price
movements reflect the fact that farmers start marketing peanuts in late
December and no longer need to sell cereals to obtain cash -- this cuts
down on cereal supplies available in the markets.
In comparing prices across zones, one notes that millet prices are
consistently higher in the Niakhar zone of the Central Peanut Basin
during both years.
This is true even though the Northern Peanut Basin
had a much lower production sufficiency ratio in 1988/89 than Niakhar
and is more distant from the cereal production zones.
Prices in the
Southeastern Peanut Basin are most often the lowest while prices in
Passy and Colobane tend to move around between the two extremes.
Maize prices are illustrated in Figure 9.2. Again, the highest
prices are in Niakhar zone of the Central Basin, with Colobane and
Sagatta prices being also quite high.
The Colobane price for 1989/90,
however,
is quite a bit lower than it was in 1988/89; we have no
explanation for this.
The lowest prices are in the production zone of
Senegal Oriental (Missirah), the Southwest Basin (Passy), and the
Southeast (Dioly).
The price series for Dioly remains incomplete for a
substantial number of months.
The seasonal trends are about as
anticipated,
with Missirah exhibiting almost as much volatility in
1988/89 maize prices as it did in millet/sorghum prices.
Maize prices are consistently greater than millet and sorghum
prices in all zones.
This is illustrated more clearly in Figures A9.1-
A9.6. The fact that more maize than millet is sold in already processed
form could be causing a slight upward bias to the maize prices (see the
discussion in Appendix 9.2), but this should not be a major problem
because much of the maize price series comes from market rather than
TDPs.
' Figures A9.1-A9.6 in Appendix 9.3 present price trends individually for
each zone.
Zone-specific trends are easier to see on these figures than on the
cross-zone comparisons presented in the text.
.’
:.
.
.
:’
‘.
’
.:
.
Figure 9.3 -- Consumer price trends for shelled peanuts. (Source: ISRA/IFPRI
survey data, adjusted.)
300-
2 8 Cl-
2 6 0l-
2 4 0
2 0 0
180
1 6 0
140,
IZO-
1 oo-
80- I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
Oct88
I
Jan89
I
I
Apr89
Jul89
octa9
Jan90
Apr90
Jul90
Oct90
-U- SAGATTA
-+- NIAKHAR
-3++. COLOBANE
-El-- MISSIRAH -X-- PASSY
-.A- DIOLY
Figure 9.4 -- Consumer price trends for cowpeas.
(Source: ISRA/IFPRI survey
data, adjusted.)
26CI-
24C)-
22cI-
200I-
180
160
. .
100
80
_:
60,
40,- -I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
-
~~
Oct88
Jan89
Apr89
Jul89
Oct89
Jan90
Apr90
Jul90
Oct90
-1c- SAGATTA
-+-- NIAKHAR
-316.~ COLOBANE
-B-- MISSIRAH ---X--. PASSY
-A- DIOLY
198
Figure 9.3 presents the rather complex picture of consumer prices
for shelled peanuts.
When they are not sold in official channels,
peanuts are often sold in shelled form. On average, a kilo of unshelled
peanuts yields 700 grams of shelled peanuts.
This implies that the
shelled price should be at least 1.43 times the unshelled price -- about
100 FCFA if the official price prevails for unshelled peanuts. In
developing the consumer price index for peanuts, quantities of unshelled
peanuts were converted to the shelled equivalent.
Prices tend to be lowest in November after the harvest has begun
and before the official marketing campaign has opened.
Prices
frequently peak during the planting season (May-July, depending on the
rains) and then again before the harvest (September-October).
Once again, Niakhar appears to have the highest prices.
The
Central Basin (Colobane) and the Southeastern Basin (Dioly) tend to have
the lowest prices.
As Chapter 5 revealed, these are the biggest peanut-
producing zones in the sample.
Figure 9.4 presents the consumer price series for cowpeas. The
series is very incomplete, but it does illustrate that prices in the
three northernmost zones where most production takes place are at
similar levels and follow similar patterns of seasonality.
Tables A9.9-A9.14 present the producer price series based entirely
on transaction-derived prices.
Average producer prices for peanuts are
70 FCFA/kilo in all zones because such a large share of peanuts is
marketed at the official price.
Producer prices of cereals are more
volatile and also more difficult to explain without doing some
econometric analyses.
In some cases there are producer prices that are
higher than consumer prices in the same zone. This is often due to the
thinness of the data.
It can also be due to lumping different types of
sales together.
For example,
there are numerous cases of sales to
friends and neighbors.
The prices for these sales may give consumer
prices that are equal to or even lower than normal producer prices.
More time is needed to thoroughly examine the implications of the trends
in the producer price series.
Additional Work Planned
It is difficult to develop complete price series from household
survey data.
It is very important that it be done correctly because
mistakes in the data or incorrect decisions made to fill in missing
values can lead to serious problems in demand and supply analysis. The
199
fact that market survey data are available in Senegal renders the task
somewhat more tractable than it would have been otherwise, even though
observations that are missing from household data are frequently also
missing from market survey data.
On the demand side, the next step is to begin examining
relationships between purchases and sales behavior presented in Chapter
6 and the price trends presented in this chapter.
Tables A9.1-14
provide a feeling for what such an analysis will yield because the
kilograms transacted each month are presented with the price data.
Demand in rural households, however, includes purchases as well as
consumption from home production.
To conduct a complete analysis of
demand, one would want to know how consumption of home-produced cereals
as well as purchased cereals responds to price changes.
On the supply side, a great deal of work remains because of the
many gaps in the producer price series.
Ultimately, we hope to examine
the relationship between aggregate production (discussed in Chapter
5),
producer prices, and marketed surplus of cereals and peanuts.
In addition to the demand and supply analysis, it would be useful
to look at price formation patterns and market integration within and
across agroclimatic regions.
This type of analysis is not planned as a
part of the current project but the data lend itself to such analyses.
One issue that could be addressed by this type of analysis is why
consumer prices are consistently higher in the Niakhar zone of the
Central Peanut Basin even though the roads and commercial infrastructure
in the area appear to be better than that in other zones.
HOUSE-HOLD LEVEL IMPACTS OF CHANGES IN RELATIVE PRICES
Background
Over the last decade two issues concerning trade regime policy
have been hotly contested in Senegal:
(i) devaluation of the franc CFA
and (ii) increases in the tariff on rice imports. Both of these issues
have been central to the debate on Structural Adjustment policies. To
date, devaluation has not been attempted in Senegal. During the 1980s
a number of changes were made in the consumer price of rice, with
substantial increases in 1982, 1983 and 1985.
These increases were
partially rescinded in 1988 and again in 1989 following protests in
urban areas.
.
.
.,
.’
.’
200
It can be argued that trade regime changes are necessary for
efficient long-term growth, both of the food and cash crop subsectors
and of the overall economy.
Timmer (1981, p. 209), however, exhorts us
by stressing that an important part of policy analysis is "facing
squarely the short-run welfare consequences of efficient long-run food
sector development strategies".
Timmer's observation is particularly important in the Senegalese
context where data are seldom available to evaluate the potential
household-level impact of trade regime changes.
Our inability to
predict potential impacts for different groups of people makes it
extremely difficult to develop temporary policy interventions that will
protect
'at risk' groups from bearing an unanticipated share of the
negative household-level consequences
associated with such policy
interventions.
For this reason, governments tend to find themselves
initiating policy changes that must later be rescinded in the face of
popular uprisings that threaten political stability.
The objective of the analyses is to simulate the potential impact
of selected trade regime changes on different types of Senegalese
households.
The objective is not to provide arguments against reforms
in macro-economic policy that are clearly needed for long run economic
development but rather to provide policy makers with information that
can be used to ensure that macro-economic policies succeed.
Using different hypothesis about devaluation and changes in rice
tariffs, we try to better understand which types of households will
gain, which types will lose and what will be the magnitude of the gains
and loses.
The method used is partial equilibrium and comparative static. We
limit ourselves to identifying the upper bounds for potential direct
effects of policy changes on peanut, rice, and coarse grain prices.
First, we present a maintained hypothesis (not tested) of how a given
trade regime change would affect the prices households face (e.g. how a
devaluation would affect the rice price for consumers). Second, using
information on patterns of household behavior described in previous
chapters (e.g. what is the share of rice in the budgets of various
groups of consumers), we surmise how the price change would change
nominal income on the production side (how much more cash the households
income is worth), or reduce real income on the consumption side (how
much more cash households have to spend to get the same amount of
product).
2
,
a’
:
.
:
201
The results summarized here are based on analyses conducted in
November 1992 and reported in a seminar paper "Potential Welfare Impacts
of Trade Regime Changes on Households in Senegal" (Reardon et al. 1992)
presented at Saly Portugal (Seminar on Regional Integration of
Agricultural Markets in West Africa). We have not been able to revise
the analyses to incorporate very useful suggestions made by conference
participants.
Some of the revisions subsequently made to the base data
files (particularly changes in consumption levels and composition for
the southeastern Peanut Basin) have also not been included.
It is
unlikely, however, that these revisions will substantially change the
overall conclusions of the original paper.
We turn now to a brief
review of hypothesis and assumptions underlying the analysis.
This is
followed by a summary of the results and a brief discussion of the
policy implications.
Hypothesis and Assumptions
The analytical question is what is the welfare impact on rural
households of a change in a price facing the households?
In this
analysis the price change is caused by a devaluation (change in tariff
for example), but the reason for the change could be anything.
We focus on the key tradeables and non-tradeable in the food
sector.
Devaluation would:
(a)
Raise the export border price of peanuts in CFA Franc equivalent
and -- with no government intervention -- raise the farmgate
price, given a world dollar price of peanuts;
(b)
Raise the imported price of rice and of comestible oil (at border,
and with no government intervention, to the consumer), given a
world dollar price of rice and of comestible oil;
(c)
Depending on the transmission effect of a rise in the price of
tradeable rice on non-tradeable coarse grains (call it millet
here), raise the price of millet
(4
Raise the price of imported inputs such as fertilizer and
fungicides.
The impacts of (a)-(d) on household welfare are conceptually as
follows. In each case we assume commerce and manufacturing absolute
margins stay the same, for simplicity.
(a)
An increase of the peanut price raises producer nominal incomes,
and decreases consumer real incomes (to the extent that peanuts or
peanut oil is consumed);
(b)
An increase in the rice price and imported comestible oil price
decreases real incomes of consumers;
‘.
.
.
I
.
.
‘.
,’
.
202
(c)
An increase in the rice price might or might not raise the millet
price;
assuming
it raises millet prices,
but weakly (weak
substitutes, see below), this will increase producer nominal
incomes and reduce consumer real incomes.
(d) An increase in imported crop input prices decreases producer
nominal incomes.
For any "average household" in a specific zone, the combination of
(a)-(d) will produce a net effect on nominal and on real income. These
four channels are simulated, given consumption, production, and income
patterns (drawn from survey data) and based on certain assumptions
concerning the 'transmission effect' from rice price to millet price,
and the 'passthrough policy' from FOB price change to producer price
change for peanuts. These latter two are discussed below.
The above direct, instantaneous effects in practice depend on the
answers to three key empirical questions, and one policy/empirical
question, as follows.
(i)
What is the share of each of the three goods above in consumption
and in production over groups and over zones? If one assumes that
home consumed goods can be given an imputed value of the market
price, then a change in price, regardless of how much of a good is
sold or bought by a household, will change the nominal income and
real income of the household in proportion to their entire
consumption or production of the good in question. For example, if
the price of millet rises, and a third of a household's income is
composed of millet sales plus home consumption of millet, then the
price rise increases the value of the entire production of millet,
regardless of marketed share; the household could trade all the
millet for more of another good than it could previously.
The evidence for these shares for consuming/producing households
in rural Senegal has been presented in previous chapters of this
report.
ii)
What is the share of imported inputs in agricultural production,
and how does this differ by zone?
iii) What is the "transmission effect" of a rise in the price of one
consumable or producible to another? This question applies to rice
on the consumption side and imported inputs (fertilizer and crop
protection chemicals) on the production side.
The most pertinent but also the most difficult question here is
whether and to what degree a rise in the price of rice (occasioned
203
by devaluation or protection) is transmitted to the price of
millet. In theory, if the goods are strong substitutes, the price
effect on millet will be strong (if substitutes, the price of
millet will rise sharply; if complements, it will fall sharply).'
