; vv—vwwm
.‘ 4

"I

a“

V‘_. '35”-

IT’I‘F Irv

 

‘

l ‘ I 'V 7‘ v
Wu) _' |: l' ”I
. 0" ~
'51:." 2. . L urn

, ‘n‘liélix \‘A‘ . l

V,“

. " ’H‘ . 1
"‘ ‘ .. i "I '.‘r:-:"'
- . » '"' ' " “ ; "2' ‘ Eut’myfi
""‘f:: . V ' ‘ ‘ r :4 ' V ' .{F "‘.“"."l J" - --Lu «‘1

3,23,]- ‘ ‘1‘1‘4 HJil“:

Po
5»:
. (’

 

...

£- -- "I
‘ . . v» £9 ;
you . " '

1 9

b1»
s w

.- L?
.»‘-.. Vi?

‘ ‘ o -.
fifflfii-{hm
"r

"at

‘1

t‘

“I

" ., I I! :*
‘4‘
L)“

.

J
(<7
.1

I

‘ HI.
. l r‘ ‘ . ,.' " t ‘ - ‘3 3
n ‘ ' '

. "3 9‘

l
H v_

U4

3 .'

all“ 3-" 11;)!“7
#43:!!11]

. 3‘ II "

 

- -w
—-—-a.

B7‘7

,’r{r'
,.

.1“ s? ‘ ,7
. : . zwr‘b'V-vu
.v '9 1-" 1

   

 

 

»~— “WWIIMiflixmmum“

‘,-_ .H ., 3 1293 3106513

[.3390 r'
1

war Cw—I’.d.uw—w: g-bw“:
I‘ ,9,‘ —91,
-I

I mans“ -

This is to certify that the

thesis entitled

A COMPARATIVE ANALYSIS OF ALTERNATIVE IRRIGATION
SCHEMES AND THE OBJECTIVE OF FOOD SECURITY: THE
CASE OF THE FLEUVE REGION IN SENEGAL

presented by

Fadel Ndiame

has been accepted towards fulfillment
of the requirements for

 

 

Master's degree in Agricultural Economics

M49074

Date March 22, 1985

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

MSU

 

 

 

RETURNING MATERIALS:
Place in book drop to

 

 

 

LJBRARJES remove this checkout from
.—::_. your record. FINES W'i'l'l
be charged if book is
returned after the date
stamped below.
:77 , gm: fir
tubs—eye] ]
fiazn
/ 72

, 1
NOV 0 4 2002

 

 

A COMPARATIVE ANALYSIS OF ALTERNATIVE IRRIGATION
SCHEMES AND THE OBJECTIVE OF FOOD SECURITY:
THE CASE OF THE FLEUVE REGION IN SENEGAL

By

Fadel Ndiame

A Thesis

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Agricultural Economics

1985

ABSTRACT
A COMPARATIVE ANALYSIS OF ALTERNATIVE IRRIGATION

SCHEMES AND THE OBJECTIVE OF FOOD SECURITY:
THE CASE OF THE FLEUVE REGION IN SENEGAL

By
Fadel Ndiame

The large and small irrigation perimeters developed
for rice production in Senegal's Fleuve 'region are evaluated
for their contribution to the country's objective of food
security. An index of food security, based on the per
capital availability of cereals in producing households, is
used to compare the two schemes. An econometric analysis is

also developed to approximate the quantitative relationships

involved in paddy rice production. The data used here were

collected in a Policy-oriented farm
Because of the larger farm

per capita availability of cereals

survey.
plots in large perimeters,

is more important there

than in small perimeters, despite higher yields in the

latter. The

estimated production

equations reveal the

impact of inputs and other localized factors on paddy

production. The findings lead to

economical strategies of food

recommendations for more

security based on the

simultaneous development of irrigation with the rainfed and

flood-recession agriculture.

TABLE OF CONTENTS

List of Tables . . . . . . . . . . . . .
List Of Figures. 0 O O O O O O O O O O 0

Chapter

1

THE RESEARCH OBJECTIVES AND COMPONENTS

A.
B.

C.

D.
E.

F.

A BRIEF HISTORY OF IRRIGATION IN THE

Introduction . . . . . . . .

Concepts of Food Security and Their

to Our Research Problem . . . .
Objectives of the Study . . . .
The Conceptual Framework. . . .
Data Requirement and Collection

1. Objectives of ”Politique Agricole“

2. Strategy of Data Collection.

Organization of the Study . . .

FLEUVE REGION. 0 O I O O O O C O O

A. The Experiences with Irrigation in the

B. The Recent Experiences with Irrigation in

Pre- Independence Era. . .

the Fleuve Region . . . . . . .

19

19
22
23

1. The Experience with Large Perimeters .
2. The Recent History of the Village
Irrigation Systems . . . . . . . . . .

3 Structure of Productions and Cereals
Ava11ab11ity O O O O O O O 0 O O O O O I O O

A. Output, Transactions and Cereals

B.

Availability. . . . . . . . . .

1. Production and Yields. . . .
2. Transactions . .

3. Indices of Cereal Availability

Areas Cultivated. . . .

C. Power Source and Level of Technology.

ii

35
35
44

47
49

Chapter

D.

1. Human Labor. 0 O O O O O O O O O O O
2. Animal Traction. . . . . . . . . . .
3. Tractor Mechanization. . . . . . . .

The Other Inputs. . . . . . . . . . . .
1. SeEds. O O O 0 O O O 0 O O O O O O O

1.1 Origins of Seeds . . .
1.2 Quantity and Quality of

Seeds. .

2. Ferti11zers. O O O O O O O O O O O O

4 STRUCTURE OF PRODUCTION AND CEREALS
AVAILABILITY IN THE SMALL PERIMETERS . . .

A.

B.
C.

Cereals Availability, Productions,
Transactions. . . . . . . . . . . . . .

1. Productions and Yields . . . . . . .
2. Transactions . . . . . . . . . . . .

2.1 Sales and Purchases of Cereals .
2.2 Gifts and Tithe. . . . . . . . .
2.3 Debt Reimbursements. . . . . . .
3. Indices of Cereal Availability . . .

Areas Cultivated. . . . . . . . . . . .
other Inputs. 0 O O O O O O O O 0 O O ..

1. Labor and Animal Traction. . . . . .
2. Seeds and Fertilizers. . . . . . . .

2.1 Origins of Seeds . . . . . . . .
2.2 Technical Factors. . . . . . . .

5 THE DETERMINANTS OF RICE PRODUCTION IN
THE FLEUVE REGION. 0 O O O O O O O O O O O

A.

The Physical and Technical Determinants

1. Physical Factors . . . . . . . . . .
2. Technical Factors. . . . . . . . . .

The Economic and Policy Factors . . . .
Some Sociological and Institutional

Determinants. . . . . . . . . . . . . .
Elements of Methodology . . . . . . . .

iii

 

60

6O

60
66

66
68
70
72

75
77

77
77

77
79
84
86

86
86

90

96
100

Chapter Page
6 THE ECONOMETRIC MODEL. . . . . . . . . . . . . . 102

A. Introduction. . . . . . . . . . . . . . . . . 102
8. Presentation of the Models. . . . . . . . . . 104

1. The Economic Model . . . . . . . . . . . . 104
2. The Statistical Model. . . . . . . . . . . 109

C. Analysis of the Empirical Results . . . . . . 112

1. Analysis of the Estimated Parameters . . . 112
2. some Examples. 0 I O O O O O O O O O 0 O O 121

0. Implications of the Econometric Model . . . . 128
E. Limitations of the Study. . . . . . . . . . . 131

7 SUMMARY, CONCLUSIONS AND RECOMMENDATIONS . . . . 134

A. Introduction. . . . . . . . . . . . . . . . . 134
B. Summary and Conclusions . . . . . . . . . . . 135
c. Recommendations 0 O O O O O O 0 O O O O O I O 139

1. Policy Recommendations . . . . . . . . . . 139
2. Needed Research. . . . . . . . . . . . . . 145

APPENDICES. O O O O O O O O O O O O O O O O O O O O O O 148

Appendix A Questionnaire. . . . . . . . . . . . 148
Appendix B Codes of Zones, Perimeters and

Producers Groups . . . . . . . . . . 152
Appendix C Presentation of the Econometric

Analysis Results . . . . . . . . . . 155

C.1 Summary Table of the Different

Area and Yield Equations . . . . 155
C.2 Presentation of the Most Satis-

factory Area and Yield Equation. 158
C.3 Transformation Procedures of

the Econometric Analysis'

Results. . . . . . . . . . . . . 151

SELECTED BIBLIOGRAPHY o o o o o o o o o o o o o o o o o 166

LIST OF TABLES

 

Table Page
1 Evolution of Selected Indicators of SAED's I 28
Performance from 1980-1981 to 1982-1983.
2 Average Production and Yields in Cereals-Growing 36
Households in the Large Perimeters by Zone and by
Crop.
3 The Transactions of Cereals Per Household in Large 43

Perimeters by Crap and by Zone (1981/82).

4 Indices of Cereal Availability in Producing 46
Households by Crap and by Zone (In Millet
Equivalents).

5 Distribution of Areas Cultivated Per Producing 48
Household by Zone and by Crop (In Hectares).

6 Average Number of People Per Household Growing a 50
Particular Crop by Zone

7 Structure of the Debt in Large Perimeters: Percent 52
of Cases Hhere Debts were Incurred for Various
Activities by Crop and by Zone.

8 Percentage of Farmers who Received Their Seeds 54
From Various Sources by Crap and by Zone.

9 Seeding Rates by Crop and by Zone (In Kilograms 56
Per Hectare).

10 Fertilizers Available Per Household and Rate of 57
Fertilization by Crop and by Zone.

11 Average Production and Yields Achieved in Cereals- 61
Growing Households in the Small Perimeters by Zone
and by Crop.

12 Average Sales of Cereals Per Household to Various 67
Buyers in the Small Perimeters by Crap and by Zone.

H m
w 0'

14

15

16

17

18

19

20

21

22

 

Average Quantities Provided as Gifts Per Households

(In Kilograms) and Percentage of Households who
Provided the Tithe (In Parentheses)

The Structure of Debts in Irrigated Perimeters

A. A Percentage of Cases in Which Debts were
Incurred for Various Activities of the Small
Perimeters.

8. Modes of Reimbursement: Percentage of Cases
in Which Debts were Reimbursed in Cash or in
Kind, by Zone.

C. Average Quantities of Rice and Amounts
Reimbursed Per Household and by Zone.

Indices of Cereals Availability in Producing
Households Operating in Small Perimeters, by
Crop and by Zone.

Distribution of Areas Cultivated Per Household
on the Small Perimeters by Zone and by Crap
(In Hectare).

Average Numbers of Pe0ple Living in Cereals-
Growing Households of the Small Perimeters of
the Zone.