We perceive the conventional wisdom in the Sahel as being
that increases in imported rice prices would have immediate strong
impacts on coarse grain prices. Very little empirical research has
been done for the Sahel on the determinants of grain prices, and
transmission effects of macro policy changes on sectoral prices.
Consumption studies in the Sahel, including in Senegal, generally
show the cross-price elasticity of rice and coarse grains to be
either very low, insignificant, or negative and significant.
Preliminary analyses of ISRA/IFPRI survey data for rural
areas suggest that some substitution of coarse grains for rice
does occur when millet becomes less expensive relative to rice.
There is ample evidence,
however, that non-price factors (tastes
and access to commercial infrastructure,
for example) also
influence the relative shares of rice and coarse grains in the
diet.
Evidence from ISRA/IFPRI data also suggests that the
tendency to substitute in urban areas is much lower as meals based
on coarse grains are rarely eaten at noon.
In our simulations of a rice price increase in rural Senegal
we opt for a middle ground assumption that rice and coarse grains
are weak substitutes.
This implies that part of the rice price
increase will be passed on to coarse grain prices as consumers
move away from rice and increase demand for coarse grains.
In the
simulations below, we assume that 20 percent of the rice price
increase is passed on to millet.
The second aspect of the 'transmission' question concerns
inputs:
To what extent is a price increase in (imported)
fertilizer,
fungicide, and other production inputs used by
different agroclimatic zones transmitted to the prices of other
products? One can also consider indirect effects that might work
through price increases in imported fuel (transport costs, for
example). Increases in prices of imported inputs should also be
expected to be transmitted to some degree (again, depending on
degree of substitutability) to other (non-traded) inputs such as
' By incorporating a passthrough effect on mi llet prices, we dev iate
from a strictly "static" analysis, but we believe
it better reflects the
way rural households will respond.
.,;.
.
.
:
.’
204
labor.
iv)
The combined policy and empirical issue concerns how devaluation
will affect domestic peanut prices.
We assume that Senegal is a "small country" in the world peanut
market.
This implies that Senegal can market as much as it wants at the
current world market price and therefore has no interest in selling at
a lower price.
In other words, Senegal maximizes peanut revenues by
selling at the established world price rather than by lowering its price
and increasing the quantity of its sales.
Given the decline in peanut
production in recent years, and indications that worldwide demand for
vegetable oil products is on the rise,
this is a reasonable short-term
assumption.
The alternative assumption is that Senegal cannot market all of
its production at the world market price and has an interest in cutting
prices to increase the volume of exports. For example, Senegal could
undercut other sellers with a lower price in a contract for peanut
deliveries to Nigeria.
We do not use this assumption directly in the
simulations.
Returning now to the small-country assumption, let us also assume
that a marketing board does the actual exporting of peanuts. Then, a
devaluation would increase the number of FCFA that the producer or
marketing board or both could earn given the world dollar peanut price.
The equity impact would then depend on the distribution of the
gain, which we call the "passthrouoh po1ic.v". Assuming a 50 percent
devaluation,
and the marketing board passed the whole increase on to
producers, the farmgate price in FCFA would rise 50 percent. If only
half were passed on (with the marketing board keeping the difference, a
more common scenario in developing countries that export agricultural
commodities), then the producer price increase would only be 25 percent.
We ignore here what the marketing board does with its share of the
increase.
As mentioned above, this analysis is static and partial. There
are a number of important issues concerning the dynamic, indirect, and
intersectoral effects of trade regime changes that are not addressed in
these simulations.
Simulatinq Devaluation
Two points made in previous chapters of this report are
particularly pertinent to the present discussion of the potential
impacts of devaluation:
2 0 5
(1) Average levels of caloric intake are, at the margin, barely
adequate in both rural and urban areas.
This suggests that any
policy that reduces real income will be likely to put more
households in the 'at risk' category unless non-food expenditures
are reduced by the full amount of real income lost and transferred
entirely to food expenditures.
(2)
Rural household income comes from a variety of sources. The share
of crop production income in total household income ranges from 22
to 80 percent.
The smaller the share of crop revenues in total
revenue, and the smaller the share of peanut production in total
crop revenues, the less likely a rural household will be to
benefit from the positive effects that a devaluation might have on
producer prices.
Table 9.1 shows the average share of household income derived from
crop production for each rural zone.'
Part A of the table shows the
shares of different crops in total gross agricultural revenue.
Peanuts
account for 60-70 percent of agricultural revenues in four of the five
zones in the Peanut Basin.
Cereals account for most of the remaining
income,
with very small shares of income coming from cowpeas and
cotton.3
Part B of Table 9.1 shows that the share of gross crop revenue in
total household revenue ranges from a low of 20% (this was during a very
bad harvest year
in the northern Peanut Basin) to 75% in the
southeastern Peanut Basin.
In all zones but the northern Peanut Basin,
crop revenues averaged at least 50% of total income during the two year
period (1988/89-1989/90).
During the poor harvest year 50-60% of income
came from crops in most zones.
In the good harvest year, 70-80% of
income came from crop production in most zones.
Table 9 C and D show gross peanut and coarse grain incomes as
shares of total household income.
Colobane in the central Peanut Basin
and in the southeastern Basin earned approximately 50 percent of their
income from peanuts during the two year survey period.
Niakhar in the
central basin and Senegal Oriental earn the largest share from cereal
' Note that these figures differ from results in chapter 4 which
have been revised to show net rather than gross cropping income. By
doing the simulation with gross cropping income, the positive impact of
an increase in producer price of peanuts is upwardly biased.
3 Data for the 1988 cotton harvest in Missirah is not included in
this analysis; in 1989 sample households in Missirah boycotted cotton
entirely.
.
:
.
‘.
206
production, about 40 percent during the survey period.
iable 9.1 -- Incane Patterns: (Averages 1988 & 1989)
N. P. 8.
Central P. B.
S. P. B.
Sen. Or.
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
Peanuts
0.69
0.44
0.68
0.60
0.66
0.29
Cereals
0.23
0.57
0.31
0.41
0.31
0.70
Cowpeas
0.09
0.00
0.01
0.00
0.00
0.02
Cotton
0.00
0.00
0.00
0.00
0.03
0.00
N. P. 6.
Central P. B.
S. P. 8.
Sen. Or.
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
0.20
0.68
0.70
0.50
0.75
0.58
N. P. 8.
Central P. 8.
S. P. 8.
Sen. Or.
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
0.14
0.30
0.48
0.30
0.50
0.17
N. P. 8.
Central P. B.
S. P. B.
Sen. Or.
Sagatta
Niakhar
Colobane
Passy
Dioly
Mssirah
0.05
0.39
0.22
0.21
0.23
0.41
Source:
ISRA/IFPRI survey data, 1988-1990.
Table 9.2 uses the information in Table 9.1 to simulate the
potential effect of devaluation on nominal income.
We present two
assumptions about how the devaluation will affect producer price of
peanuts and one concerning a passthrough effect that increases the
millet price.
Assuming a 50% increase in the domestic producer price of peanuts
with full passthrough to the producer, the producer's nominal income
will increase from 7 to 25% depending on the zone.
The lowest increase
is in the northern Peanut Basin, the highest in the central (Colobane)
and southeastern Peanut Basin.
207
TabLe 9.2 -- Policy Effects Via Production Output Side (average 1988 and 1989)
Part A: Specific peanut and cereal effects
5D%'hikcin.domestic producer price of peanuts due to devaluation;
fixed dollar price on world market; full passthrough
Increase in nominal income due to full passthrough on peanut price
(1)
N. P. Ei.
Central P. B.
S. P . B .
Sen. O r .
Sagatta
Niakhar
c01obane
Passy
Dioly
Hissirah
Peanuts
0.07
0.15
0.24
0.15
0.25
0.08
.,
,.
,. .,.
.,
A$smnjLDw 2:
5O%"hike'in"bomestic
producer price of peanuts due to devaluation;
fixed dollar price on world market; half passthrough
Increase in nominal income due half passthrough on peanut price
(2)
N. P. B.
Central P. B.
s . P . 8.
Sen. O r .
Sagatta
Niakhar
Colobane
Passy
Dioly
Missirah
Peanuts
0.03
0.07
0.12
0.08
0.12
0.04
Arsstmprl q i‘3 ;
Devaluation raises rice price by 50% causing coarse grain price to
increase by 10% thereby increasing nominal income from cereal production
Increase in nominal income due to cereal passthrough effect (3)
N. P. B.
Central P. El.
s. P . B.
Sen. O r .
Sagatta
Niakhar
c01obane
Passy
Dioly
Wissirah
Cereals
0.00
0.04
0.02
0.02
0.02
0.04
Full passthrough plus cereal effect (4)
N. P. 6.
Central P. B.
s. P . B.
Sen. Or.
Sagatta
Niakhar
Colobane
Passy
Dioty
Missirah
0.07
0.19
0.26
0.17
0.27
0.12
Half passthrough pius cereal effect (5)
N. P. B.
Central P. B.
s. P. B.
Sen. O r .
Sagatta
Niakhar
c01otxme
Passy
Dioty
Hissirah
0.04
0.11
0.14
0.10
0.15
0.08
..;. :
.
.
.
;
208
If only half of the benefits of devaluation are passed through to
the producer, the nominal income effect is halved and ranges from 3 to
12 percent.
In hypothesizing about effects on cereal prices, we assume that
the devaluation raises imported rice prices by 50% and that there is a
20% passthrough effect on millet prices which increases the millet price
by 10%.
This price change increases nominal income by only 0 to 4
percent, depending on the zone.
Part B of Table 9.2 shows the total production effect after
combining the peanut and millet effects. With full pass through on the
peanut side, nominal income increases from 7 to 27 percent. Zones with
peanuts accounting for more than 60 percent of crop income (northern
basin, Colobane in the central basin, and the southeastern basin), a@
crop income accounting for more than 70 percent of total income
(Colobane and the southeastern basin only) realize increases greater
than 25%.
With the half passthrough assumption on peanuts -- a more likely
scenario -- increases range from 4 to 15 percent,
The total production effect in Part B is also the net production
effect after accounting
for anticipated price effects on the
agricultural
input side.
This occurs because use of tradeable
agricultural
inputs (fertilizer,
insecticide,
herbicide) is at an
extremely low level in Senegalese rainfed agriculture.
Table 9.3
illustrates that the negative effect via purchases of tradeable inputs
is insignificant given rates of input use reflected in data for the
1989/90 agricultural season.
Table 9.4 illustrates consumption patterns during the 1988-89
period, showing share of expenditure devoted to rice, coarse grains,
peanuts,
vegetable oil, and all other foods,
We concentrate on the
first four product groups as they are the ones with prices likely to be
affected by devaluation.
Across all zones, cereals as a group account for about 50 percent
of total expenditure (Part C).
About a third of total expenditure is on
coarse grains in all zones but the northern Peanut Basin (18 percent)
and the Niakhar zone of the central Peanut Basin where the Serer
maintain strong preferences for millet, spending as much as 44 percent
on coarse grains.
209
Table 9.3 -- Producticm Input Expenditure Patterns and Policy Effect.
Centrat P. B.
S. P. 8.
Sen. Or.
Sagatta
Niakhar
Colobane
Passy
Dioly
Missirah
1.
Gross ag
rev 1989
5.15
7.92
15.49
10.20
22.16
11.96
season in
millions FCFA
2.
Level of expen.
on fert. in
insig
insig
insig
0.13
0.15
insig
millions FCFA
3.
Level of expen.
on other var.
inputs (ext.
0.13
insig
0.04
insig
0.15
insig
seed and labor) in
millions FCFA
4.
Total above inputs
millions FCFA
0.13
insig
0.04
0.13
0.30
insig
5.
Share of above
inputs in gross
ag rev
0.02
nsig
insig
0.01
0.01
insig
6.
Share of above
inputs in total
income
insig
nsig
insig
0.01
0.01
insig
Effect
insig
insig
insig
insig
insig
insig
of 50%
deval. on income
via inputs
Source:
ISRA/IFPRI Survey
Note:
FCFA values are suns over all househoLds in the sample for 1989/90 production season.
Data not available for production season 1988/89.
‘.
.
210
Table 9.4 -- Expenditure Patterns: Selected Products as Shares of Total Expenditure
Part A: Share of Rice in Total Expenditure
II. P . B.
Central P. 8.
S.P.B
Sen. Or-U
Sagatta
Yiakhar
Colobane
Passy
Dioly
Missirah
1988
0.27
0.09
0.19
0.18
0.08
0.06
1989
not available
0.09
0.14
0.14
0.04
0.03
1988-89
not available
0.09
0.16
0.16
0.05
0.04
Share of Coarse Grains in Total Expenditure
Part B:
N. P. B.
Central P. B.
S.P.B
Sen. 0r.U
Sagatta
Yiakhar
c01obarle
Passy
Dioly
Missirah
1988
0.18
0.44
0.26
0.38
0.41
0.43
1989
not available
0.44
0.20
0.29
0.29
0.36
1988-89
not available
0.44
0.23
0.33
0.34
0.39
Part C* Share of Cereals in Total Expenditure (A +
A
8)
N. P. B.
Central P. B.
S.P.B
Sm. 0r.U
Sagatta
Niakhar
c01obane
Passy
Dioly
Wissirah
1988
0.45
0.53
0.45
0.56
0.49
0.49
1989
not available
0.53
0.34
0.43
0.33
0.39
1988-89
not available
0.53
0.39
0.49
0.39
0.43
Part D.
- Share of Peanuts and Vegetable Oil in Total Expenditure
Y. P. B.
Central P. 8.
S.P.B
Sea 0r.U
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
1988
0.16
0.10
0.15
0.11
0.10
0.12
1989
not available
0.12
0.10
0.12
0.05
0.11
1988-89
not available
0.11
0.13
0.12
0.07
0.11
Part Em
A Share of All Other Foods in Total Expenditure
N. P. B.
Central P. B.
S.P.B
Sen. 0r.U
Sagatta
N i akhar
cotobane
Passy
Dioly
Hissirah
1988
0.24
0.25
0.23
0.18
0.20
0.18
1989not
not available
0.22
0.32
0.17
0.26
0.21
1988-89
not available
0.23
0.27
0.18
0.23
0.20
Part F.
A Share of All Food in Total Expenditure (C + D + E)+
N. P . B .
Central P. B.
S. P. B.
Sen. 0r.U
Sagatta
Niakhar
Colobane
Passy
Dioly
Missirah
1988
0.85
0.88
0.83
0.85
0.79
0.78
1989
not available
0.87
0.76
0.73
0.65
0.71
1988-89
not availabLe
0.87
0.79
0.78
0.70
0.74
Source:
ISRA/IFPRI survey data 1988-1990
.!
211
Rice expenditure reaches a high in the northern Peanut Basin (27
percent), ranges from 9 to 16 percent in the central and southwestern
Peanut Basin, and declines as one moves eastward to the Southeastern
Peanut Basin and Senegal Oriental (4 to 5 percent).
Peanuts and oil range from 7 to 16 percent of expenditures, with
most zones being in the 11 to 13 percent range.
All foods combined {including those not specifically mentioned
above) account for 70 to 87 percent of total expenditure.
The share
allocated to food is generally greater (7 to 15 percent) following the
poor harvest of 1988.
Table 9.5 presents the hypothesized direct effects of a 50%
increase in the price of imported rice and the indirect price effect
that is passed through to coarse grains.
As mentioned earlier, we
examine these impacts assuming that the relative shares of rice an
coarse grains in the diet do not change in response to the price
changes.
While this is not an entirely unreasonable assumption in urban
areas where coarse grains are very unlikely to be substituted for rice
at the noon meal, it is a less tenable assumption for rural areas where
a variety of coarse grain dishes are consumed at all times of the day.
We see that the rice price effect results in a decline in real
income ranging from 2 percent in areas of low rice consumption (Senegal
Oriental and the southeastern Peanut Basin) to 14 percent in the
northern Peanut Basin.
Income declines by 5 to 8 percent in other rural
zones.
Although we do not present a full analysis for the urban areas
of Kaolack and Tambacounda, we expect the effect in these towns to be
comparable to that of the northern Peanut Basin or greater as the shares
of calories obtained from rice are equal or greater.
Given our hypothesis about the 20% passthrough to coarse grain
prices, the drop in real income increases another 2 to 4 percent across
all zones.
The size of the passthrough effect in the short-run depends
on how price elastic demand for rice and coarse grains are.
If small
increases in rice price push consumers to reduce rice consumption and
increase coarse grain consumption substantially the price of the latter
good will rise rapidly.
The medium-run size of the passthrough effect
depends to a large extent on the supply responsiveness to increases in
millet prices.
Assuming that the devaluation increases the producer
price of peanuts more than it increases the producer price of coarse
grains,
there is unlikely to be a large coarse grain supply response
because peanuts will continue to be a more profitable crop.
.’
.
212
Table 9.5 -- Policy Effects via Decreases in Real Income on Cmsurption Sick
.,.
Paf-jz:~~:.I$ff&%s”id&, :~~~:::~i~Fcj:;product~:by'~~~r::~~::'::~~
.,... . . . .,.,.,
,. . . ,.
.
AS,+HPTIdi( 3:
Rice price increases by 50%
Decrease in real income due to change in rice price (I)
N. P. B.
Central P. 8.
S. P. 8.
Set-~. 0r.U
Sagatta
Niakhar
c01obane
Passy
Dioly
Hissirah
1988
0.14
0.05
0.09
0.09
0.04
0.03
1989
not available
0.04
0.07
0.07
0.02
0.02
1988-89 not available
0.05
0.08
0.08
0.03
0.02
ASW.TXDN..Zi
..,. . . . . . . . . ..A. :;....i. ..A.., . . . . .
Coarse grain price up by 10% c&z to passthrough frm rice (2)
Decrease in real income due passthrough effect on coarse grains
N. P. B.
Central P. B.
S. P. B.
Sea 0r.U
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
1988
0.02
0.04
0.03
0.04
0.04
0.04
1989
not available
0.04
0.02
0.03
0.03
0.04
1988-89 not available
0.04
0.02
0.03
0.03
0.04
@,.$k&$J~~~~j3 Price of peanuts and peanut oil increase by 50% (3)
Decrease in real income due to peanut and oil price effect
N. P. 8.
Central P. 8.
S. P . B .
Set-~. 0r.U
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
1988
0.08
0.05
0.08
0.06
0.05
0.06
1989
not
available
0.06
0.05
0.06
0.03
0.06
1988-89 not available
0.06
0.07
0.06
0.04
0.06
Part'?: ~fotal:‘d~rea&in reat’:iya+ ,,F”average durinq:S988;1989 period,(&)
. . . ::. .“. ,.
N. P. B.
Central P. 8.
s. P. 0.
Sen. 0r.U
Sagatta
Niakhar
Colobane
Passy
Dioly
Hissirah
Rice
0.14
0.05
0.08
0.08
0.03
0.02
MMS
0.02
0.04
0.02
0.03
0.03
0.04
PN/Oi1
0.08
0.06
0.07
0.06
0.04
0.06
Total
0.24
0.14
0.17
0.17
0.09
0.11
.
‘
.
.
213
Although peanuts serve as the major cash crop throughout most of the
rainfed areas of Senegal, the crop is also an important food crop.
Peanuts are consumed directly as a condiment in sauces or a snack food
and indirectly as processed oil.
Vegetable oil consumed is of three
types: home-processed, locally produced peanuts, industrially processed
peanuts,
industrially processed oils from imported oil seed products.
The price of all three types of vegetable oil will be influenced by a
devaluation.
As rice consumption has increased over time, there has been a
tendency for oil consumption to also increase (large quantities of oil
are used to prepare the most popular Senegalese rice dish 'tcheb u
jen').
The ISRA/IFPRI data suggest that there is a secular trend toward
increased rice and oil consumption as rural areas tend to be adopting
more urban consumption patterns.
In the current analysis, we hypothesize that the price of all peanut
and oil products
increase by 50 percent due to the devaluation
(assumption 3, Table 9.5).
This results in a 6 to 7 percent decrease in
real income in most rural zones.
Part B of Table 9.5 summarizes and adds up the three consumption-side
policy effects described above.
Using the average effect across the two
years
of survey data we find that real income declines by 9
(southeastern Peanut Basin) to 24 percent (northern Peanut Basin). The
effect is strongest in the north, west, and center of the Peanut Basin
and diminishes as one moves east to the southeastern basin and to
Senegal Oriental -- the two zones with lowest rice and oil consumption.
Net real income effects followinq a hypothesized devaluation
Table 9.6 presents the final picture of the overall impact on rural
households.
Using our best case scenario (full passthrough for peanuts
and 20% passthrough from rice to millet prices) the net effect on real
household income ranges from a 17% decline (northern low production and
high rice consuming zone) to an 18 % increase (southern high peanut
producing and low rice consuming zone).
Two of the relatively high
potential zones (the southwestern basin and Senegal Oriental) are
virtually unaffected by the change,
In the central Peanut Basin, the
Colobane zone, which produces more peanuts and consumes more rice than
the nearby zone of Niakhar,
fairs much better (9% increase in Colobane
verses a 4% increase in Niakhar).
‘.
~
:
:
:
‘.
214
214
Table 9.6:
Calculation of Net Effects on Real Household Inccm 1988-89
N. P. B.
C e n t r a l P . 6.
S . P . 8.
Sen. Or.
S a g a t t a
Niakhar
Colobane
Passy
D i o l y
Hissirah
Total
Negative
-0.24
-0.14
-0.17
-0.17
-0.09
-0.11
Consumption
Effect (1)
Total
Positive
Production
Effect
0 . 0 7
0.19
0.26
0.17
0.27
0.12
Full
Passthrough (2)
Net
Effect
Ful I
- 0 . 1 7
0.04
0.09
-0.00
0.18
0 . 0 1
Passthrough (3)
Total
Positive
Production
Effect
0.04
0.11
0.14
0.10
0.15
0.08
Half
Passthrough (4)
Net
Effect
H a l f
- 0 . 2 0
-0.03
-0.03
-0-08
0.05
-0.03
Passthrough (5)
,
:
_,;
:
215
The last line of Table 9.6, using the half passthrough option,
presents a more realistic picture of what might really happen following
a devaluation.
All zones but the southeastern Peanut Basin experience
a loss in real income.
The increase in income is 5%. The decline in
income ranges from an extreme of 20% in the north, to the more typical
case of 3% declines in the central basin and Senegal Oriental.
The
southwestern basin suffers slightly more with an 8% decline in
purchasing power.
This analysis suggests that the ranking of zones by
income prior to a devaluation will no longer hold after devaluation.
The northern Peanut Basin will clearly move from one of the wealthier
zones to one of the poorest.
The southwestern Peanut Basin may also
drop from being one of the top three income earners. The southeastern
Peanut Basin, the wealthiest zone prior to devaluation, will remain the
wealthiest after devaluation.
Although we do not do a separate simulation for the urban zones,
their consumption patterns are similar to those in the northern Peanut
Basin.
Given that urban households do not realize the gains associated
with peanut production, we would anticipate the net effect in urban
areas to be somewhat more negative than that for the northern Peanut
Basin -- at least a 25 percent drop in real income.
To repeat, the key factors in determining whether one gains or losses
following this hypothesized devaluation are (1) the importance of
peanuts on the production side (those with a greater share of peanut
revenues gain proportionately more) and (2) the importance of rice on
the consumption side (the more rice consumption the more one's real
income declines).
Given the descriptive information on consumption and income patterns
presented in the rest of this report, one can draw some tentative
conclusions about types of households and individuals most likely to
gain or lose from a devaluation.
Rural households in the poorest tercile are more likely to produce a
larger share of cereals than of peanuts because they cannot afford
peanut seed (poor access to credit) and they need to devote a large
share of land to cereal crops to assure a minimum level of food
security.
This means that poorer households will be less likely to
benefit from the positive production side effects associated with peanut
production.
Rural households in the poorest tercile are net cereal purchasers.
Of cereals purchased, imported rice accounts for 40-60 percent.
In the
northern, central, and southwest Peanut Basin rice accounts for anywhere
from 9 to 39 percent of total cereal/pulse calories, depending on the
,i/.