Origins of the Seeds in the Small Perimeters.
Percentage of Farmers Hho Received Their Seeds
From Various Sources by Crap and by Zone.

Seeding and Fertilization Rates in Small
Perimeters, by Crop and by Zone (In Kilograms
per Hectare).

Empirical Estimates of the Regression's
Coefficients.

Private Profitability of Fertilizers at the
Margin In Three Selected Producer Groups During
Three Seasons.

Results of the Different Estimation Procedures.

vi

Page

69

71

71

71

73

76

78

80

81

113

126

156

Figure

m

LIST OF ILLUSTRATIONS

Sampling Design Used in the Fleuve Region

Distribution of Paddy Production in the
Delta and Middle Valley Zones.

Distribution of the Yield of Paddy Rice in
the Delta and Middle Valley Zones.

Distribution of Paddy Production in the Aere
Lao and Matam Zones.

Distribution of the Yield of Paddy Rice in
the Delta and Middle Valley Zones.

Presentation of the Different Zones of the
Study.

viii

Page
17

39

4O

64

65

14

CHAPTER ONE
THE RESEARCH OBJECTIVES AND COMPONENTS

A. Introduction

The agricultural sector plays an important role in the
Senegalese economy, as it employs an important part of the
country's population and makes a significant contribution to
its gross domestic product. A major characteristic of
Senegalese agriculture is its dependence on natural factors
which remain the main determinants of production. In the
1960's, the Senegalese authorities pursued a strategy
initiated by the French colonists and based on the
production of peanuts for exports. Meanwhile, the foreign
exchange earnings allowed the country to import rice.
Despite the policies of diversification undertaken during
the last two decades, peanuts and products based on them
remain dominant in the Senegalese economy.

In the 1970's, the combination of drought conditions,
unfavorable deveIOpments in the international economic
envirbnment and inapprOpriate domestic policies resulted in
a profound food crisis, which is still being felt by the
country. The growing cereals deficits represent the most
obvious manifestation of Senegal's inability to feed

itself. To tackle the problem, the Senegalese authorities

_r_._,....—--— ..._

have setflgpfighflambitious program to_redefineg the components

of 19 129!_ agricultural policy aimed at achieving food

2

security. Indeed, the VIth plan provides the objectives of
the agricultural policy and expresses the need for more
investigations of important policy and institutional
issues. Two important objectives mentioned by the plan are
the improvement of the country's food self-sufficiency and
the promotion and development of irrigation systems able to
assure greater food security. In this respect, the
development and exploitation of the Senegal River resources
constitute a major component of a strategy to lessen the
impacts of adverse climatic conditions and to promote a
modern agro-industry.

In fact, such a program 15' being carried out both at
the national and regional levels through the operations of
SAED1 and 0.M.V.S.2 SAED was created in 1965 and was
assigned three main functions:

1. To develop and manage the state-owned lands
to allow their profitable exploitation.

2. To develop irrigated rice by assisting and
providing extension services to groups of
farmers organized in cooperatives.

3. To organize the processing and marketing of
harvested output.

During the years, SAED has gone through many changes

both in its status and role. The most important ones were

 

1Societe d'Amenagement et d'Exploitation des Terres du
Delta et des Vallees du Fleuve Senegal et du Faleme.

2The “Organisation Pour la Mise en Valeur du Fleuve
SenegalI was created on March 11, 1972 and includes Mali,
Mauritania and Senegal.

3

brought about by the institutional reforms that took place
with the 1984 New Agricultural Policy (NAP). Basically, the
NAP reiterated the Senegalese government's decision to
sharply reduce the state's intervention in the rural economy
and to transfer more responsibilities to farmer organi-
zations and the private sector. This implied an important
reduction in the size and responsibilities of the develop-
ment agencies. For SAED, the measures introduced by
the NAP created additional constraints on its irrigation
programs.

At the regional level, the OMVS countries attempt to
join their efforts to struggle against hunger, desertifi-
cation and to exploit the resources of the common river.
Four Specific development objectives were mandated at OMVS:

1. To increase and secure the income of the

river basin's population (1,612,000 people,
representing 16% of the population of the
three countries involved).

2. To guarantee the ecological equilibrium of

the zone. This reflects the concern that the
rocess of production is likely to have
external“ effects detrimental to the
physical environment and calls for a rational
system of exploitation.

3. To alleviate the dependency of the region's

economies on weather-related factors. For
this purpose, the OMVS countries decided to
build the Diama and Manantalie dams which
will allow them, among other things, to
regulate the flow of the Senegal River.

4. To accelerate economic development through an

intensive regional cooperation involving the
harmonization of domestic policies and,

ultimately, the complete integration of the
involved economies.

4

The scarcity of local capital and the scope of the
different programs require an efficient allocation of
existing resources. Accordingly, the question of the
comparative performance of different irrigation schemes is
especially important. But, to the extent that the various
policies pursued by the ‘national and regional institutions
attempt to achieve ”food security“, it is relevant to
analyze the concept and to examine its meanings in the
Senegalese context.
8. Concepts of Food Security and Their

Relevance to Our Research Problem

The issue of food security has been associated in the
literature with various elements. Some authors emphasize
the need to stabilize food availability, mainly via the
minimization of the variance of quantities and prices
through time and space. Other authors analyze the question
from the viewpoint of particular participants in the food
system and focus on questions ranging from the needs to
ensure rural self-sufficiency, and/or guarantee cheap food
to urban consumers to the demands for the political
liberation of peasants (Lele et al., 1984; de Janvry,
1981; Dumont, 1982). Finally, other authors analyze food
security in a macro-economic context, insisting on the
requirements to link policies on agricultural policies,

prices and employment (Mellor, 1984; Timmer et al., 1983).

5
All those aspects are important and this makes an analysis
of “food security“ extremely complex.

The views of different authors, complementary or
conflicting, are closely associated with their conception of
underdevelopment and their vision about the process of
economic development.3 It is, however, very important to
keep in mind that the analytical process itself involves
sets of choices that cannot be made exclusively on technical
grounds. The choices are relative to elements ranging from
the definition of the research problem, the selection of
analytical categories, and techniques to the choice of
relevant performance variables. Under conditions of
uncertainty and when interests conflict, those choices
determine whose interests will count (Schmid, 1984).
Furthermore, because the choices made in the analytical
process can affect the way alternative projects compare,
they can be systematically based on the objectives of the
macro-economic policies.

In this study, we conceive food security as Senegal's
ability to assure adequate per capita food supplies to the
different segments of the country's population at any point
in time. Thus, our conception of food security requires
that each Senegalese has access to the adequate quantity and

quality of food during his lifetime. For the sake of

 

3See Eicher and Staatz (1984) for an overview of the
evolution of ideas on agricultural development.

6
simplicity and because of resource limitations, we shall
limit our research problem to the analysis of cereals
productions and transactions in the Fleuve region and their
relationships to food security. We will, however, keep the
other related issues in mind.

Several points are important in the analysis of cereal
production in Senegal. The first point relates to the crop
which is produced. The importance of this point stems from
the dichotomy existing between the national demand ofcereal
and the country's current production potential. Indeed, the
national demand for cereal, for historical, economic and
psychological reasons, tends to favor imported rice against
the more easily produced millet and sorghum. In fact, under
current technologies and cost structures, the opportunity
cost of producing rice in Senegal is hypothesized to be
high, both with respect to the cost of importing it and the
cost of producing millet and sorghum (Dumont, 1982. p. 36).
Yet, the production of irrigated rice in large perimeters,
conceived to generate enough surplus to meet the country's
needs, represents a major component of Senegal's import
substitution strategy (Founou, 1980). On the other hand,
rice production in village irrigated systems allows peasants
to secure an important source of food, mainly during years
of droughts.

The second related point deals with the question of
irrigated versus rainfed agriculture. The promotion of

irrigated agriculture constitutes a major step in mastering

7

water and helps to eliminate the disastrous consequences of
drought. As such, irrigation, by eliminating weather—
determined variations in production, significantly
contributes to food security. However, it is important to
pay attention to the costs of irrigation, mainly in the
large perimeters. The scarcity of local resources and the
existence of cheaper sources of water create the need for an
awareness of the Opportunity costs of irrigation.

The last point deals with the relationships between
cereals production and several types of inputs. The
elements that are relevant in explaining the level of
production of each cereal and its' distribution through time
and space include the physical and climatic environment, the
quantities and distribution of' inputs, the production
techniques, the structure of relative prices and the
existence of market outlets for a given crop. In addition,
organizational factors, embodied in the type of perimeter
involved, are thought to influence the level of yield and
the quantities produced in each cereal.

In sum, an analysis of cereal production and food
security in Senegal should pay attention to three important
issues.

First, the current production patterns reflect
producers' perceptions of their food security. In addition,
the distribution of production among several crops and the

structure of the national demand of cereals reflect the

8

value that participants of the food system place on the
different cereals. Thus, the various cereal promotion
policies should account for the mix of production which is
the most consistent with the different participants'
preferences.

Second, irrigation constitutes only one component of a
more global agricultural system including the traditional
rainfed and flood-recession agriculture, animal husbandry
and other off-farm activities. The various components of
the system are related by complex and interdependent
relationships. Therefore, any assessment of irrigation in
the Fleuve region should incorporate those relationships.

The third issue deals with the analysis of the
relationships between "food security" and "food self-
sufficiency" in the context of Senegal. Such analysis
would determine, given the production potential and the
structure of national and international prices, the
appropriate mix of local production, imports and food aids
that contributes most effectively to food security.

In the next section, we shall define the general and

specific objectives of this study.

C. Objectives of the Study

Our study has two general objectives:

To understand the relationships between alternative
irrigation schemes and cereal production in the Fleuve
region.

To discover how the policy variables involved in
the cereal sector can be influenced to achieve food security
in Senegal.

More Specific objectives are set to help meet the general
ones:

(1) The first specific objective is to understand
the evolution of the Fleuve'irrigation schemes and to
analyze their influence on current and future
perimeters.

(2) The second specific objective is to present
the distribution by type of perimeter of cereal
production in the Fleuve region and to test for the
existence of significant differences in performance
variables across perimeters. Our performance variables
are the quantities of cereals available per capita in
producing households. The yields and production of
each crop and the transactions that take place are
major determinants of cereal availability.

(3) The third specific objective is to
investigate the relationships between the variations of

the performance in paddy rice production with the

10

changes in selected characteristics of the production
units: the variations in input availability, the type
of perimeter, the location along the river and other
organizational factors operating at the producer group
level. Such an analysis is important because of
the high value attached to this crop by Senegalese
government officials and the need to assess the
effectiveness of the resources allocated to its
production.

(4) A final specific objective is to discover how
the design and organization of the irrigated perimeters
can be influenced to achieve the objectives of the

Senegalese agricultural policies.

0. The Conceptual Framework

Our analysis will involve several steps using various
analytical techniques. The first step will provide a brief
presentation of the origin, evolution and current state of
irrigation in the Fleuve region. The second step of the
analysis will describe the performances of selected
perimeters during the 1981/82 season. The tabular
presentation of the performance variables will allow us to
compare the selected perimeters by crop cultivated. The
third step involves the use of standard multiple regression
analysis to approximate the relationships involved in paddy

rice production.

11

The descriptive analysis of the second step involved
comparisons of average figures computed for perimeters
having certain defined characteristics. But a large
variability is hypothesized to exist within perimeters of a
given type. Such variance is related to factors such as
input availability, differences in location, and differences
in production techniques among perimeters of the same size.
Furthermore, important organizational factors are believed
to operate at the producer group level where important
decisions relative to the management of the perimeters are
taken. The information related to those elements is "lost“
because of the aggregation at the perimeter level.

In contrast, by using the households as basic unit of
analysis, we can retain the information related to the
variations internal to perimeters. To do this, dummy
variables will be used to represent such qualitative
elements as type of perimeter, location along the river and
the organizational factors operating in the producer group.
To the extent that these dummy variables are significantly
different from zero, the coefficients will allow us to
isolate the specific effects of the qualitative factors

which influence paddy rice production in the Fleuve region.

E. Data Requirements and Collection
The data used in this study were collected in 1983 for

a study of agricultural policy in Senegal. The researcher

12
participated in the conception and execution of the data
collection in the Fleuve region. In order to see how this
data can be used in our current study, we shall briefly
present the objectives and the data collection strategy of
the 1982 “Politique Agricole" study.

1. Objectives of “Politique Agricole“

The general objective of “Politique Agricole" was to
describe the structure of cereal production in Senegal.
More specifically, three types of information were needed:

(a) The factors of production available and

utilized during the reference year 1981/1982.

(b) The levels of production achieved and their
relationships with the factors of production.

I (c) The transactions of cereals and, more

Specifically, the movements of products in different

areas of the country during the year.

Policy makers wanted this information in order to
understand relationships between inputs and output on one
hand, output and transactions on the other hand. The
knowledge of these relationships could help guide policies
to promote cereal production in Senegal.

The researcher was in charge of organizing the data
collection in the Fleuve region. The strategy used is

explained in the next section.

13

2. Strategy of data collection.

The data were collected in a survey using stratified
sampling methods. For this purpose, it was necessary to
classify the Fleuve region in strata that were homogenous
with respect to relevant agro-economic criteria. Such a
scheme allows one, in theory, to obtain reliable information
from randomly selected subsamples that can be used to make
inferences about the larger population.

In the case of the Fleuve region, however, the
administrative limits could not be criteria of
classification because they do not provide homogenous
strata. Indeed, all along the river, from Saint Louis to
Bakel, there is a set of physical, demographic,
administrative and organizational variables whose
simultaneous variations provide a noticeable Specificity to
any given location.

With the collaboration of SAED, the Fleuve region was
classified according to three criteria:

(a) Location along the Senegal river.

(b) Type of perimeters involved: large or small
perimeters.

(c) Type of agriculture: irrigated or rainfed.

The combination of these criteria allowed us to

establish the following stratificationz4

 

4See the situation of the zones on the map.

14

mwmfi .xmmou ESL; umuaou< "mucaom

552.com 425...... a

$355.. 3:3 as;

     

«omme

......
............
..o.

    

£4249

.- o.
...........
.......

no

         
 

3...: no 339: .. a 2:3 .5

-Pcma Pmem on; mem<
co mcauas »a__a> a_uuwz au_ao
nwcma _Pmsm 6:» ho mcmpms on“ Co mcmpms
-Pcma mace; u_cma omen;

       

mac... .pm

.....
.....
I...
.....
ooooo
.....
Oo-

 

 

 

mz «145

  

1.40— cmdzu

 

auzum mzu mo mmcoN acmcmumpo as» we copumucommca

15

1. The large perimeters of the Delta.

2. The large perimeters of the Middle
Valley.

3. The small perimeters of the Aere Lao
zone.

4. The small perimeters of the Matam zone.

5. The zone of the “traditional“
agriculture, which is a fictive zone corresponding
to the rainfed agriculture undertaken in all the
areas of the sample.

In each zone, a sample of three perimeters was selected
and five 'groupements de producteurs“ were randomly chosen
from each perimeter.5 Finally, a sample of twenty members
of each “groupement de producteurs” was randomly selected
and surveyed.6 The “groupements” constitute the major
farmer groups which deal with SAED about the operations of
each perimeter. A groupement is generally composed of
people of the same village and ethnic group.

A farmer who was selected to participate in the study
was interviewed about his own fields of cereals, whether
irrigated or not. Then members of his household were also
surveyed about their fields. Separate records were kept for

each crop grown in a different field by each member of the

 

5A list of the perimeters and ”groupements" selected is
provided in Appendix B.

6In some 'groupements“ more than 20 members were
interviewed and in others fewer members were surveyed.

16

household. During the data analysis, the information
relative to different fields of the same crop was aggregated
at the household level. As a result, the most basic case is
represented by a given crap grown in a given household.

Every farmer was interviewed about his resources
allocated to cereal production, the output achieved and the
ways in which it was used. A capy of the questionnaire used
is attached in Appendix A. The diagram of the sampling

design in the Fleuve region is presented in Figure 1.

F. Organization of the Study

The remaining chapters of this study are organized as
follows:

Chapter Two provides a brief presentation of the
history of irrigation in the Fleuve region and summarizes
the major characteristics of the currently irrigated
perimeters. ' Chapters Three and Four deal with the
description of the major results achieved _in selected large
and small perimeters during the 1981/82 season. Chapter
Five presents the determinants of rice production in the
Fleuve region. The elements analyzed include physical and
technical factors, economic and policy variables and some
institutional and sociological determinants of production.
Chapter Six presents an econometric model that approximates
the quantitative relationships involved in paddy rice
production. Area and yield equations were estimated and

they allowed us to derive a production equation relating

Nozm

 

 

 

v m x — m 4 a m m

21.x

 

 

 

 

 

 

 

 

 

 

 

 

a o c v m 3 m z m c m t c c c n d m c x

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

N: :ccmmgc.am

m.c=sm a. muse—Lac www.ca cum: ‘: nzm m.~:<o amnioa

18
changes of output to the variation of several types of
inputs. Finally, the estimation of a production equation
allowed us to investigate the implications of the econometric
analysis both .for future studies and for policy options.
Chapter Seven summarizes the findings and conclusions of this

study and provides some recommendations.

CHAPTER THO

A BRIEF HISTORY OF IRRIGATION IN
THE FLEUVE REGION

The current irrigation schemes existing in the Fleuve
region are the most recent forms of a series of attempts to
develop this form of agriculture in the region. AIn fact, as
early as 1824 the French colonists initiated, with the
creation of Richard Toll, the various irrigation projects
that were implemented throughout the last two centuries.
The earlier projects have generally served as basis of
development of the most recent ones, not only by providing
the technical basis, but also through the teachings of their
successes and failures. For these reasons, it is useful to
briefly view the earlier experiences of irrigation conducted
during the colonial period. Such a presentation will
facilitate the analysis of the current schemes.
A. The Experiences with Irrigation in

the Pre-Independence Era

The development of irrigation in the Fleuve region was
a component of the Agricultural Colonization scheme. This
project was undertaken in a ge0political context marked by
the Napoleonic wars in Europe and its ramifications in Hest

Africa, the abolishment of the slave trade and the political

19

20
emancipation of the former French colony of Saint Dominque,
which became Haiti.1

As those events denied France important sources of
revenue, the Agricultural Colonization scheme was initiated
as an attempt to develop irrigation in the Senegal River.
The project was aimed at establishing a modern agriculture,
oriented towards the exports of tropical commodities. The
year 1822 is generally mentioned as the beginning of the
project. In fact, it corresponds to the year during which
the French governor of Saint Louis, Baron Jacques Roger,
hired a gardener named Richard to execute the project. The
current city of Richard Toll (Richard's garden in Ouoloff)
is named after him.

An important component of the project dealt with the
adaptation of European crops to the environment of Senegal
in addition to research on local plants. The labor used in
the Agricultural Colonization scheme included French
technicians and several categories of local workers
involving forced labor emanating from former slaves.

The irrigation techniques used was based on the
chain-pump which involved the development of a series of
small dams to provide water to plots. But, the Agricultural

Colonization scheme did not have the expected success and

 

1Hilliam McClinton Patterson (1984, p. 35 and
following) provides a more complete treatment of the
historical context. This part draws heavily on the
Patterson dissertation.

21

was terminated in 1831. Its failure has been attributed to
technical factors such as the salinity of the soil, problems
associated ~with the irrigation technique and the local
climate to _which European settlers could not adjust well.
Beside these problems, some sociological factors contributed
to the failure, including conflicts over land tenure and
labor shortage in an area depopulated by three centuries of
slave trade.

After the Agricultural Colonization scheme, the next
project took place a century later with the develOpment of
several irrigation systems using a “controlled submersion"
technique. This practice involved “tiered waterbeds
surrounded by dikes to keep the water in higher parts, and
to control the inundation of the lower parts. When the
flood of the river was high, a system of flood gates allowed
the .water to enter the waterbed using the force of
gravity.“ (Patterson, 1984, p. 43).

This technique was carried out by the “Mission
d'Amenagement du Senegal“ at Guede, where 1,000 hectares
were irrigated. AS in the earlier Agricultural Colonization
scheme, forced labor was used in the Guede project. Hhile
the focus was earlier put on growing cotton for export, the
project concentrated its effort on growing rice in order to
lessen the shortages of the colony during World War II.

After thewar, the “Mission d'Amenagement du Senegal"

initiated an ambitious project of 50,000 hectares of rice at

22
Richard Toll. The plan was to be executed in 10 years
using pump irrigation and with full mechanization. But, the
actual results were much more modest: only 6,000 were
finally developed.

At the independence in 1960 several development
agencies were successively created with the primary
objective to develOp irrigation in the region. Indeed, in
the five years following the country's independence, three
organizations were set up: the "Organisation Autonome du
Delta“ (0A0), the "Organisation Autonome de la Vallee"
(OAV), and, finally, SAED. All these organizations were
charged with initiating and training peasants in irrigated
agriculture, with the primary objective being to produce
enough rice to reduce the country's imports of this
commodity. They also built on the infrastructural and
technical legacy of the earlier projects.

The organizations have had various degrees of success
in carrying out their assigned missions. The experience of
SAED will be analyzed in our next section.