:
,
.!
216
survey year.
Unless these households are able to increase coarse grain
production or purchase coarse grains at predevaluation prices, they will
experience substantial loss in purchasing power after devaluation.
Given that poor households were unable to significantly increase crop
production during the favorable 1989 season and given that they are not
big peanut producers, the net effect is likely to be quite negative for
the poor tercile in these zones.
The fact that these are among the most
densely populated regions in Senegal magnifies the problem.
Women do not generally earn a great deal of income from peanut and
cereal production.
When women do have their own fields, however, they
usually produce a cash crop such as peanuts or cowpeas rather than
cereals (this pattern has been less pronounced since the government
stopped their seed distribution program in the mid-1980s). As women are
generally not responsible for purchasing family cereal supplies with
their own income, they stand to realize a net gain from their peanut
production.
This same scenario would be true for other household
members who grow peanuts and have little or no responsibility for cereal
supply ("sourga," unmarried adult males, dependent household heads).
On the other hand, the household head on whose fields the family
cereal supply is grown, will gain less than other household members to
the extent that he specializes in cereal production rather than peanuts.
Most household heads grow both cereal and cash crops.
One scenario
might be for the household head to increase peanut production in
response to the rise in nominal peanut prices and cut back on millet
production believing that the increased revenue will be adequate to
cover the cereal shortfall with purchases.
If enough producers do this,
coarse grain supplies are likely to decrease putting further pressure on
cereal prices.
The advocates of devaluation have the high ground in that it is clear
that there are very severe fiscal and balance of payment deficits and
currency overvaluation in the Sahel.
On the other hand, evidence presented in this chapter suggests that
only one rural zone survives a half-passthrough devaluation with a net
gain for the average household.
It is clear that urban households on
average risk a sharp drop (20-25 percent) in real income as most urban
households are in the service sector or local commerce and not likely to
realize any compensatory production side benefits.
We have pointed out that one of the benefits of examining potential
effects of a devaluation on households is that it permits one to design
policy interventions that lessen the negative impacts for households
’
5
.
.’
,
.
:
217
likely to experience large loses in real income and to stimulate the
positive impacts from the production side.
Three broad categories of interventions come to mind:
Investments in reversing structural constraints to agricultural
supply response, especially in the high potential zones.
Second is investment in ways to make coarse grain processing
cheaper and more accessible so that local grains can gradually replace
rice as the grain that meets the needs of the poor in growing cities.
Third is investments in infrastructure and transport, and
certainly protection from spending cuts in this domain, so as not to
allow devaluation and public spending cuts to increase price instability
and transaction costs, thus choking off supply response.
It is evident that in the very short-run, supply response will not be
adequate to provide consumers with large stocks of low cost coarse
grains that will be able to substitute for more expensive rice. It is
here that better organized intra-regional markets could help if some of
the excess production in neighboring countries (perhaps Mali or Niger)
could be made available in Senegal.
One might also consider some type
of targeted program to assure that poor households in the zones most
likely to realize strong negative impacts (urban areas and the northern
Peanut Basin) do not have to support the full impact of the devaluation.
Our data suggest that poor households in the north, for example, rely
most heavily on crop production and livestock for their incomes. We
also found that poor households ate a much larger share of cowpeas than
wealthy households following the poor 1988 harvest. These facts suggest
that programs to increase cowpea production and marketing systems in the
northern Peanut Basin could be a significant help to poor farmers.
Cowpea production could improve caloric intake, provide a reliable cash
crop, and also provide high quality feed supplements for livestock in
the form of cowpea hay.
IFPRI/ISRA data show that livestock production is frequently the
second most important source of income for poor households in rural
areas.
Policy interventions
to make livestock enterprises more
profitable for poor households could also protect them from the negative
effects of devaluation -- credit for fattening activities and better
market information systems are two interventions that come to mind.
Timely investments in infrastructure could also improve production and
marketing response.
At the beginning of the ISRA/IFPRI survey one of
our study zones was just 20 minutes from a paved road; by the end of the
218
survey the road had deteriorated so much that it took about an hour to
get to the study zone.
Water is another problem. Much time that could
be devoted to crop production is now allocated to transporting water due
to inadequate well systems.
The above is clearly not an exhaustive list of possible policy
interventions but merely suggestive of the types of measures that could
be undertaken to protect 'at risk' groups from the negative consequences
of a devaluation and encourage producers to realize the full benefits of
better nominal prices for some of their products.
,..
.
.
I .
..
;
I
219
REFERENCES
Fall, A.A., V. Kelly, and T. Reardon. 1989. Household Survey Methods
used in the IFPRI/ISRA study of consumption and supply impacts of
aqricultural price policies in Seneqal.
Document 2, IFPRI,
Washington, D.C.
Timmer, C.P. 1981.
"Developing a Food Strategy," in Proceedinqs of the
Conference on Food Security in a Hunqrv World, International Food
Policy Conference, San Francisco.
.’
‘.
:
.
I
’
220
SUMMARY OF PRINCIPAL RESULTS AND POLICY IMPLICATIONS
This report has served as a vehicle for presenting the most salient
characteristics of income (levels, composition, and sources), crop
production, agricultural product and livestock transactions (purchases,
sales, gifts, and prices), and food consumption for rural and urban
households in the Peanut Basin and Senegal Oriental during the period
October 1988 through July 1991.
The data were collected to permit
econometric modeling of household demand and supply for agricultural
products, particularly cereals and coarse grains. Although the modeling
work is still under way, there are a number of important findings that
are unlikely to change during the course of future analysis.
These
findings are summarized below.
A discussion of the insights of these
findings for agricultural price policies follows.
The next section
describes patterns of income diversification and their implications for
rural development policies.
A discussion of the policy implications can
be found at the end of the summary.
PRINCIPAL RESULTS
Food consumption
Adequacy:
Average daily caloric intake per adult equivalent
(AE)' for each zone was generally below recommended levels (3000
kcal) but close to minimum needs of 2400 kca1.l Exceptions were
' Adult equivalents are calculated by weighing different persons
in the household by a coefficient representing their relative demands
in consumption.
For example, an adult male will have a much larger
weight than a 5 year old boy.
FA0 adult equivalent standards were
used.
Many results are standardized and reported in adult equivalent
terms (income per AE, for example) to facilitate comparisons across
households of different sizes.
' Minimum daily needs of 2400 kcal/AE are equivalent to 80
percent of the 3000 kcal/AE norms recommended by the Organisme de
Recherches sur 1'Alimentation et la Nutrition Africaine (ORANA). When ,
the average intake per AE falls below 2400 kcal/day, the group is
considered to be "at risk".
IFPRI/ISRA estimates of caloric intake
include all cereals, pulses, tubers, milk, vegetable oil and bread.
encountered in the central Peanut Basin zone of Niakhar (2200 kcal),
in Senegal Oriental (2100 kcal), and in the town of Kaolack (2100
kcal ) .
Given that products not included in the analysis (see footnote 2)
were consumed in very small quantities in Niakhar and Senegal
Oriental, it is unlikely that average intake for these zones would
have reached 2400 kcal if all products were considered. Diets being
much more diversified in urban areas means that average consumption in
Kaolack may have approached, but probably not surpassed, the 2400 kcal
level if all products had been considered.
Although these zone averages serve as a vehicle for making
interzone and interannual comparisons, they tend to mask important
differences among households within the same zone.
About half of all rural households in the sample fell into the
"at risk" category consuming less than their minimum needs of 2400
kcal in both years (59 percent in 1988/89 and 47 percent in 1989/90).
Approximately 25 percent of the entire sample averaged less than 2000
kcal/AE.
The southeastern Peanut Basin had the most households falling
below 2400 kcal in both years (71 percent in 1988/89 and 60 percent in
1989/90), while the southwestern Peanut Basin had the fewest in
1988/89 (44 percent) and the Colobane zone of the central Peanut Basin
the fewest in 1989/90 (32 percent).
In the urban zones, half of the
households interviewed had average annual caloric levels lower than
2400 kcal/day/AE, illustrating that undernutrition is both a rural and
an urban phenomenon.
Composition:
The important role of imported rice has long been
acknowledged for Dakar but generally considered of little importance
outside the capital city.
The IFPRI/ISRA data now confirm that
imported rice provides more than half of all household consumption of
cereal calories in the secondary towns of Kaolack (62 percent) and
Tambacounda (70 percent), as well as in rural areas of the northern
Peanut Basin (70 percent).
In the other zones sampled in the central
and southwestern Peanut Basin, rice provides from 10 to 33 percent of
cereal calories, with higher consumption following poorer harvests.
Rice provides only 5-9 percent of cereal calories in Senegal Oriental
and the southeastern Peanut Basin.
Millet and sorghum account for 99 percent of coarse grains
consumed by sample households throughout the Peanut Basin.
Maize
consumption is important only in Senegal Oriental (the principal maize
production zone), where it provides about 50 percent of cereal
calories in rural areas and 12 percent in the town of Tambacounda.
Prices for imported rice compared to coarse grains: For coarse
grains to provide the same amount of consumable calories per CFA franc
as imported rice, the retail price of whole grain millet, sorghum, and
Products not included are sugar, fruits, vegetables, meat and fish.
222
maize would need to be below 92 FCFA/kilo.3 There are instances in
the IFPRI/ISRA data when the coarse grain price exceeded this
threshold
-- primarily in the northern Peanut Basin and the Niakhar
zone of the central basin from April through September 1989. Across
all rural zones and years, however, coarse grains generally provided
less expensive calories at prevailing grices than imported rice, even
when processing costs are factored in.
The large share of imported rice consumed in the northern basin
(70 percent of cereal calories) can be explained to some extent by
relative prices of rice and millet.
Price, however, does not appear
to explain the large share (about 30 percent) of rice consumed in the
central Peanut Basin (Colobane) and the southwestern Peanut Basin, nor
the small share of rice (15 percent) consumed in the central Peanut
Basin (Niakhar) which exhibited millet prices higher than those in the
northern basin where substantial amounts of rice were eaten.
This
behavior suggests that factors other than price (dietary preferences
and access to markets where rice is readily available, for example)
play an important role in decisions about cereal consumption made by
households in most rural zones.
In urban areas, the relative importance of price and nonprice
factors is less clear.
During the survey year average prices in both
zones exceed the threshold price (which underestimates somewhat the
urban processing costs).
Average annual millet prices in Kaolack were
115 FCFA/kilo.
Average annual sorghum and maize prices in Tambacounda
were about 96 FCFA/kilo.
Given that rice was generally less expensive
than coarse grains during the entire urban survey, it is not
surprising to find 60 to 70 percent of cereal calories coming from
rice.
This makes it difficult, however, to evaluate the extent to
which an increase in rice price might increase coarse grain
consumption.
Rice consumption per AE in market villages is consistently higher
(1.5 to 8 times greater, depending on the zone) than that in nonmarket
villages.
Consumption patterns in market villages serve as an
indication of the secular evolution in rural consumption patterns
given increasing levels of commercialization. The survey provides
little grounds for believing that the evolution of rural consumption
patterns over time will reverse the current trend toward greater rice
consumption.
3 The 92 FCFA/kilo price factors in the FCFA/kilo cost of
processing manually millet at a village millet mill. If processing
costs are not considered, the threshold price would be about 107
FCFA/kilo.
4 Imported rice consumed by sample households was almost
exclusively Thai brokens.
Only one instance was enumerated of higher
quality rice being eaten (an urban household in Tambacounda).
:
.
_‘,’
:
,I,
.
./
.~
223
Cereal purchases:
Rural households, in general, purchase 20 to
40 percent of all cereals consumed.
Although the share of purchases
in total cereal consumption is lower following the much better 1989
harvest, the drop in all but one zone is only 2-3 percent. This
suggests that household cereal purchasing behavior is heavily
influenced by factors other than household cereal harvests.
Higher
income levels in 1989 (see income section below) might have encouraged
the higher levels of consumption of purchased rice in some households.
Preferences for diversified diets or easy access to markets (which
increases opportunities for off-farm income and reduces transactions
costs) are other factors that may keep purchases high even when
production is adequate.