B. The Recent Experiences with Irrigation
in the Fleuve Region

The operations of SAED have been characterized by
important changes in its intervention, its technical
procedures and management methods. In fact, one can
distinguish the earlier model of the large perimeters,

mainly deveIOped in the Delta, from the more recent schemes

23
of village irrigation systems commonly called small

perimeters.

He shall examine these two types of schemes

successively.

1. The Experience with Large Perimeters

The large perimeters represent the dominant type of
irrigation scheme in terms of area exploited. In 1981 they
encompassed 10,061 hectares, or three-quarters of the
irrigated area in the SAED intervention zone (Bonnefond et
al, p 6). AS suggested by their name, the large perimeters
are characterized by their dimension, which varies from 250

to 2,400 hectares; their relatively high levels of

mechanization; their large operating costs and the size of
the individual plots (1.5 hectares on the average).
Initially, the efforts in the large perimeters were directed
toward the development of rice and tomatoes in the Delta.
Indeed, in 1965 SAED undertook an ambitious project to
exploit 30,000 hectares in order to produce 80,000 metric
tons of paddy rice per year. The operation required
displacing 8,000 families from the Middle Valley to the

Delta. Up to 1 hectare of irrigated land per family was

provided by SAED, which also supplied some inputs and

services to peasants, such as seeds, fertilizers, land

preparation and irrigation.

The following results were recorded in 1975: 8,034

hectares were exploited instead of the planned 30.000

24
hectares. Only 8,500 settlers were involved in the project
while 40,000 people were initially targeted. Total rice
production amounted to 11,000 metric tons per year (Lam,
1983). This relatively modest performance has been
attributed to two factors: the irrigation models used and
policy mistakes made by SAED.

The water control systems used by SAED were modeled on
the controlled flooding polders of Indochina. This
irrigation model appeared to be totally inapplicable to the
semi-arid conditions of the Senegal River Basin during the
years of drought. The following problems were particularly
noted: the flooding was late or ~inadequate and salt water
intruded from the sea. As a result of these problems, the
investments in irrigation were shifted toward systems of
total .water control which required pumping and leveling.
However, this option increased both capital and operating
costs very markedly. Such cost increases required double
cropping in order to spread fixed costs over greater
quantities of output. But double crappingwas extremely
difficult under the specific conditions of the river.
Indeed, the intrusion of salt water made it impossible to
double crap in the Delta. It was also very difficult to
implement it upstream during the dry season because greater
demands were placed on the limited volume of water available

in the Middle Valley.

25

In addition, the late arrival of rains in the region
and the lack of water for pre-irrigation severely
constrained the crapping calendar and required the
introduction of mechanized techniques (of production. These
costly techniques created a heavy burden on maintenance and
input delivery systems (AID/OMVS, 1982).

The policy mistakes were related to the highly
centralized mode of management adapted by SAED. No
initiative was left to farmers with respect to the operation
of their plots. Extension agents who were receiving bonuses
per area cultivated were not motivated to care about the
quality of the work. In general, the yields achieved by
farmers varied from one perimeter to another depending upon
the quality ’of the irrigation system and the services
provided by SAED.

Farmers were charged 25,000 CFA francs per hectare for
irrigation services while the actual costs are estimated at
80.000 CFA francs per hectare, implying subsidies of 55,000
CFA francs per hectare. Dumont (1980, p. 201) indicates
that farmers could generally afford to pay their subsidized
expenses when the harvests were good. But, in cases where
the yields were low, most settlers had to incur debts that
were payable in the following harvests. As a result, most
farmers could not keep enough rice to meet the consumption

requirements of their family (Dumont 1980. p. 202).

26

In the 1980's, important changes were brought about in
the role, mode of operation and performance of SAED.
Indeed, in 1981, the “lettre de Mission“, a contractual
document jointly conceived by the Senegalese government,
SAED and various donors agencies redefined the objectives of
SAED. The emphasis was put on the following points:

(a) A better management of the existing equipment
and infrastructure in order to ensure the long term
viability of the investments.

(b) The need to reduce the operating costs of
SAED via a progressive reduction in its direct
involvement in production “and marketing activities.
Those functions would be transferred to producer groups
and to the private sector.

The actions of SAED were supposed to be complemented by
a set of short-term and long—term policies undertaken by the
Senegalese state to ensure a consistent cereal strategy,
equilibrate its finances, elaborate a policy of management
of the water resources resulting from the construction of
the future dams of Diama and Manantali, and promote
research.

Since the early 1980's, SAED has accomplished
noticeable progress in its mode of operation and its
economic performances. This progress was apparent in the
evolution of several indicators: production yields,

percentage of nonharvested hectares, level of debt

27

repayment, and quantities of paddy rice marketed through the
official channels. These indications are presented in Table
1.
These improvements are credited to a better utilization
of the equipment thanks to the greater emphasis that SAED
put on maintenance. These efforts allowed SAED to
rehabilitate and bring into production parts of the
developed land that were initially abandoned. Such progress
was encouraged by the institutional reforms brought about by
the “Lettre de Mission" which provided SAED with a greater
flexibility in the management of the perimeters. Meanwhile,
a policy of farmers' participation was pursued through an
effort to decentralize the decision-making process and
promote farmers' organization. As a result, the operating
losses were reduced, as SAED's deficits were less than the
levels allowed for by the “Lettre de Mission“ both in
1982/83 and in 1983/84.2

It is still early to fully assess the real impact of
those reforms, as they have been brought about by external
pressures because of the financial problems of SAED and the
difficulties of the state. In fact, the viability of SAED's
current plans will be tested by the ability and willingness

of the private sector and farmers' organizations to

 

2See SAED's “Elements de Reponse au Memorandum 'Note de
Synthese Etablie a la Suite de la Reunion de Concertation
C.C.C.E.lFAC/IDA/US AID les 9 et 10 Mai 1983 a Nashington',"
July. 1983.

28
Table 1

Evolution of Selected Indicators of SAED'S
Performance from 1980-1981 to 1982-1983

 

 

 

 

 

 

 

YEARS
Variables 1980-1981 1981-1982 1982-1983
Cropping Intensity
(In x) 75 78 92
Area Allocated to
Rainy Season's Paddy
Production (In ha) 8,857 9,097 12,220
Now Harvested Area
(In 1) 21 20 8
Paddy Production -
(In kg) 27.683 31.050 49.990
Yieldsa (In kg/ha) 4,000 4,300 4,400
Quantities Marketed
through Official
Channels (In kg) 8,864 9,558 19,375

 

aBased on harvested area source: SAED, July, 1983.

29
successfully carry out SAED's previous functions. An
analysis of the village, irrigation system will help
understand where the difficulties lie and whether or not
private individuals have the incentives to carry out these
functions.

2. The Recent History of the Village
Irrigation Systems

The small-scale irrigation systems were introduced
along the river in 1974. They are characterized by their
relatively small dimensions, with an average size of 18
hectares per perimeter. The size of individual plots is
also small, averaging .20 hectare. In 1981, the small
perimeters covered 3,465 hectares, representing one quarter
of the area developed by SAED.

A noticeable feature of the small perimeters is their
relatively low capital costs: 300,000 CFA francs per
hectare versus 1,500,000 CFA francs per hectare in large
perimeters (Bonneford et al. 1981, p. 6). Beside the lower
costs involved, the small village perimeters present other
advantages over the large ones. Dumont (1980, p. 202)
estimates the yields at 4.5 to 5 tons per hectare, with
peaks of 10 tons per hectare for the paddy in contrast to
average yields of 1 to 2 tons in large perimeters. In
addition, he indicates that in the majority of cases,
peasants can make two harvests of rice per year, with annual
yields between 8 and 9 tons: i.e., 3 to 4 times the levels

reached in the large perimeters. As a result, the ratio of

30
invested capital to gross income was estimated at 1:1 in
small perimeters versus 21:1 in the large perimeters
(Dumont, 1980, p. 202).

The greater efficiency of the small village perimeters
has been credited to their particular conception and their
suitability to local conditions. Diemer et al (1983, p. 3)
characterize the schemes as a special case of bottom-up
development. They stress two factors in their rapid
expansion. The first is that village management capa-
bilities have been developed by existing social insti-
tutions. The initiation and operation of a small perimeter
are done as follows:

SAED supplies a group of peasants with an irrigation
pump, inputs and marketing facilities. It also provides
technical advice on irrigation structures, channel layout on
a mutually agreed-upon Site and supervises the division of
land into plots. However, the other management tasks are
left to farmers. They make the decisions on. the distri-
bution of water, the financial operations and on the
physical maintenance of the schemes. The group of farmers
undertake land clearance, site preparation, construction of
channels and banks, etc. Representatives of the group
organize the collective works, raise funds for fuel,
replacement of the pump, and order the fuel and inputs
needed from SAED. Despite variability from one producer

group to another, the yields on the small perimeters are

31

higher than those obtained in the large schemes_where SAED
applies more direct methods of management.

The second important feature of the small-scale
irrigation is the correSpondence of the perimeter's size to
the scales on which village institutions Operate. Diemer et
al. (op. cit., p. 7) reveal, in a study of the Halpulaar
ethnic group, the way in which authority in the schemes is
modeled on the village power structure.3 The distribution
of power in Halpulaar village reflects the organization of
society along rights acquired through birth in a certain
lineage. For instance, certain functions of leaders are
reserved to members of the free4born lineage. The authors
found Similar arrangements in the schemes, where only a
certain group of people is eligible for important
positions. Age groups and youth clubs constitute social
institutions which extend and reinforce traditional
authority in the operation of the irrigation schemes. Their
principal function is to provide a framework where members
of the group learn how to participate in collective decision
making and how to enforce those decisions. These social
institutions are believed to play an important role in the
operation of the schemes by developing standard operating
procedures which guarantee the resolution of conflicts and

the enforcement of order.

 

3Casey Clyde Franklin (1983) made similar discoveries
in his study of the Soninke Experience in the upper valley
of the river.

32

In this respect, Tiffen (1983. p. 2) argues that the
initial conception and operation of the small-scale
irrigation perimeters corresponded to the objectives of
meeting the needs of a subsistence agriculture, which was
severely affected by the series of droughts of the early
1970's. As long as it had those objectives, the small scale
irrigation schemes could mobilize the village resources, and
the traditional structures could regulate the conflicts
between participants. The point emphasizes the possibility
that movements toward a cash-crop oriented agriculture may
encourage individualistic behaviors by farmers and could,
therefore, weaken the incentives for village cooperation and
collective works.

But the producer groups, which constitute the current
framework of operation of the village irrigation systems,
are far from homogenous groups. In fact, Fresson (178,
p. 23) describes how conflicts of kinship and differences
of ward complicated by political differences appeared when
Nguidjilone 1, one of the early SAED perimeters, was started
in 1975. Attempts to constitute homogenous groups according
to both residential locality and political persuasion were
not successful.

The issues related to land rights are generally solved
by villagers. Initially, the location of perimeters on
“Fonde” lands where the traditional agriculture does not

normally take place helps to minimize the land tenure

33

problems. It is likely that tenure disputes will become
more likely in the future as the “Fonde” lands become more
valuable because of irrigation. Meanwhile random assignment
of the irrigation plots to the households constitutes a
device which guarantees peasants equal access to irrigated
land. But, the small-scale irrigation system still presents
several problems that need to be addressed.

One category of problems has to do with the size of the
perimeters and the difficulties associated with plans to
increase it. Despite the relatively high yields, the output
per capita is still very small and does not allow farmers to
meet their consumption needs. Consequently, they keep their
traditional dry land and flood recession fields. As a
result, there exists a complex set of interdependent
relationships between the different components of the
production system.4

Because of hypothesized positive link between area
cultivated and output, it is generally felt that plots of
.75 hectare per family would be the appropriate size for the
small perimeters (Sy et al, 1981, p. 22). But the extension
of current plots is constrained by the fact that irrigation
sites are exhausted. In addition, the extension of existing

perimeters would necessitate the use of mechanized

 

4A detailed analysis of the production system is
presented in the general report issued by the French Caisse
Centrale de C00peration Economique: "Evaluation Economique
de l'Amenagement de la Rive Gauche du Senegal.“ December,
1982.

34
equipment, mostly for land preparation. Furthermore, the
possibility of extending the current system brings up again
the need to discover the positive features of the current
perimeters and to build on them. In this respect, it is
important to understand the process by which several
categories of inputs contribute to the production process.
The next two chapters analyze the performances of the two

types of perimeters during the 1981/82 season.

CHAPTER THREE

STRUCTURE OF PRODUCTIONS AND CEREALS
AVAILABILITY IN THE LARGE PERIMETERS

The objective of this chapter is to present the level
Of "food security" that farmers operating in large
perimeters achieved during the 1981/82 season. The large
perimeters studied are those Operating in the Delta and
Middle Valley zones. They correspond to the perimeters of
Boundoum, Lampsar and Richard Toll for the Delta, and
Dagana, Nianga and Guede for the Middle Valley.

The per capita availability Of cereals in producing
households is considered here as -the major index of food
security. As this variable is closely related to the
quantities produced and the various transactions Of cereals
undertaken by farmers, we shall briefly analyze the process

Of production and transactions.

A. Output, Transactions and Cereals Availability

1. Productions and Yields

The major results relative to the production of various
cereals in the two zones are summarized in Table 2. Table 2
diSplays noticeable features with respect to the cropping
pattern in the zones and the differences in production
and yields among the various cereals involved. Only rice

and millet were grown by the households interviewed in the

35

36
TABLE 2
Average Production and Yields in Cereals-Growing

Households in the Large Perimeters by
Zone and by CrOp

 

 

 

 

DELTA ZONE
Standard Standard
Number Production Deviation Yield Deviation
CrOps of Cases (In kg) Of Production1 (In kg/ha) Of Yield1
Paddy Rice 383 4917 215 1938 78
Millet 68 310 41 187 30

MIDDLE VALLEY ZONE

Paddy Rice 411 2834 '280 3788 234
Millet 185 710 61 467 34
Sorghum 235 1551 273 712 78
Maize 67 711 450 697 283
Cowpeas 125 344 117 296 42

 

1The standard deviations reported correspond to those of the mean
productions and yields. They were computed by dividing the standard
deviations Of the households' productions and yields by the Square root
Of the number Of cases. These standard errors can be used to test
whether mean production and yield were significantly different in Aere
Lao and Matam zones.

37
Delta while sorghum, maize and cowpeas were also grown in
the Middle Valley zone.

With respect tO paddy rice, a large variability in
production and yields was recorded in the two zones. In the
Delta zone, paddy production per household ranged from a low
Of O to a high of 30 tons and yield varied from 0 to about
16 tons per hectare. An average quantity Of 4917 kilograms
was produced per household, corresponding to an average
yield of 1938 kilograms per hectare in the Delta zone. In
the Middle Valley zone, paddy production per household
varied from 0 to 85 tons, whereas the yields achieved in
that zone ranged from a minimum Of O to a maximum Of about
20 tons per hectare.1 The average production per producing
household in the Middle Valley zone was equal to 2834
kilograms and the average yield was 3788 kilograms per
hectare.

The wide range Of values recorded for production and
yield in the two zones is believed to reflect the diSpari-
ties which exist among producing households in terms
Of their access tO important resources: land, water, labor,
seeds, fertilizers, etc. In addition, an important factor
which can explain the large variations Of paddy production
has to do with the fact that some households Operating in

the Middle Valley zone can have two crops Of paddy rice per

 

1A yield Of 40 tons per hectare was recorded for one
household Of the Middle Valley zone.

38
year. The different levels of production and yields
achieved in both zones and their frequency distributions are
presented in Figures 3 and 4.

Figure 3 presents the frequency distribution of paddy
rice production in the Delta and Middle Valley zone. Figure
3 indicates that 56 percent of the Delta zone's
paddy-growing households produced between 0 and 5 tons.
Thus, paddy production was larger than 5 tons in 44 percent
Of the cases. For the Middle Valley zone, 86 percent of the
producing households Obtained quantities between 0 and 4
tons. Quantities higher than 5 tons were Obtained only by 8
percent of the households involved in paddy production in
the Middle Valley zone.

The levels Of production achieved in the different
zones are believed to be highly related to the households'
access to inputs, especially land. Therefore, the
quantities produced, per se, do not provide a good picture
of how well resources were allocated in the different
zones. In this respect, the yields recorded in the
different zones constitute a better measure of performance.
The frequency distributions Of the yield Of paddy rice in
the Delta and Middle Valley zone are presented in Figure 4.

Figure 4 indicates that 58 percent Of the Delta zone's
paddy growers recorded yields between 0 and 2 tons per
hectare. Yields between 2 tons and 5 tons per hectare were

recorded in 41 percent Of cases. Only 1 percent of the

39

 

 

 

 

l N

DISTRIBUTION OF PADDY PRODUCTION

THE DELTA AND MIDDLE VALLEY ZONES

 

DJ
2
O
N
>-
B
m 4
Z
o§
N
<5
l-O
gde
:3 (3 E
V
m\'
4I
[

 
  

N\\\\\\\\\\\\\\V

10+

9—10

PRODUCTION (IN TONS)

 

.\\\\\\\‘

«Raw ‘

’ \\\\\\\\\\\\\\‘

  
  
  
 
   
 
   
 
   
 
   

\\\\\\\\\\\\\\V

\\\\\\\\\\\\\\\\\\\V

N\\\\\\\\\\\\V

k\\\\\\\\\\\\\\\\\\‘

\\\\\\\\\\‘\\\\V

\\\\\\\\\\\\\\\\\\\\\\\\‘

t\\\\\\\\\\\\\\\\\\\\\\\\\\\\‘ 1

 

 

4o

:54

30*)

25“r

 

r
In 0

20~r
5-P-
10-1'

’

SUTOHESDOH 30 39V1N3083d

7-8

——.- .——._——-u “r... _- .

-_.._ . _.

._ -- —— - _--v"-— -.

Figure 3

4O

Isi‘.‘l§1"l '1 1' 0.. . 'h 1"

UZON >muj<> UJODE‘ I

§

OZMOMI.

 

MZON (POND

 

 

It‘ll-I‘l'll'.v I

 

‘ '1 9,1Il‘u1" .10.. 1" It | .

Amm<POmI\mZOP 20
+5 him win mid din nlN NIP T...

1..--..

a mc:m_u

 

 

99!- III! 1’-

.‘ll

.."|u.‘l\‘ll IA’I.I|.‘|..|I|' I'll-1.1.0

'1.

WWW“

\\\\\\\V\

m

.\\\\\\\\\\\ \\\\\\\\“

      

WW\\\\\\\\'\\V

\\

  

     
 

\

 

 

            

0.4m;

O

 

mmmmm

mm

   
 

  

I'll" l..

 
 

Lrnu

Ifon

LA on

 

l r00.

WMZON %m1..1_<> HEDGE; OZ< <15m0 MT; 7:
MOE \rOD<O EO 04m; MID; .nLO ZOCIDmEFmD

rllii a l|.ll u! Itali‘-

SG'IOHBSDOH :IO BOVINEIOHBO

 

41

Delta zone's paddy producing households achieved yields
larger than 5 tons per hectare. In contrast, only 29
percent Of the Middle Valley's paddy growers recorded yields
lower than or equal to 2 tons per hectare. On the other
hand, 52 percent of the Middle Valley's producers Obtained
yields of paddy greater than 5 tons per hectare. It is
interesting tO analyze these results in relation to the
characteristics Of the two zones and the different
households' endowments Of resources. Such an analysis will
be undertaken later in this study. Meanwhile, we can
analyze the performance recorded by producers of cereals
other than rice in the large perimeters. The figures are
presented in Table 2.

With respect to millet, higher yields were obtained in
the Middle Valley than in the Delta. Indeed, an average
yield of 467 kilograms per hectare was recorded in the
Middle Valley zone, versus 187 kilograms per hectare in the
Delta zone. The other cereals were grown only in the Middle
Valley zone. The following yields were recorded: 712
kilograms per hectare for sorghum, 697 kilograms per hectare
for maize and 296 kilograms per hectare for cowpeas. The
performance in the production Of different cereals
determined, to a large extent, the transactions that tOOk

place.

42

2. Transactions

In the Fleuve region, the marketing Of cereals is
characterized by the dominant role played by SAED in the
assembly and processing Of paddy. Besides purchases
undertaken at the Officially fixed producers' price, SAED
collects important quantities Of paddy rice in reimbursement
of its input deliveries and services to farmers. The
quantities assembled are processed in SAED'S mills and
distributed to consumers through licensed wholesalers and
retailers.

Up to the early 1980's, SAED had a legal monopoly on
paddy marketing. The other cereals are essentially handled
by informal channels. In fact, there is increasing evidence
that important quantities Of paddy go through those
unofficial channels. The characteristics and the roles
played by the parallel market are not generally well
understood. In this respect, the cereal marketing study
being conducted by I.S.R.A.'s macro-economic unit should
provide useful insights.

Our own study revealed that important transactions of
cereals took place in the large perimeters during the
1981/82 season. Those transactions of cereals included
sales, purchases and the quantities used to reimburse
various types Of debts. Table 3 presents the different

transactions that occurred in the different crops.