Policy implications of food consumption patterns: The fact that
a large share of households do not cover minimum caloric needs implies
that some type of policy adjustment is necessary. Because more
households are at nutritional risk in higher potential agricultural
zones, nutritional inadequacy may be linked to problems of income
smoothing rather than total income per se (see discussion on income
diversification below).
Credit or food for work programs should be
considered in such situations.
Imported rice is an important component of the diet in most rural
households.
Any policy that increases relative rice prices has the
potential to negatively affect real incomes of rice consuming
households.
Crop production and marketinq
Production technology:
Farm size varies significantly across
zones with the smallest farms (4-5 hectares) found in the zones with
the lowest cropping incomes per AE (northern basin, Niakhar in the
central basin, and Senegal Oriental).
The correlation between small
size and low cropping income per AE suggests that the smaller farm
sizes are not being compensated for by more fertile soils or higher
yields as is sometimes the case in higher potential zones that have
small farms.
Results from a pre-survey village reconnaissance suggest that
small 'size in the central basin zone of Niakhar is due to a land
constraint, while small size in Senegal Oriental is related to labor
and equipment constraints.
Median farm size in other zones ranges
from 7 hectares in the southwest to 11 in Colobane and 13 in the
southeastern Peanut Basin.
Use of "modern inputs" such as fertilizer, pesticides, and
herbicides is extremely limited,
being found primarily 'in households
with production contracts for cotton or confectionery peanuts.
Fewer
than 10 percent of sample households benefitted from these contracts
in any given year.
Most participating households contracted for a ,
hectare or less.
224
Use of fertilizer has declined dramatically since the price
increased in the mid-1980s, following the reduction and then the
1
elimination of fertilizer subsidies.
Credit constraints also
discourage farmers from using fertilizer and from investing in new
traction equipment.
Peanut seeds account for the largest single input cost;
approximately half the peanut seed comes from household stocks and
half is purchased from neighbors or at weekly markets. Despite the
relatively low levels of modern inputs used, input costs represent 14
to 21 percent of the gross value of all cropping output; this is due
in large part to the seed intensive nature of peanut production.5
"Informal" markets based on labor exchange among neighboring
households with inking payments or reciprocity provide most nonfamily
cropping labor, particularly at harvest time.
"Formal" markets for
day laborers do not appear to be well developed. This observation is
based on the fact that hired labor is seldom used and wages paid for
similar tasks are extremely variable.
Production levels:
Zone averages for cereal sufficiency ratios
(ranging from -14 to 1.03) show that households could have covered
from 2 to 12 months of coarse grain "needs" with home production,
depending on year and zone.6
All zones failed to cover minimum
coarse grain "needs" from their own production in 1988/89. Following
the better than average 1989/90 harvest, three zones (Colobane in the
central basin and both zones in the southern basin) had average
sufficiency levels of 12 months while the remaining zones covered 3 to
9 months.
Producer prices of cereals and peanuts: The government
controlled producer price for peanuts was found to be generally
effective during the official marketing period (December - March) when
most peanuts were sold.
The price during the survey period was highly
subsidized (at least 25 percent) given the prevailing world price for
5 Peanut cultivation is particularly demanding in seed. Farmers
use at least 100 kilos of unshelled peanuts to seed one hectare; a
typical yield is 800 kilos (reproduction rate of 8). For cereals only
4 kilos of millet seed are needed to harvest about 400 kilos
(reproduction rate of 100).
6 The cereal sufficiency ratio is calculated by dividing
calories available per day from home production by a household's
average daily coarse grain needs, estimated from IFPRI/ISRA
consumption data to be 1900 kcal/ae.
Use of 1900 kcal makes allowance
for the fact that all households obtain some calories from imported
rice, pulses and other foods.
225
Senegal's peanut oil exports and the net returns realized after
processing.7
Average annual producer prices for coarse grains were in the 60
to 70 FCFA/kilo range for most zones during the two year period, with
prices at the lower end of the range following the better 1989/90
harvest.
A direct comparison of producer prices is not, however,
informative of the economic incentives carried by those prices unless
it is put in the context of returns to land and/or labor.
Returns to family labor:*
Median returns to an 8-hour workday
in household crop production ranged from 150 to 1500 FCFA/day,
depending on crop and zone, during the relatively good 1989/90 season.
The higher end of these returns compares favorably to the SMIG
(official minimum wage in the formal sector) of 1500,FCFA/day.
Analysis of returns to family labor shows that there is little
incentive for most farmers to increase cereal production given current
technology and price relationships. Cereal production was
significantly more remunerative than peanut production (almost double)
in only one zone -- Niakhar in the central Peanut Basin.' In all
other zones median returns to labor in peanut cultivation were higher
(12 to 500 percent) than those for cereals. In three .zones (the
northern Peanut Basin, Niakhar in the central Peanut Basin, and the
7 The 25 percent figure is the estimated subsidy given the
current overvalued CFA franc.
Work by Badiane and Kinteh (June 1993)
suggests that the subsidy serves in large part to compensate for
farmers' loss due to overvaluation.
* Returns to family labor are calculated for each field as net
returns divided by the number of hours of household labor used on the
field.
Labor inputs are not weighted by AE (AE is an appropriate way
of weighting consumption data but is not indicated for production
analysis).
Net returns are calculated by subtracting costs (imputed
or actual) of all variable inputs from the gross value of crop
production (valued at the average producer price during the harvest
year for each crop/zone combination).
', Numerous explanations for the higher returns to millet in this
zone can be hypothesized:
climatic and pest patterns during the study
period that were more favorable to millet than to peanuts, poor
quality peanut seed, or farming techniques that favor millet over
peanuts.
Further analysis of the IFPRI/ISRA crop production data and
additianal field work are necessary to fully understand why this zone
appears to be different.
I
‘.
:
.
3 .
226
southwestern basin) the difference in mean returns to labor for the
two crops was statistically significant at the .05 level."
Crop marketing:
Gross receipts from peanut sales account for 85
percent or more of receipts from crop sales in each year and zone
except Senegal Oriental where cotton was the most important crop in
1988/89 (86 percent of gross receipts). Zones with the largest share
of sales receipts from cereals were Colobane in the central Peanut
Basin (8 and 12 percent of sales in 1988/89 and 1989/90 respectively),
Dioly in the southeastern Peanut Basin (7 percent both years), and
Passy in the southwestern basin (2 and 6 percent)." These sales were
virtually 100 percent millet.
Although some households in Senegal Oriental had a cereal surplus
in 1988/89, they sold less cereal than deficit households in the
Peanut Basin.
This appears to be due to the zone's isolation from
cereal deficit zones and differences in consumption habits (maize and
sorghum are seldom eaten in the Peanut Basin) that limit extra-zone
demand for Senegal Oriental's maize and sorghum.
Despite the generally low cereal sufficiency ratios, more than 90
percent of all households sold some cereal each year. Given that
approximately 50 percent of sample households fall below the "at risk"
level of caloric intake, it is clear that food insecure households do
sell cereals.
Overall quantities sold per household, however, are usually
small.
More than half of the sample households sold less than 200
kilos per year.
This means that half of sample households are not
even marketing enough cereal to satisfy typical annual demand by one
adult equivalent."
To better understand which types of households are marketing
large quantities of coarse grains analysis of households selling more
lo In the Colobane zone of the central basin, the difference was
significant only at the .15 level.
" According to Commissariat de Securite Alimentaire (CSA)
records, the zone of Colobane had a net cereal deficit during the
entire two-year survey period.
ISRA and CSA personnel are currently
collaborating to determine the source of these differences.
Given the
large share of purchased rice in this zone's cereal consumption, it is
possible that farmers are selling millet and purchasing rice.
'* Coarse grain demand per AE is estimated at 1900 kcal per day
(from consumption data).
Allowing for loss of calories in
transforming whole grain to its consumable equivalent, one kilo of
coarse grain provides about 2800 calories:
(200 kilos * 2800
kca1)/365 days = 1534 kcal per day.
227
than 200 kilos of cereal was conducted. These households are more
likely than others to:
be located in the Peanut Basin
be located in villages without weekly markets
belong to the Wolof or Fulani ethnic group
belong to the highest income tercile
earn more than 40 percent of total income in cropping
activities
have a cereal sufficiency ratio greater than .75
cultivate more than 6 hectares per household and more than 1
hectare per adult equivalent
consume less than 10 percent of their total calories from
imported rice.
The most unanticipated result here is that households in market
villages are not selling more quantities of cereals than those in
villages without markets.
This can be explained by a combination of two factors:
1)
Households living in market villages have more diversified
opportunities and rely on non-agricultural income rather than
cereal sales to supplement cash crop income.
This observation is
supported by the fact that farm size and the share of income from
crops both tend to be lower in market than in non-market
villages.
2)
The relatively dense market infrastructure in the Peanut Basin
provides most non-market village households with easy access to a
weekly market where they can market crops.
This ease of access
is confirmed by the fact that most households had some members
making purchases at nearby weekly markets every week.
As market
infrastructure is much thinner in Senegal Oriental it may be
constraining cereal sales in that zone.
Policy implications of crop production and marketing behavior:
Farmers' production and marketing patterns during the 1988-1990 survey
period suggest that GOS cereal production goals (80 percent self
sufficiency by the year 2000) are unlikely to be met if current price
relationships and marketing policies continue.
The small-scale,
largely non-commercial nature of smallholder participation in coarse
grain supply (unlike the case for peanuts) suggests that attempts to
encourage greater coarse grain consumption by increasing rice prices
may require prohibitively large relative price changes before major
changes are observed in coarse grain supply.
Low levels of cereal
production and marketing suggest that a guaranteed outlet and price
for peanuts and cotton may be diminishing the relative incentive for
cereals, which, unlike the two export crops, have no guaranteed price.
and purchaser.
Thin commercial infrastructure in Senegal Oriental
also appears to be constraining cereal production and marketing in
.’
..
.’
228
that zone.
The relative importance of infrastructure versus other
factors (demand for maize or production technologies, for example)
requires further investigation,
however.
Household Income: Levels, distribution, and sources
The level of income is traditionally considered one of the most
important determinants of a household's consumption behavior.
It can
also influence production behavior, particularly a household's ability
to make capital investments or to purchase costly inputs when credit
markets do not function well.
The share of income earned in nonfarm activities can also
influence consumption and production behavior, yet rural household
studies rarely address the issue of multiple income sources and what
this implies for household response to policy changes. Without the
intensive survey methods used in the IFPRI/ISRA study, there is a
considerable risk of deriving an erroneous impression of the relative
importance of different economic activities in household income
strategies.
For example, assuming equal levels of income, a Senegalese
household earning 30 percent of income from cropping and purchasing 80
percent of cereal needs will have a different reaction to a rice price
increase than a household that earns 80 percent of its income from
cropping and only purchases 10 percent of its cereals.
These two
households could also be expected to respond differently to programs
designed to improve agricultural productivity by decreasing soil
erosion if such programs competed with nonfarm activities for
household labor.
Levels and sources of incomes in rural Senegal differ across
zones due to differences in the natural resource endowment (climate,
soils), the infrastructure (transportation, communication), and the
institutional environment (savings, credit, markets).
Understanding
how both levels and sources differ across zones is an essential first
step to understanding how households are likely to react to various
policy initiatives.
In the IFPRI/ISRA study, "income" means net household income for
one harvest year.
The complete flow of income for all household
members and from all sources is taken into account; this is only
doable in the case of bi-weekly interviewing as was the case for the
IFPRI/ISRA study.
Cropping income is calculated by valuing total
harvest at the annual producer price for the zone. The costs of
variable inputs for all activities are netted out as is depreciation
when relevant.
The household income is standardized and reported as
income per AE to facilitate interhousehold comparisons.
Levels:
Average household income in rural zones ranges from
27000 FCFA ($95) per AE in the central Peanut Basin zone of Niakhar i'n
1988/89 to 72000 FCFA ($250) per AE in the southeastern Peanut Basin
229
during harvest year 1989/90.
Average income for urban households is
higher -- about 105,000 FCFA per AE ($365) in both Kaolack and
Tambacounda in 1990/91.13
Following the poor 1988/89 harvest, income did not cover minimum
needs14 for 50 percent of households in the Niakhar zone of the
central basin.
In the Colobane zone of the central basin, the
southwestern basin and Senegal Oriental, 20 to 30 percent of
households had incomes less than minimum needs.
Following generally better 1989/90 harvests, Niakhar (central
Peanut Basin) and Senegal Oriental still had a substantial share of
households with combined income from all sources less than "minimum
needs" levels (35 and 21 percent, respectively).
Interannual differences in levels of rural income are large.
Following the better 1989/90 harvest, average household incomes (from
all sources) increased by 20 to 70 percent throughout the Peanut
Basin.
The central Peanut Basin zones had both the lowest (Niakhar)
and the highest (Colobane) increases.
Most of the increase came from
better cropping outcomes which increased by 50 to 100 percent between
the two years, but some also came from higher off-farm incomes.
Only Senegal Oriental failed to experience an important increase
in income in 1989/90.
This was due in part to poor rainfall but also
because farmers in the zone boycotted cotton production the second
year.
Distribution across households: The distribution of income
across households within each zone is more skewed than that found in
other Sahelian countries for which comparable data are available.
Gini coefficients measure the extent to which income is concentrated
in a small share of households; the larger the coefficient, the
greater the concentration of income.
Gini coefficients calculated
for each rural zone and year are most often in the .35 to .45 range.
Estimates for Nigeria, Niger and Burkina Faso are -28, .30 and .33
(Matlon, 1979; Reardon,
1989; Hopkins and Reardon 1993). The urban
distribution is more skewed than the rural with Gini coefficients in
the .5 range.
Sources of income:
In urban areas, the commerce and services
sectors dominate.
Commerce accounts for 46 percent of net household
13
The per capita equivalents of the Niakhar and southeastern
Peanut Basin figures cited above are 20,000 and 52,000 FCFA,
respectively ($70 and $180); the urban per capita figure is 82,000
FCFA ($280).
During the survey period US$l.OO was worth approximately
290 FCFA.
l4 Minimum needs is defined as the cost (using zone-specific
1
prices) of 1900 kcal/AE/day of coarse grains plus 20 percent to cover
other essential food and nonfood expenditures.
.
~
,’
.
.’
.
:
230
income in Kaolack and services for 22 percent.
In Tambacounda, the
relative importance of the two sectors is reversed -- the service
sector (which includes civil servants) provides 75 percent while
commerce contributes only 12 percent. The contribution by sector is
fairly constant across income terciles.
In rural areas, cropping activities contributed from 20 to 70
percent of rural household income in 1988/89 and 40 to 80 percent in
3989/90.
The extreme case of only 20 percent cropping income was
found following the locust invasion of the northern Peanut Basin in
1988/89.
Consistently large market shares of cropping income (greater
than 70 percent) were found only in the southeastern Peanut Basin.
Most zones were in the 40 to 50 percent range in 1988/89 and 70 to 80
percent range in 1989/90.
Interannual differences in levels of cropping income are quite
large.
In the central Peanut Basin, levels of cropping income doubled
between 1988/89 and 1989/90.
In the southern Peanut Basin zones, the
level increased by 50 percent.
In Senegal Oriental the level of
cropping income dropped between the two years by about 20 percent (due
largely to the cotton boycott and poor rains in 1989/90).
Net income from livestock transactions was an important share of
total income (13 to 23 percent) in all zones but Niakhar in the
central basin and the southwestern basin (5 percent) following poor
1988/89 harvests.15
The following year, livestock income accounted
for less than 5 percent of income except in the southeastern basin and
Senegal Oriental where it represented about 20 percent. The tendency
for livestock income to vary inversely with cropping income and,
therefore, smooth total income is clearly evident.
Off farm income has a variety of sources; commerce and service
activities account for the largest share in most zones while
handicrafts (northern Peanut Basin) and migratory remittances
(northern basin and Senegal Oriental) tend to be more zone specific.
Findings concerning differences in income diversification patterns
across zones and the policy implications of these differences are
presented following the discussion of price policies.
Insights of consumption, production and income results for
aqricultural price policies
Results from simulation analysis show economy-wide policy reforms
(nominal devaluation, for example) have the potential, in the short-
run, to substantially reduce current real incomes of households in
l5 The extreme reluctance of the Serere households in the
Niakhar zone of the central Peanut Basin to sell their livestock is
evident in the IFPRI/ISRA data.
This zone has the largest livestock
herds, the lowest income levels, and the lowest caloric intake in
1988/89 yet they sold very little livestock to supplement income.
:
.’
;
231
urban areas and those in rural areas (northern Peanut Basin, for
example) that obtain a large share of cereal calories from imported
rice.
The same type of policy reform also has the potential to
increase real income for households that consume little rice and
produce large quantities of peanuts (southeastern Peanut Basin, for
example).
These changes may be quite different once associated changes
occur in factor markets such as labor and capital, most particularly
if accompanied by a liberalization of the rice market that would tend
to reduce any devaluation-induced increase in the price of rice. Even
if the associated changes were allowed to occur, however, substantial
disruption of consumption patterns can be anticipated, in the short-
to-medium run.
Both production and consumption patterns may be
difficult to change in less than one harvest year given the realities
of cropping cycles.
The simulation conducted was based on the anticipated price
changes that would occur if border prices for rice and peanuts rose by
50 percent.
Because concrete numbers on how recommended structural
adjustment measures are likely to affect relative prices are not
available, the simulation analysis uses educated assumptions about how
much a devaluation would be likely to raise both the domestic retail
price of imported rice and the domestic producer price of peanuts.
Assuming that the full impact of the real income effect on both
the consumption and production sides is passed through to consumers,
one would anticipate the following consequences:
Households in the northern basin will be worse off (17
percent drop in real income);
Households in the southwestern basin will have no change in
real income;
Households in the remaining four rural zones will be better
off (1-18 percent).
Using a more probable scenario that permits only half of the
price benefit on the production side to be passed through to
producers, one finds that:
Households in the southeastern Peanut Basin are the only
ones to realize a net increase in real income (5 percent);
Households in the northern basin experience greatest loss
(20 percent) and those in southwest the second greatest loss
(8 percent);
Households in the remaining three zones experience 3 percent
drop in real income.
;
232
Urban households in Kaolack and Tambacounda will clearly be
affected negatively as they are big consumers of imported rice and
they do not produce exportable crops.
These results show the average effect for all households in a
zone.
Some preliminary assessment of the differential impact on rich
and poor households has been conducted using the different consumption
and production patterns observed for these groups.
Results show that
imported rice provides an important share (19 to 39 percent) of total
calories in the diets of the poorest quartile of households in three
zones (the northern Peanut Basin,
Colobane in the central basin, and
the southwestern basin).
This implies that the negative income
effects linked to more expensive rice consumption are likely to affect
poor as well as rich households in rural areas -- rice consumption is
not limited to the wealthy who are better able to survive a drop in
real income.
The purpose of this analysis is not to provide arguments against
reforms in macro-economic policy that are clearly needed for long-run
development.
The analysis provides policy makers with information
about which groups of people are most likely to be negatively affected
by the policy changes in the short run, alerting the GOS to the
advisability of developing countervailing measures to protect the most
at-risk groups from the negative short run effects.
Implementation of
such countervailing measures provides greater potential for the policy
changes to survive and elicit the desired long run macroeconomic
effects.16
Patterns of household income diversification
Knowing what factors are driving households to diversify income
and the long-run economic implications of existing patterns of
diversification is an important prerequisite to designing rural
development policies that will foster growth in rural incomes and
increased agricultural production.
Furthermore, understanding income
diversification strategies of rural households permits one to better
judge household-level response to proposed policy changes. To better
grasp these elements, a typology of income diversification patterns
and a model of factors driving households to diversify were
developed.17
l6 In the past, policies that have increased the price of
imported rice have been met with civil disturbances and subsequent
lowering of the rice price.
The most recent case was in 1988 when the
imported rice price of 160 FCFA/kilo was reduced to 130 FCFA/kilo.
" There is an extensive literature on income diversification in
Asia, but evidence from Nigeria (Matlon), Burkina Faso (Reardon,
Delgado, and Matlon), and Niger (Hopkins and Reardon) suggests that
there are some important differences between Asian and Sahelian
diversification patterns.
The topology and model presented here are
:
:
:
.
‘.
233
Topology of diversification patterns: The topology provides a
framework that facilitates the tasks of (1) diagnosing the underlying
causes of the diversification in each zone, (2) understanding what
factors are motivating production and consumption behavior in each
zone, and (3) anticipating the type of household response that
different policy interventions are likely to elicit. Three major
styles of diversification have been identified:
(1) an outward-oriented
"coping" pattern characterized by:
declining levels of cropping income over time due, in
large part, to a poor, deteriorating natural resource
base;
a dearth of activities with production and consumption
linkages to agriculture
an important share of household income coming from off-
farm activities only in emergency situations following
severe crop shortfalls
consistently low total household income due to poor
cropping outcomes and limited off-farm opportunities;
(2)
an outward-oriented, lonq-run, adaptive pattern characterized by:
declining levels of cropping income over time (as in
the "coping" case) or low levels of cropping income
because technological transformation has not yet taken
place and/or price risk is high due to poorly developed
markets and physical isolation from centers of demand
nonfarm income often earned outside of the zone of
residence;
little reinvestment, if any, of nonfarm income into
cropping activities;
total incomes that are low but generally adequate
(higher than those for the "coping" case)
in the long-run, households may move out of crop
production entirely;
specific to Senegal and based on survey findings; the
conceptualization of the different factors to consider in the topology
and model drew heavily on the analyses of the other West African
research cited above.
.
.
234
(3) an inward-oriented, lonq-run, adaptive pattern, characterized by:
good cropping performance stimulating growth of
production and consumption linkages;
off-farm activities that complement rather than
substitute for cropping income;
income from off-farm activities that is reinvested in
crop production;
the presence of well developed transportation and
market infrastructure
a policy environment that fosters development of factor
markets (particularly credit and labor).
The "copinq" pattern is found in the Niakhar zone of the central
Peanut Basin.
Households in this zone had the lowest average incomes
in the study during both survey years.
Following the extremely poor
1988/89 harvest, off-farm income was 53 percent of total income.
Following the better 1989/90 harvest, cropping income doubled and off-
farm income dropped to only 20 percent of total income. Even with the
doubling of cropping income, however, total income remained low.
An example of the outwardlv-oriented, lonq-run, adaptive pattern
is found in the northern Peanut Basin, where cropping potential is
extremely low due to poor rainfall and degraded soils.
Diversification is based to a large extent on migration and provision
of products and services to Dakar.
Another zone that appears to be moving toward an outwardlv-
oriented strateqy is Senegal Oriental.
In this case, however, the
strategy does not appear to be well adapted to the zone.
Senegal
Oriental has promising cropping and livestock potential given high
average rainfall, less degraded soils than the Peanut Basin, and
abundant supplies of crop and pasture land due to low population
density.
Nevertheless, households earn a relatively low cropping
income ker AE and a high share of "outwardly-oriented" off-farm
income.
There are many indications that market failures and lack of
infrastructure are constraining agricultural productivity and
unnecessarily pushing Senegal Oriental's households into migration (22
percent of household income in 1989/90).
Senegal Oriental clearly has agricultural potential. The policy
issue here is how to design policies that will encourage households to
" The zone had the lowest average cropping income in 1989/90
and the third lowest in 1988/89.
,’
.
.
:
,‘.
.’
235
invest in the intensification of agriculture rather than in outward-
oriented off-farm activities, particularly migration.
An inwardly-oriented adapted strategy is currently found in the
two zones in the southern Peanut Basin and Colobane in the central
basin.
These zones have relatively low levels of migration income.
Much of the off farm income in these three zones is linked to crop
production activities (peanut marketing and processing, cereal
marketing, restaurants).
Animal husbandry is also an important source
of revenue in Colobane and the southeast but not in the southwest.
The challenge in these zones is to provide the incentives
necessary to improve agricultural productivity.
Without growth in
agriculture the impetus for growth in the local non-farm sector will
be lost.
If this happens, the prevailing credit constraints will be
magnified because reinvestment of off-farm income in crop production
will decline.
Determinants of off-farm income shares: Understanding why
diversification off-farm is taking place is important to policy
analysts because households will respond differently if the factors
pushing them to diversify are different. Information about factors
driving diversification can be used to develop nonprice policies that
will speed up and increase market response to price signals.