43

Table 3

The Transactions of Cereals per Household in Large
Perimeters by Crop and Zone. (1981/82)

 

 

 

 

 

 

TRANSACTIONS
Sales of Cereals Purchases of Cereals In kind Reimbursement
Quantities t of Quantities t of Quantities t of

Crops and Zones In kg Production ln kg Production In kg PrOduction

Delta 1191 24 281' 6 1169 24
Rice -

Middle Valley 205 7 132 5 518 18

Delta 4 1 614 198 35 11
Millet

Middle Valley 27 4 51 a s7 3

Delta NA‘1 NA NA vA NA NA
Sorghum
° Middle Valley 25 1.6 328 21 0 3

Delta NA NA NA NA NA NA
Maize

Middle Valley 1 .1 29 4 17 2

Delta NA NA NA NA NA NA
Cowpeas

Middle Valley 23 6 43 12 O 0
Note:

‘NA means not applicable i.e. the crop was not grown in the zone.

44

In the Delta zone, most transactions involved rice, as
relatively important quantities of this crop were sold,
purchased and used to pay back debts. With respect tO
sales, only marginal quantities of cereals other than rice
were involved in the two zones. 0n the other hand,
purchases .were more common. In both zones significant
quantities Of ricewere used to pay back debts, suggesting
that most debts were incurred in paddy production.

The transactions, along with the quantities produced,
determined what was available for consumption in producing
households.

3. Indices Of Cereal Availability

The per capita availability Of cereal was considered as
the major index Of food security. To measure such
availability, three ratios were computed per household in
the two zones:

(1) Ratio 1 represents total production divided
by the population Of the household. This ratio shows
how much of each cereal was potentially available for
consumption in the households. But, as indicated
earlier, several types of transactions take place and
change cereal availability in the households. To
account for these transactions, a second ratio was
calculated.

(2) Ratio 2 represents the production retained on

a per capita basis, after the sales, the in-kind

45

reimbursement of debts and the gifts performed by
farmers. It measures the extent to which farmers are
able to feed themselves and their family. However,
farmers have the Option to buy cereals either to
complement or to replace their own production. For
this reason, it is necessary to take purchases into
account.

(3) The third ratio represents the overall
availability Of cereals when the quantities purchased
per capita are added to production retained.

The three ratios, which are presented in Table 4,
suggest the following: A

First, both zones have a comparable level of rice
availability. Production per capita was higher in the Delta
but larger quantities were used up there by sales and debt
reimbursement. Even though we do not have an absolute norm
to measure an adequate level Of cereal availability, the
quantities recorded in both zones were low compared to the
210 kg per capita per year generally considered as
desirable.2 The households which undertook some multi-

cropping had more cereals available on a per capita basis.

 

2The Ministry Of Rural Development's estimates of food
needs were the following in 1983 (in kilograms per capita):

Millet/sorghum 110.6
Rice 47.3
Corn 12.3
Hheat 27.9
Others 12.2

Total 210.3

46

’able 4

Indices Of Zereal Availability in PrOducing Households
3y Crop and by Zone (in milled equivalents)a

 

 

CATEGORIES

 

 

‘ _ 3ur:nases of’ Ratio 3“ Tots?
Ratio 1' Ratio 2‘ Cereals per Availability oer
W “‘9 39' C401") ”‘9 99" caoita) Caoita ;Ir hzg,‘ :aoita 'rn vi
Delta 305 173 23 195
II: . 3\:2\
31:9
“izdie Jailey 259 17? 11 191
'3 - 111‘
Zeita 23 21 £0 51
H - 52‘
lilic
loo 9 Kelley 54 33 l -2
- 1:5‘
Ieita 'A p v: _
\l . 3‘
Sorghum
'1dc7e valley 155 1g. :1 .9,
“ . 235‘
Zefza vs NA :A ;
3‘: - 1‘
“aize
“130 e Valley 41 ' 3S 3 33
1 - =71
:eiza vA NA NA 2:
.v - ~13
Sewoeas
Middle Valley 119 56 6 52
IN - 125)

 

“'53 milled equivalents of the different cereals were derived using FAQ'S extraction
rates: 57: for paddy rice, 85% for maize, and 55: for miliet and sorghum. 'FAO'S :CCO
3alance Sheets. 1975.77. Average and per capita fooo Subplies. 1961-65; Average 1367-19’T.Z

{Ratio 1 - Production of the household/population of the household.

'Ratio 2 - 'Proouction - sales - debts)- gifts) , 3Opulation.

“Ratio 3 . Ratio 2 e Purchases per capita.

47
But, as indicated by the number of households involved with
each cereal, not all farmers grew several crops.

DeSpite the importance of purchases, own production
constituted the bulk Of cereals availability in large
perimeters. This makes the analysis of the production
process particularly relevant as it allows us to understand
the relationships between output and several types of

inputs.

8. Areas Cultivated

The allocation Of land to the production of any cereal
crop is believed to be a major determinant Of output
variations across zones. The distribution Of areas
cultivated by crop and by zone is given in Table 5.

The comparison Of areas devoted tO crops grown in both
zones reveals varying patterns according to the crop
involved. Rice fields were much larger in the Delta than in
the Middle Valley while the reverse was true for millet.
Sorghum fields of the Middle Valley also have, on the
average, the same size as those Of millet, 2.5 hectares per
household. Such relatively large sizes are characteristic
Of a "traditional“ agriculture in which increases Of output
are primarily achieved through extensions Of the area
cultivated. Maize and cowpeas, also grown under the
traditional extensive methods, were devoted relatively large

areas, 1 and 1.5 hectares, respectively.

Distribution

48
Table 5

Of Areas Cultivated per Producing

Household by Zone and by CrOp (in hectares).

 

 

Crops
Rice

Millet

Sorghum

Maize

Cowpeas

 

 

 

ZONES
[Erge Perimeters TLarge Perimeters
of the Delta of the Middle Valley

2.5 .8
(383)a (441)
1.9 2.5
(68) (187)

NA ' 2.5
(235)

NA 1

(67)

NA 1.5
(127)

 

aThe number of

cases involved is given in parentheses.

49

Because Of the important areas devoted to cereal
production, it is interesting to analyze the sources Of

power and the levels of technology involved.

C. Power Source and Level of Technology

In the context Of the Fleuve region, three types Of
power sources are possible: human labor, animal traction
and tractor mechanization.

He shall examine these three sources successively.

1. Human Labor

An ideal study Of labor would include quantitative and
qualitative aspects. The quantitative aspects would include
the total number Of mandays available per household and its
distribution among several activities. The qualitative
aspects would distinguish several categories of labor
according to the qualification Of workers for different
productive tasks. In our survey we did not collect such
detailed information. Instead, some global information on
the demographic characteristics Of the different cereals-
growing households was collected. The figures involved are
summarized in Table 6. Even though the aggregate figures on
population are not a perfect indicator the actual avail-
ability of labor within the household, they dO give some
idea of labor availability and are important in determining
what the production per capita will be. Besides human
labor, animal traction constitutes a potential power

SOUI‘CE.

50
Table 6

Average Number Of People per Household Growing”
a Particular CrOp by Zone

 

 

 

 

 

ZONES
£3225 Delta MiddTe VaTTey
Rice 12.2 9.2
Millet 10 10.3
Sorghum NA 7.9
Maize NA 8
Cowpeas NA 6.8

 

51

2. Animal Traction

The use of animal traction in cereal production was
almost inexistent. In fact, it is in this reSpect thatthere
was the greatest homogeneity between the two zones.
Indeed, in none of the 794 cases Of rice production in the
two zones were animals used. It was only in rare instances
that some households reported using animals, mainly for
transportation. These findings are not very surprising in
the context of large perimeters where tractor mechanization
is a dominant characteristic.

3. Tractor Mechanization

The situation in terms Of tractor mechanization varies
considerably depending on the crop involved. As indicated
earlier, rice is grown under irrigation while the other
cereals remain produced under rainfed conditions. The
striking differences that these two types Of agriculture
exhibit are apparent in the structure Of the debts incurred
by various cereals growers and presented in Table 7.

Almost all rice growers Of both zones incurred debts on
the highly mechanized land preparation and irrigation
services. In addition, significant proportions Of rice
growers reported debts on threshing. Those services are
provided by SAED and can be reimbursed either in kind or in
cash. The Situation is different for other cereals for
which only relatively small fractions of growers incurred

debts on fertilizer and insecticides deliveries. It is

52
Table 7

Structure of the Debt in Large Perimeters:
Hhere Debts Here Incurred for Various Activities

by Crop and by Zone.

Percent of Cases

 

 

Crop; and Zones

Delta
Rice
Middle Valley

Delta
Millet
Middle Valley

Delta
Sorghum
Middle Valley

Delta
Maize
Middle Valley

Delta
Cowpeas
Middle Valley

TYPES OF DEBTS

 

Land
Preparation Irrigation Fertilizers Insecticides Threshing
99.1 100 99.7 70.9 35
98.6 98.8 98.8 86.3 66.7
0 O 14 14 O
0 0 9.3 9.3 0
NA NA NA NA NA
0 0 2.4 2.4 0
NA NA NA NA NA
4.2 4.2 4.2 4.2 0
NA NA NA NA NA
5.6 0 0 5.6 0

 

53

noticeable that similar differences appear between rice and
the other crOps with respect to the patterns Of debts
onfertilizers and insecticides. This will be apparent in

the quantities of inputs used.

0. The Other Inputs

Seeds and fertilizers represent the other major inputs
used in cereal production. We shall analyze them
successively.

1. Seeds

Several aspects of seeds are important: origins,
quantities used, quality and prices. We shall analyze the
origins of seeds used in the Fleuve region. To a larger
extent, the quantity, quality and price Of the seeds depend
on where farmers got them.

1.1 Origins of Seeds

An analysis of origins can provide interesting insights
into the structure of cereal production and distribution.
Indeed, such analysis helps to identify bottlenecks in the
distribution System. In addition, it allows one to
understand various transactions taking place in the
production as farmers reimburse their debts. The origins of
the seed used in the large perimeters are shown in Table 8.

Two things are apparent from Table 8. First, rice
growers received their seeds almost exclusively from SAED.

The reliance on SAED is, however, greater in the Delta zone,

54

Table 8

Percentage of Farmers Hho Received Their Seeds From
Various Sources by Crop and by Zone.

 

 

Crops and Zones

 

Delta
Rice
Middle Valley

Delta
Millet
Middle Valley

Delta
Sorghum
Middle Valley

Delta
Maize
Middle Valley

Delta
Cowpeas
Middle Valley

 

 

ORIGINS
Production *Other
Retained Farmers SAED Others

0 O 98 2
4 2 92 1
28 7O 0 2
48 47 O 5
NA -NA NA NA
60 35 0 5
NA NA NA NA
50 43 0 7
NA NA NA NA
40 50 2 8

 

55
where 98% of rice growers received their seeds from the
agency, than in the Middle Valley, where 93% obtained
their seeds from SAED. Second, the growers of other cereals
either kept their own seeds or relied on other farmers to
get them.

Beside the origins, the quantities of seeds, in
relation to the area cultivated, was analyzed.