It is relatively easy to understand the role of climate and
natural resource constraints in pushing households into "outwardly-
oriented" off-farm activities in zones such as the northern basin. It
is more difficult to assess the role of inadequate infrastructure,
credit constraints and other market failures.
Regression analysis was
used to gain a better understanding of the relative importance of
these factors.
Elasticities estimated for these variables suggest that:
*
A 10 percent increase in the cereal sufficiency ratio
decreases the share of off-farm income by 7 percent;
*
A 10 percent increase in land assets (proxied by cultivated
area) decreases the off farm share by 5 percent;
*
A 10 percent increase in livestock assets decreases the off-
farm share by 3 percent;
*
A 10 percent increase in agricultural and food credit
decreases the off-farm share by 1 percent.
The elasticity for the cereal sufficiency ratio is large and
negative, as expected.
Households better able to meet food security
needs from home production have less incentive to seek off-farm
,
236
income, all else being equal." This suggests that programs or
policies which reduce food insecurity (food aid, food banks, or
credit, for example) could encourage households in zones with good
agricultural potential to concentrate more on agricultural production.
The same types of programs in zones with an outwardly-oriented
adaptive pattern of diversification or a coping pattern would be
unlikely to elicit greater agricultural productivity.
The elasticity for land assets is also large and negative as
anticipated.
All else equal, more land permits households to earn
more cropping income by producing the desired mix of cereals and cash
crops; higher cropping income reduces the pressure to move off-farm to
supplement income.
This result is not easy to translate into policy
because it involves land tenure issues that were not covered by the
study.
The implication is, however, that larger farms should be
encouraged if the objective is to reduce unnecessary diversification
that reduces productive efficiency.
Larger herd size also decreases the share of off-farm income
because livestock provide savings that are easily converted to cash
when households need liquidity for food or input purchases, reducing
the need to move off-farm.
Policies to improve animal health and
water supplies (a major constraint to animal husbandry in several
zones) when the carrying capacity permits larger herd sizes would be
most likely to diminish off-farm diversification and promote more
integrated cropping/livestock systems in zones with inwardly-oriented
diversification patterns.
Such policies would be unlikely to improve
crop productivity in zones with coping or outwardly-oriented, adapted
strategies.
Although the size of the elasticity for credit is smaller than
that for other factors,
it is an extremely important variable from a
policy perspective.
The credit elasticity implies that better access
to credit permits households to increase the share of cropping income
in total income (that is, it diminishes the need to seek off-farm
income).
This suggests that improving access to credit might permit
households
to use labor which is now engaged in off-farm activities to
improve farm productivity -- planting trees or building windbreaks,
for example.
In sum, the significant coefficients on three of the above
variables -- cereal sufficiency ratios, livestock holdings, and access
to credit -- provide ample support for the hypothesis that growth in
agricultural productivity is being constrained by liquidity problems
that push farmers into greater income diversification than would be
the case were institutions for credit and savings better developed.
" Cereal sufficiency is defined relative to beginning stocks of
cereals, avoiding problems of reverse causality.
237
Although these regression results provide some insights into the
nature of rural income diversification in the Peanut Basin, the model
did not directly address the links between risk and the share of off
farm income in total income.
Additional work on the risk aspects of
diversification is clearly needed.
POLICY IMPLICATIONS OF THE STUDY
The most important policy contribution of the IFPRI/ISRA study
are the insights it provides on the radical differences in household
consumption and production patterns found across rainfed areas of
Senegal that have long been considered relatively homogeneous for the
purposes of agricultural policy design.
Differences in production and
consumption patterns explain,
to a large extent, why households do not
respond to price policy incentives in anticipated ways.
The major
production and consumption patterns found and the implications of
these findings for policy design are summarized below for each zone.
Implications for hiqh potential areas
In the IFPRI/ISRA study, Senegal Oriental and the southern Peanut
Basin zones are located in high potential areas. These zones have
relatively good rainfall and less degraded soils than the rest of the
Peanut Basin.
The study clearly shows that Senegal Oriental is embarking on a
path of "outwardly-oriented" income diversification. failing to
realize its agricultural potential,
The zone is characterized by very
low returns to agricultural labor, very low participation in cereal
markets, low caloric intake, low total income, high outmigration, and
low reinvestment of migration or other non-cropping incomes in
cropping activities.
Study results suggest that implementation of price policies
(increasing producer prices or decreasing input prices) without
attention to increasing market infrastructure, decreasing
transportation costs, and improving crop production technologies is
unlikely to increase agricultural production in the zone.
Low returns to agricultural labor suggest that labor enhancing
technologies need to be further developed and encouraged with
appropriate extension and marketing programs.
Animal traction, for
example, is much less advanced here than in the Peanut Basin.
The
high cost of modern inputs is also a problem, reducing net returns to
cropping labor and discouraging input use, even when credit is
available.
Although production technologies are constraining, the major
cause of the unrealized potential appears to be physical isolation
that limits inter-regional trade and keeps input costs high. Policies
to reduce transportation and transactions costs associated with inter-
.’
:
:
:
I
.
. ‘.
238
regional trade could reduce the wholesale price of Senegal Oriental's
maize in other zones of Senegal and make it more competitive.
A more
competitive price might not stimulate increased human consumption in
the areas of Senegal that consume millet and rice, but there are other
sources of potential demand (poultry producers, for example) that one
can expect to expand rapidly in the near future.
Relatively thin market infrastructure within the zone also
mitigates against intra-regional cereal marketing, making it difficult
for cereal deficit households to purchase local production from other
farmers in the zone and unnecessarily encouraging consumption of
cereals imported into the zone.
Removal of constraints to the
creation of local markets could contribute to better intra-regional
cereal trade that, in turn, might reduce seasonal fluctuations in
caloric intake.
The southeastern Peanut Basin is the most striking example of a
zone pursuing an "inwardly-oriented adaptive" strategy of income
diversification.
It is a high potential zone that comes closer to
realizing its agricultural potential than Senegal Oriental, but
remains constrained by a number of factors.
Incomes in this zone are
the highest in the IFPRI/ISRA survey, yet caloric intake is frequently
inadequate and returns to family agricultural labor remain lower than
those in the lower potential zones of the central Peanut Basin. There
is a need for improved technologies that enhance labor productivity
and increase yields.
Land clearing and weeding, for example, are
extremely labor intensive in this zone.
Fertility enhancing inputs
(fertilizer, manure, compost) are rarely used.
Results show that animal husbandry is a major source of regular
income in the southeastern Peanut Basin rather than simply a savings
account that is depleted when harvests are poor or ceremonies take
place.
Research on improved agricultural technologies should take
livestock assets into account (better use of manure, for example).
Policies that increase returns to animal husbandry (marketing
information systems and permanent sources of water, for example) would
also increase incomes and overall productivity in this zone.
The major overriding constraint in the southeastern basin,
however, appears to be a lack of liquidity. Households in market
villages are able to overcome liquidity constraints, to some extent,
through off-farm activities.
As a result, households in market
villages have higher total income per capita, generally higher
cropping and livestock incomes, and better levels of caloric intake.
Households in nonmarket villages, conversely, have much lower levels
of participation in the off-farm sector, lower total incomes, and
lower caloric intake.
Two types of policies can help reduce the liquidity constraint in
this zone -- savings and credit programs and better infrastructure
(roads and markets).
.
.
,.
239
Policies that encourage savings and credit are the most direct
way of addressing the liquidity problem.
Such programs should be
designed to increase investment in agriculture. The relatively high
cropping incomes currently realized suggest that the potential for
credit repayment exists.
The low returns to labor and low levels of
modern inputs used suggest that credit could encourage greater
investment in existing technologies (animal traction, fertilizer,
improved seeds) and therefore improve cropping productivity.
Increases in cropping productivity stimulated by credit and
savings programs would also increase off-farm incomes through
consumption and production linkages.
For these linkages to be fully
exploited policies are needed to foster better roads and markets
within the zone.
Although the southeastern basin is better endowed
with markets and roads than Senegal Oriental, this type of
infrastructure is still less developed than in the other areas of the
Peanut Basin.
This is the only Peanut Basin zone where households in
nonmarket villages need to make relatively long trips over difficult
roads to get to market villages.
This diminishes the incentives to
engage in off-farm activities that thrive in market village settings
due to the consumption and production linkages fostered by the zone's
relatively dynamic agricultural sector.
The southwestern Peanut Basin is also following an "inwardly-
oriented adaptive" diversification strategy. Households in this high
potential zone are characterized by relatively high levels of total
income, fifty percent of which is earned in off-farm activities.
Many
of the off-farm activities are directly linked to agricultural
production (cereal and peanut marketing, in particular).
Returns to family labor used in cereal production are among the
highest in the study, but even higher returns (by 50 percent) are
realized in peanut production.
Average levels of caloric intake are among the highest in the
study.
The share of imported rice in cereal consumption is also
relatively high but typical of areas with important shares of off-farm
income.
Unlike the southeastern basin, the southwest benefits from
relatively dense market and road infrastructure, although the latter
deteriorated significantly during the survey period.
Given the extremely large difference in returns to labor for
peanut and cereal production, producer cereal price increases would
have to be substantial to elicit even small increases in cereal
production.
It is even conceivable that an increase in the consumer
price of imported rice might have a greater impact on cereal
production than a change in the producer price of coarse grains given
the relatively high rice consumption in this zone.
\\
240
Part, but not all, of the difference in returns to labor for
different crops is due to the production of confectionery peanuts and
the accompanying fertilizer credit. The extent to which the crop
(confectionery peanuts) or the fertilizer enhances returns to labor
needs to be more thoroughly examined and the implications for input
and commodity policies reviewed.
The unusually large share of off-farm income in this zone of high
returns to cropping labor reflects both the liquidity constraint found
in the southeastern basin and a land constraint that is beginning to
make itself felt.
Credit programs are also indicated to assure that
capital is available to invest in land enhancing technologies and that
labor which could be used to improve land productivity during the dry
season (composting, tree planting, and windbreaks, for example) is not
needlessly allocated to noncropping activities.
Despite the high reliance on off-farm income in this zone, the
IFPRI/ISRA data suggest that most of the off-farm income is dependent
on the level of productivity in the agricultural sector.
For many
households, a generalized drop in crop production in the zone leads to
a drop in off-farm income.
This correlation highlights the importance
of investing in agricultural research to improve crop and land
productivity thereby encouraging general economic growth in the zone.
Implications for lower agricultural potential areas
The northern and central Peanut Basin zones are considered the
low potential zones in the IFPRI/ISRA study due to less reliable
rainfall and seriously degraded soils.
Among the three study zones in these low potential areas, one has
developed a successful pattern of diversification away from crop
production, another has been able to maintain a relatively good (but
precarious) level of income based on a mix of cropping, livestock, and
off-farm activities, and the third is barely coping.
These zones each
require a different set of policy initiatives to assure food security
and encourage economic growth.
The northern Peanut Basin has adapted well to recurrent droughts
and low cropping productivity by following an "outwardly-oriented,
long-run adaptive" pattern of diversification. Income per capita and
caloric intake levels fall about mid-way in the distribution across
zones.
Share of noncropping income is, however, the highest (about 70
percent) as is consumption of imported rice (about 70 percent of
cereals consumed).
Crop production constraints are severe in this
zone and households are clearly moving into other sectors of the
economy.
The livestock sector is particularly important, both as a
source of income from animal husbandry and as a source of income from
activities linked to livestock production (sale of byproducts,
commerce and transport of animals to markets in Dakar).
~
.,
.
.
:
:.
‘.
241
Because households in this zone have adapted to the adverse
cropping environment, food security is currently less of a problem for
them than it is for some households in higher productivity zones.
Incomes are generally adequate to purchase food and imported rice is
readily available in weekly markets and village shops,.
This situation
could change dramatically, however,
if the consumer price of imported
rice were to increase.
Economic growth in the zone is highly dependent on transportation
and trade links with Dakar where a large share of production from off-
farm activities is sold.
Policies that reduce real incomes in Dakar
(a sharp increase in the price of imported rice, for example) could
reduce demand for livestock products produced in this zone.