1.2 Quantity and Quality of Seeds

The seeding densities of the different crops are
presented in Table 9. These densities will be later
compared with those recorded in the small perimeters, as
their differences may explain variations of output across
zones. The percentages of farmers who incurred debts on
insecticides presented earlier in Table 7 constitute an
indication of the quality of the seeds. Indeed, some
insecticides were applied to seeds. In addition, the seeds
provided by SAED are selected varieties that are already
treated.

The seeding densities are much lower for other cereals,
reflecting differences of production techniques with rice,
and differences in input availability. But, the impact Of
the quantity and quality of seeds on production depends also
on the availability Of complementary inputs. Fertilizers,

which belong to this category, will be analyzed next.

56

Table 9

Seeding Rates by CrOp and by Zone
(in kilograms per hectare).

 

 

£212
Rice
Millet
Sorghum
Maize

Cowpeas

 

 

 

ZONES
Delta Zone MiddTe Valley Zone
120.7 122.9
9.2 7.4
NA 9.4
NA 12.2
NA 38.2

 

57
Table 10

Fertilizers Available per Household and Rate
of Fertilization by Crop and Zone

 

 

 

 

 

ZONES

Crogs Delta Middle Valley
Quantities Available 372 179.5
in kg

Rice
Fertilization Rate 145.8 221.2
in kg/ha
Quantities Available 2.97 128.3
in kg

Millet
Fertilization Rate 1.47 35.3
in kg/ha
Quantities Available NA 22.5
in kg

Sorghum
Fertilization Rate NA 16
in kg/ha
Quantities Available NA 74.6
in kg

Maize
Fertilization Rate NA 124.3
in kg/ha
Quantities Available NA 137.8
in,kg

Cowpeas
Fertilization Rate NA 45.3

in kg/ha

 

58

2. Fertilizers

The situation with respect to fertilizers is summarized
in Table 10. Fertilizers also display a dual pattern
between rice and the other cereal crops. Indeed,
relativelyimportant quantities of fertilizers were available
to rice producing households. Those quantities were larger
in the Delta than in the Middle Valley. But, when those
quantities are related to the areas cultivated, higher
fertilization rates are recorded in the latter zone.

Relatively small quantities of fertilizers were used in
the other cereal crops. But, the fertilization rate is
noticeably high in millet in the 'Middle Valley compared to
its level in the Delta zone: 35 kilograms per hectare
versus 1.47. Relatively large quantities of fertilizers
were also applied to maize in the Middle Valley.

The relationships between the inputs used and outputs
will be investigated later. Meanwhile, some general
features of the large perimeters can be summarized.

It is apparent that there are both similarities and
differences between the two zones which have the large
perimeters. The similarities are relative to the high level
of commitment of resources to paddy production, with limited
use of modern inputs on other cereals. In addition, the two
zones present similarities in terms of absence of animal
traction, their debt structure, and patterns of transactions

on rice and other cereal. To a large extent, the common

59
characteristics of the Delta and the Middle Valley stem from
the similarities in their organization and the fact that A
SAED manages them using similar principles and methods.

0n the other hand, significant differences exist
between the two zones. A greater diversification of the
cr0pping pattern is apparent in the Middle Valley, with a
significant representation of several cereals other than
rice. This has a major impact on cereal availability in the
sense that the presence of cereals other than rice more than
makes up for the smaller quantities of rice produced in the
Middle Valley. Differences in yields reflect differences in
location and the corresponding- changes in natural and
environmental factors. Indeed, the quantity of rainfall,
quality of soils, availability of inputs are thought to
play, along with other human factors, a major role in
determining the inter-zone differences. Further
differences, in the type, organization, and structure of the
production system, are expected with the small perimeters,

which will be examined in our next chapter.

CHAPTER FOUR

STRUCTURE OF PRODUCTION AND CEREALS
AVAILABILITY IN SMALL PERIMETERS
The small perimeters covered in this study are located
in the Aere Lao and Matam zones. The Aere Lao zone includes
the perimeters of Ndioum, Pete and Demete. The Matam zone
covers the following perimeters: Matam, Boki Diawe and
Nguidjilene. The two zones share some organizational and
technical features. But, because of differences of location
along the river basin and ethnic differences, each zone has
idiosyncratic characteristics as well. The impact of these
characteristics on performance will be investigated in our
,analysis of production, transactions and cereals

availability in the small perimeters.

A. Production, Transactions and Cereals Availability

1. Production and Yields

The quantities produced of each crop and the
corresponding yields are summarized in Table 11. The
performance varied by crap and by zone. For paddy rice, an
average quantity of 499 kilograms was recorded per producing
household in the Aere Lao zone. Production of paddy per
household was significantly more important in the Matam
zone, where an average of 1203 kilograms was recorded. It
is useful to complement the average figures with the

frequency distribution of paddy production in the small

60

61
TABLE 11
Average Production and Yields in Cereals-Growing

Households in the Small Perimeters by
Zone and by Crop

 

 

AERE LAO ZONE

 

 

 

Standard Standard
Number Production Deviation Yield Deviation
Crops of Cases (In kg) of Production1 (In kg/ha) of Yield1
Paddy Rice 362 499 18 3633 59
Millet 302 971 66 659 64
Sorghum 106 567 46 1013 127
Maize 89 317 . 53 1198 215
Cowpeas 130 198 29 184 32
MATAM
Paddy Rice 221 1203 50 3540 115
Millet 392 576 31 610 64
Sorghum 102 523 41 307 22
Maize 146 438 28 1732 19
Cowpeas 57 237 28 1426 297

 

1The standard deviations reported correSpond to those of the mean
productions and yields. They were computed by dividing the standard
deviations of the households' productions and yields by the square root
of the number of cases. These standard errors can be used to test
whether mean production and yield were significantly different in Aere
Lao and Matam zones.

62
perimeters. Figure 5 presents such a distribution for the
Aere Lao and Matam zones.

Figure 5 indicates that 92 percent of the Aere Lao's
paddy-producing households obtained between 0 and 1 ton of
production. Therefore, only 8 percent of the producing
households recorded quantities of production between 1 and 2
tons. A different situation existed at Matam, where 50
percent of all producing households obtained quantities
higher than 1 ton of paddy. In fact, 37 percent obtained
quantities between 1 and 2 tons and quantities larger than 2
tons were recorded by about 13 percent of the Matam zone's
producing households.

The average yields achieved in both the Aere Lao and
the Matam zones are presented in Table 11. Comparable
average yields were recorded for the two zones: 3633
kilograms per hectare at Aere Lao and 3540 kilograms per
hectare at Matam. It is interesting to note that a yield of
paddy rice comparable to those recorded at Aere Lao and
Matam was also found in the Middle Valley zone (3788
kilograms per hectare). However, the yields were, on the
average, much lower in the Delta zone, with 1938 kilograms
per hectare. The yield differences among the different
zones and within each zone remain to be explained. This
question will be examined in the next chapter. For the
moment, it is useful to analyze the frequency distribution

of the yield of paddy in the small perimeters, presented in

63

Figure 6. The most frequent yields recorded in the Aere
Lao zone were between 4 and 5 tons per hectare. Indeed,
about 25 percent of the paddy producers operating in that
zone achieved yields between 4 and 5 tons per hectare. In
addition, 60 percent of producers recorded yields between 0
and 4 tons per hectare, whereas 15 percent of paddy growers
obtained yields higher than 5 tons per hectare. In the
Matam zone, the most frequent yields were between 3 and 4
tons per hectare, as 24 percent of the producing households
achieved yields situated in this range. On the other hand,
42 percent of producing households recorded yields in the 0
- 3 ton range and in 34 percent. of the cases the yield
achieved in the Matam zone was larger than 4 tons per
hectare.

Besides paddy, other cereals were grown in the small
perimeters. The average production and yields recorded in
each crop were presented in Table 11. Relatively high
yields were recorded on maize in both zones, even though the
productions per household were lower than in the Middle
Valley zone: 317 kilograms at Aere Lao and 438 kilograms at
Matam versus 711 kilograms in the Middle Valley. For
millet, good yields were recorded in both zones, compared to
the results recorded in the large perimeters: 659 kg/ha at
Aere Lao and 610 kg/ha at Matam versus 187 kg/ha and 467
kg/ha for the Delta and Middle Valley zones, respectively.

The yield achieved on sorghum in the Aere Lao zone was the

64

mZON 2532 I
MZON o<._ mama. \\\\.
Dzmomj

WMZON

 

Z

 

m mesmwm

1"f1. 'I.l".0|ll|lvn II. ‘ , I 5-. tlll .I' i.il I villi...|'|llllu

nsz 25 20.53093
01¢ dlfi nlN NIP Flo

L
d

 

 

, R

 

 

rillli

row

iON

ion

Lro*

ion

r00

:05

..oo

1.. OD

 

00p

 

E<F<§ GZ< O<J mmm< MTC.

ZOEUDQOKQ >QQ<Q LO ZOFDEEFWE

SG'IOHBSl'lOH d0 BOVLNHOEBd

 

fit Ii... III al-: I . -nul . L.» ilnal’lln'l.!fi (to; u'II [no]... I, ’1' ail. ‘

 

65

 

—--—--.—~—.—.~ ._..

,—.._--u— 4 «*—.._< .

 

- p..-

 

rm.fl~

ZONES

LEGEND

DISTRIBUTION OF THE 'YIELD OF PADDY RICE

IN THE AERE LAO AND MATAM

V/fl AERE LAO ZONE

- MATAM ZONE

 

 

30

 

~I-

25

  

   
      

l

I
11' I
M“ i
I I
(Or-x?
LLII
, (9%?
\\\\\\\\\\\\\\\\\\\\\\\wg‘
mmmm\mmwm 4 g
. » . _ . _ 42;
.\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\V «I ,9
- .. .. Q. n I
mmmwmmwmw
. ~75?
N\\\\\\\\\W i
1* k\\\\\\\\\\\\\\\\\\\\\\\\\\\\io I

 

_L_
l'

O

SG'IOHBSIIOH :IO 39V1N3083d

51)-

1_
r
0

Figure 6

66
highest of the four zones. However, the lowest yield on
sorghum was recorded at Matam, with only 307 kilograms per
hectare.

It is difficult to isolate systematic patterns among
the other cereals grown in the small perimeters. It seems
clear, however, that important and complex relationships
exist among them. This fact suggests that any evaluation of
food availability in the small perimeters will have to be
based on a study of the whole production system.

Beside production, several types of transactions are
important for cereal availability. He shall analyze them
next.

2. Transactions

The transactions involved are the sales and purchases
of cereals, the gifts and tithes provided and debt
reimbursement by farmers. He shall examine these operations
successively.