Policy initiatives that can encourage further growth in the
sectors now providing a large share of northern basin income are
improved water sources for livestock,
improved livestock marketing
information, and credit for both livestock producers and traders.
The Colobane zone in the central Peanut Basin falls into the
"inwardly-oriented, adaptive" pattern of income diversification at
present.
It is surprisingly well adapted given that cropping
potential is lower here than in the southern basin and more risky.
There is, however, a need to improve incomes in the zone. Average
incomes are in the low to middle range.
The share of off-farm income
is relatively high.
Off-farm income is derived from a mix of
primarily local activities supplemented with some temporary migration.
Caloric intake is the highest in the survey both years despite
cereal production shortfalls.
The zone exhibits the second highest
level of rural rice consumption.
The zone benefits from relatively good road and market
infrastructure that encourages households to sell coarse grains and to
purchase rice to a greater extent than most other zones. Road
quality, however, deteriorated substantially during the survey period.
The highest returns to family labor used in cropping activities
are found in this zone.
Returns in peanut production are higher than
those for cereals (30 percent). Large farm sizes, extensive use of
traction equipment, and light, easily cleared and weeded soils appear
to contribute to the higher returns to family labor. This zone has
the second largest average farm size,
suggesting that quantity (if not
quality) of land is not yet a constraining factor.
The sustainability of these relatively high returns to labor is
an important issue.
No chemical fertilizer was used by any sample
household in this zone during the entire study period. Few fields
received manure.
Yields per hectare are low and soil fertility will
continue to decline if measures are not taken to stop erosion and
,
restore nutrients to the soil.
.I
.
.
*’
.
.
;
242
Efforts to improve cropping technologies in this zone should
concentrate on fertility enhancing and soil preserving objectives.
Livestock herds in the zone are relatively small limiting the
possibilities for expanded use of manure.
Liquidity problems in this zone are resolved to a large extent by
selling livestock.
Following the poor peanut harvest in 1988 a large
share of receipts. from livestock sales were used to replenish supplies
of peanut seed and secure food.
Credit programs could diminish the
need to liquidate livestock, but the issue of capacity to reimburse
cropping credit in this high risk zone is a real one.
This is the
type of situation where food for work programs could be employed
successfully in conjunction with credit and fertility enhancing
efforts to raise overall levels of income in the zone and move toward
more sustainable agricultural production.
Given the high levels of rice consumption, households in the zone
will be particularly vulnerable to increases in consumer rice prices.
The higher returns to labor in peanut production suggest, however,
that a very substantial change in relative prices will have to occur
before the zone increases the share of cereals in total production.
Niakhar in the western part of the central Peanut Basin is the
one survey zone that clearly falls into the "outwardly-oriented,
coping" pattern of diversification.
It is the zone with the most
discouraging economic prognosis.
The zone is characterized by the
lowest total incomes during both survey years. Cropping income is
constrained by low soil fertility and a shortage of land. Returns to
family labor in peanut production are the lowest in the survey, while
returns to cereal production are among the highest.
Diversification
into off-farm activities to supplement cropping incomes is poorly
developed.
The zone exhibited the lowest average levels of caloric intake
following the poor 1988 harvest,
illustrating that off-farm income did
not adequately substitute for crop shortfalls.
Rice consumption in
this zone is relatively low given the sizeable cereal production
shortfall.
Households show a strong preference for millet, even when
cereals need to be purchased.
More land intensive and fertility conserving production
technologies need to be developed for the zone. Explanations need to
be sought for the very low returns to peanuts. Given the high
population density and relatively low involvement in off-farm
activities, this zone could be expected to have the labor necessary
for labor intensive soil enhancement activities (composting, tree
plant, or ridging, for example).
Greater attention also needs to be paid to developing
opportunities for off-farm employment in this zone where few
production and consumption linkages are evident due to the low
agricultural productivity.
Survey results do not provide a clear
:
243
picture of why households in this zone have not developed the same
type of outwardly-oriented adaptive income strategy that was found in
the northern Peanut Basin.
Households in the zone own the largest livestock herds in the
entire sample, yet animals are sold primarily for food purchases and
ceremonies.
Income from sales is seldom invested in cropping or other
economic activities.
Pasture land is limited in the zone itself,
forcing households to graze animals outside of the zone most of the
year.
This situation not only diminishes the benefits to be gained
from manure but also limits a household's ability to sell animals
easily.
This area of the central Peanut Basin would be expected to need
food aid following crop shortfalls to a greater extent than the
northern basin where off-farm activities are better developed.
Food
for work programs to provide food security and encourage labor
intensive land improvements would also be appropriate.
Given the much
higher returns to millet production and the zone's preference for
millet consumption, research to increase millet yields per hectare may
be more indicated than improved peanut yields.
Relatively low levels of rice consumption mean that households in
this zone would suffer less than others from a sharp increase in the
price of imported rice.
The low levels of peanut production and low
returns to peanut labor mean, however, that the zone is also not
likely to benefit from any policy induced increases in producer prices
of peanuts.
Review of general findings of relevance across all zones
Imported rice is consumed in important quantities in many rural
and all urban households in the IFPRI/ISRA study zones; therefore, any
policy that increases the price of imported rice will harm rice
consuming households by reducing their real income.
The potential for price policy to reduce rice consumption in
rural zones does not appear to be strong given that rice was an
important share of cereals consumed even when coarse grains (including
processing costs) were considerably less expensive.
Rice consumption is likely to increase in the future as
commercialization of rural areas advances and nonmarket villages adopt
the same high rice consumption patterns currently exhibited by market
villages.
Farmers are unlikely to increase cereal production given
prevailing price incentives that result in higher returns to labor for
peanuts than cereals in most zones.
The anticipated positive income effect of policies that increase'
the producer price of cash crops will be dampened significantly by the
,’
.’
244
fact that cash crops provide considerably less than half of total
household income.
This dampening will be exacerbated if policies to
increase export prices in domestic currency are accompanied by removal
of current export price subsidies, as is likely to be the case.
Poorly developed infrastructure, marketing institutions, and
credit policies diminish the relative profitability of crop production
in high potential zones, pushing households into migration.
This is
most apparent in Senegal Oriental.
In all zones with reasonable cropping potential, growth in
agricultural productivity is constrained by a lack of liquidity that,
in the absence of reliable institutions for savings and credit, pushes
farm households further into off-farm production than would otherwise
be the case.
Failure to understand income diversification patterns can lead to
erroneous conclusions about the relative well-being (particularly the
food security status) of households in zones with poor agricultural
productivity.
This can lead to inappropriate food aid and food for
work programs.
The generally low levels of per capita income and the large share
of households failing to earn incomes that cover minimum needs,
seriously constrains investment in productivity enhancing agricultural
technologies.
The problem is compounded by poorly functioning savings
and credit markets.
Overall about 30 percent of sample households fell below the
posited "minimum needs level" in 1988/89. This remained true in two
zones following better harvest in 1989/90.
So long as income for a
large share of households fails to cover minimum needs, investment in
agriculture from deferred consumption is unlikely to take place.
There are surely limits beyond which it is not possible for
households to adjust their consumption further downwards without
incurring severe negative impacts on food security.
It is also hard
to see how households could expand their agricultural production,
without specific attention to improving the natural resource and
policy environments in which households make their production and
consumption decisions.
It is clear that rural households remain
extremely vulnerable to both weather shocks and the probability of
further austerity measures affecting their real incomes.
245
REFERENCES CITED
BADIANE, Ousmane and Sambouh Kinteh.
Aqricultural Trade Pessimism in
West AFrican Countries and the Possible Role of Reqional Markets.
Washington, D.C.: IFPRI, forthcoming.
HOPKINS, Jane and Thomas Reardon.
Aqricultural Price Policy Reform
Impacts and Food Aid Tarqetinq in Niqer. Washington, D.C.:
IFPRI,
April 1993,
MATLON, Peter.
Income distribution amonq farmers in northern Niqeria:
Empirical Results and Policy Implications. African Rural Economy
Paper No. 18, Lansing, Michigan: Michigan State University,
1979.
REARDON, Thomas.
“Income diversification in Burkina Faso, unpublished
mimeo". 1989.
REARDON, Thomas; Christopher Delgado and Peter Matlon. "Determinants
and effects of income diversification amongst farm households in
Burkina Faso".
Journal of Development Studies, January 1992.
.I
:
:
:
.
I
‘.
246
LIST OF PRINCIPAL PROJECT DOCUMENTS
Dakono, Astou. 1991.
Comparaison des revenus au capital et au travail
dans les activites aqricoles et non aqricoles au Nord du Bassin
Arachidier.
Internship Report, Project IFPRI/ISRA, Dakar.
Diagana, B. 1991. Quelques caracteristiques tendancelles du modele
de consommation en milieu urban seneqalais. IFPRI/ISRA, April,
Dakar.
Diagana, B., A. Fall, V. Kelly, and T. Reardon. 1990.
Caracteristiques des zones d'etude du projet IFPRI/ISRA:
Resultats des enquetes de reconnaissance des villaqes. IFPRI/ISRA
Project Document 3, October, Washington, D.C.
Fall, A. 1991.
Composition multisectorielle et distribution du revenu
dans le bassin arachidier:
Remuneration du capital et du
travail.
Tenure paper, ISRA.
IFRPI/ISRA Project Document 5,
January, Washington, D.C.
fall, A. and B. Diagana. 1992.
Les activites non aqricoles et la
capacite de financement de la consommation alimentaire par les
menaqes ruraux au Seneqal.
Paper prepared for regional workshop
on: activites informelles non agricoles et securite alimentaire
au Sahel, 25-28 novembre, Prisas, Bamako, Mali.
Fall, A., V. Kelly, and T. Reardon. 1989. Household-Level Survey
Methods Used in the IFPRI/ISRA Study of Cons umption and Supp1.y
Impacts of Aqricultural Price Policies, IFPRI/ISF IA Project
Document 2, September, IFPRI, Washington, D.C.
Kelly, V., B, Diagana, and A. Fall.
1991.
La consommation des
cereals en milieu rural seneqalais. IFPRI/ISRA, July, Dakar.
Kelly, V., J. Duncan, A. Fall, and B. Diagana. 1992. Transforminq
Survey Data in Demand Models.
IFPRI/ISRA/PRISAS-MSU Working
Paper, November, Washington, D.C. and E. Lansing.
Kelly, V. and T. Reardon. 1989.
Samplinq Methods Used in the
IFPRI/ISRA Study of Consumption and Supp1.y Impacts of
Aqricultural Pricinq Policies. June. IFPRI, Washington, D.C.
Kelly, V., T. Reardon, A. Fall, B. diagana, and L. McNeillv. Draft
9
Final Report for the IFPRI/ISRA Study of Consumption&dsly
Impacts of Aqricultural Price Policies. December II392, IFPRI,
Washington, DC, Kelly V., T. Reardon, A. F-..,
'~111. -.
R w,..J..,,..,
nianana , and L.
McNeilly. 1992.
Kelly, V., T. Reardon, A. Fall, B. Diagana, and L. McNeilly. P992.
Food Consumption in the Peanut Basin and Seneqal Oriental:
October 1988 - September 1989.
IFPRI/ISRA Project Document 7,
September, IFPRI, Washington, D.C.
.
.
;..
.:
,
.I’
:
I.
,‘,
.’
247
Kelly, V., T. Reardon, and L. McNeilly. 1991. Sales, Purchases, and
Gifts of Cereals and Pulses in the Peanut Basin and Seneqal
Oriental.
IFPRI/ISRA Project Document 6, May, JFPRI, Washington,
D.C.
Reardon, T., A. Fall, V. Kelly, C. Delqado, P. Matlon, and 0. Badiane.
1992a. ’ Is Income DiversjficationU'aqriculture-led'
in the WASAT?
Survey Evidence and Development Strateqy Implications. Invited
paper presented at the First Biennial International Conference on
African Economic Issues, June, Lome, Togo,.
Reardon, T., V. Kelly, B. Diagana, and A. Fall. 1992b. Potential
Welfare Impacts of Trade Reqime Chanqes on Households in Seneqal:
Focus on Devaluation.
Paper prepared for the seminar on regional
integration of agricultural markets in West Africa.
Saly
Portudal, Senegal, 4-7 December. IFPRI, Washington, D.C.