2.1 Sales and Purchases of Cereals

The situation of small perimeters with respect to sales
of cereals is summarized in Table 12. The average
quantities of the various crops sold by producing households
to various participants is given. It is apparent from Table
12 that relatively small quantities of cereals were sold by
growers. The largest sales occurred in the Matam zone and
involved rice. Most of it was purchased by SAED. The next

most important sales were of maize and cowpeas in the Matam

67

mo_u.u=a:a cams ozu cu mcmuoc copuap>ou neuucaum mth

 

 

NN em a c mg em a a swan:
momazou
o c a c o o o c on; mgo<
o~ cc o c c" we a c Ewan:
o~_~x
o c c c o o c c on; mgm<
~. _ o o as. N c o scum:
asgmccm
c c c o c c o c on; ocu<
g : c c :8. _ o c .33:
“6...:
_ "a c c mc.N ~_ c o and mcm<
e mm c o N n ma“ cm scan:
66.x
N m.o_ cc. m. N m v.~ m and acm<
:6.» ram c. Nco.uo.>ma a; c. Ncc.aa.>oo ax cp Nco—uu_>oo lax c’ mucoN was mmocu
nuance; mmpu.u=~=o eunucoum ma_u_a:o:c uguucaum moFHNucazc chateaum mo.u—u:azo

Lain

mma<m 4<hoh

mm>~p<¢m¢ccu

mhzmc< mk<>~¢m

ou<m

 

m¢u>=m

 

 

acoN he can acgu an acouoapcom __oam ozu c.
mechaa mzopca> a» u_ogum:o= can mpaocuu No mopam ouaem><

Nu upon»

I

68
zone. There were almost no sales of millet and sorghum.
These findings are consistent with the fact that production
is primarily oriented to subsistence needs.

The small quantities of rice produced by irrigation
were almost exclusively home-consumed in producing
households. 0n the other hand, purchases of cereals took
place. The quantities involved are presented in Table 15.
Relatively large quantities of rice and millet were
purchased by households Operating in the small perimeters of
the Aere Lao zone. In addition to sales, gifts and tithes
were provided by farmers. He shall examine them next.

2.2 Gifts and Tithes

Table 13 presents the quantities provided as gifts by
cereal-growing households in the four zones. The percentage
of growers who provided the religiously prescribed tithe is
also presented in the table. Relatively small quantities of
cereals were involved in the gifts. In addition, no
systematic pattern is apparent across perimeters with
respect to provision of gifts.

The tithe or “dime“ represents an annual tax that
people pay on their productive activities to religious
authorities. It corresponds generally to a fraction of
production and goes to needy people via religious leaders.

The next major source of cereals transactions is

represented by debt reimbursements.

69

Table 13

Average Quantities Provided as Gifts per Household
(in kilograms) and Percentage of Households Hho

Provided the Tithe (in parentheses)

 

 

 

 

- ZONE -
Crop Delta Middle valley Aere Lao Matam
Rice 83 11.9 5.44 12.8
(87) (94.2) (68.2) (97)
Maize NA 2.9 4.1 4.8
NA (76.1) (100) (82.2)
Millet 1.5 9 6.1 5.3
(81) (90) (92.7) (84)
Sorghum NA 12.7 5.1 5.9
NA (97) (92.4) (97.1)
Cowpeas NA 2 .6 1
NA (73.3) (76.9) (0)

 

70

2.3 Debt Reimbursements

The debts were incurred at various stages of rice
production. This can be seen in Table 14. Table 14A
presents the various activities of the production process in
which the debts were incurred. It is apparent that almost
all farmers got a fertilizer credit from SAED. Debts on
insecticides ,were also significant as 25% of farmers
incurred them in the Aere Lao zone and 45% incurred them in
the Matam zone. 0n the other hand, few, if any, farmers had
a debt on land preparation or threshing. This is consistent
with the labor intensity of the .techniques used in small
perimeters. An unanticipated high preportion of farmers
incurred a debt on irrigation services. Since those
services are organized and operated through producers'
organizations, it is difficult to tell what those debts
cover.

Table 14B shows the distribution of the mode of
reimbursement used by farmers in the four zones. He can
notice that the percentages in one zone do not add up to 100
as some farmers used both modes of reimbursement and others
defaulted. It is also apparent that in the Delta and Middle
Valley zone a majority of farmers paid their debts in kind.
The reverse is true in the Aere Lao zone, while a
substantial percentage of Matam farmers did some cash

payment. A similar pattern is apparent in Table 14C.

71
Table 14
The Structure of Debts in Irrigated Perimeters

A. Percentage of Cases in which Debts Here Incurred For
Various Activities of the Small Perimeters

 

 

 

 

ACTIVITIES
[andi Irrigation ’Fertilizer Insecticide ”Threshing
Zones Preparation Services Delivery Delivery Services
Aere Lao 1.2 62.5 100 25 0
Matam .5 0 99 45 0

 

B. Modes of Reimbursement: Percentage of Cases in Hhich
Debts Here Reimbursed in Cash or in Kind by Zone

 

 

 

 

ZONES
Delta Fiddle Podor Hitam
Reimbursements Zone Valley Zone Zone
In Kind 91.1 88.8 20.5 68.2
In Cash 0 27.1 93.1 32.3

 

C. Average Quantities of Rice and Amounts Reimbursed
per Household and by Zone

 

 

 

 

ZONES
Fiddle
Variables Delta Valley Podor Matam
Quantities (kg) 1874 518 12 86

Amount in CFA 0 9517 2560 1235

 

72

Indeed, relatively small quantities of cereals were used for
debt reimbursements in Aere Lao and Matam zones, where some
important cash payments were also made. On the other hand,
no cash payment was involved in the Delta, where important
in kind payments occurred. The Middle Valley seems to
combine the features of the two extremes with substantial
in-kind and cash payments.

Once again, it is difficult to get, from the analysis
of these figures alone, the complete story behind cereals
transactions in the Fleuve perimeters. But, the figures do
make important suggestions about the directions and
magnitudes of these transactions.. It seems clear that most
of these transactions have their origins in the production
process by which several categories of inputs are combined
to generate output. In our next section, we shall analyze
how production and transactions contributed to cereals
availability in small perimeters.

3. Indices of Cereal Availability

The results of our survey are summarized in Table 15.
The per capita availability of each crop to the households
which grew it is described for each small perimeter's zone.
Two important remarks need to be made with respect to cereal
availability in small perimeters.

First, no account has been made for the losses of
production that occur during the post-harvest operations,

because of our lack of information on their magnitude in the

73

Table 15

Indices of Cereals Availability in Producing Households
Operating in Small Perimeters by Crop and by Zone

 

 

 

Ratio 3a
Purchases of Total Cereal
of Cereals Availability
Crops and Zones Ratio 1a Ratio 2a Per Capita Per Capita
Aere Lao 42 40 36 76
(N = 362)
Rice
Matam 85 76 2 78
(N = 221)
Aere Lao 82 80 29 109
(N = 302)
Millet
Matam 48 47 4 51
(N a 292)
Aere Lao 51 50 6 56
(N = 106)
Sorghum
Matam 59 58 1 59
(N = 29)
Aere Lao 64 64 1 65
(N = 29)
Maize
Matam 42 37 3 40
(N a 438)
Aere Lao 36 34 .5 .5
(N = 130)
Cowpeas
Matam 12.7 7.8 .1 7.9
(N a 237)

 

aSee Table 4 for definitions of these ratios.

74
different zones. For this reason, the figures presented
in Table 15 exaggerates somewhat the availability of cereals
in the different zones. But, since the same principles were
used in all the zones covered in this study, the relative
distribution of cereals availability is not affected.

Second, as implied by number of the cases presented in
Table 15, few households grew all the cereals crops in the
small perimeters. In fact, the same thing was true for the
households Operating in large perimeters.

In general, comparable levels of cereals availability
were achieved by producers of the different crops in the
small perimeters. These levels were significantly less
important than those recorded by producers of the same crops
operating in large perimeters. It is important to mention
that if some households raised all the cereal crops included
in this study, they would have exceeded the Ministry of
Agriculture's target. Indeed, if such hypothetical
households were operating in the small perimeters, they
would have had a per capita availability of cereals equal to
236 kilograms in the Matam zone and 340 kilograms in the
Aere Lao zone. The per capita availability of cereals would
have been equal to 512 and 257 kilograms in the Middle
Valley and the Delta, reSpectively.

However, given the number of households involved in the
production of the different cereals crops, it was very

unlikely that many households grew all the cereal crops.

75
Therefore, it is very unlikely that many of the households
involved in cereal production in the small perimeters
achieved the levels of cereal availability targeted by the
Ministry of Agriculture. In our next section, we shall
analyze the pattern of input allocations to cereal

production in the small perimeters.

8. Areas Cultivated

The areas allocated to different crops in the two zones
covered by the small perimeters are shown in Table 16. The
most noticeable feature about areas cultivated has to do
with the tiny sizes of the plots of rice and maize in both
zones. Indeed, the size of those plots noticeably differs
from the dimensions areas allocated to the other cereals
grown in the two zones, and characterizes the intensive
irrigation practiced in the small perimeters. They also
represent the most obvious difference with large perimeters
as they correspond to higher yields but result also in lower
production per household. Besides the rainfed and irrigated
activities, a flood-recession agriculture is also practiced
in the small perimeters. As a result, the production system
is even more complex.

The relatively large size of areas allocated to millet
and sorghum reflects the fact that those crops were grown
under rainfed conditions. 0n the other hand, maize and
cowpeas were grown either under irrigation or in flood

recession.

76

Table 16

Distribution of Areas Cultivated Per Household on

The Small Perimeters by Zone and by Crop
(in hectares)

 

 

 

Aere Lao Zone

 

Crops Matam Zone
Rice .23 .35

(N = 362) (N = 221)
Millet 2.0 3.3

(N . 302) (N = 292)
Sorghum 1.5 1.8

(N = 106) (N = 102)
Maize .47 .30

(N a 29) (N = 146)
Cowpeas 2.2 .32

(N = 130) (N = 57)

 

77
The characteristics of each production subsystem will
be also apparent in the patterns of the allocation of other

inputs to the various cereals.

C. Other Inputs

The other inputs include the power sources, represented
by labor and animal traction: seeds, and fertilizers.

1. Labor and Animal Traction

As in the large perimeters, no detailed information was
collected on the availability of labor in the small
perimeters. The average numbers of people living in cereal-
growing households are presented in Table 17. In the
absence of more specific information, the figures of Table
17 constitute the only indicators of labor availability in
the small perimeters.
2. Seeds and Fertilizers

The origins and quantities of inputs used represent the
two most relevant questions in the small perimeters. The
analysis done earlier in Table 14A suggested that almost all
farmers had incurred a debt on fertilizer delivery from
SAED. Therefore, what remains to be discovered are the
origins of the seeds used and levels of fertilizer use.

2.1 Origins of Seeds

The information relative to seed origins is given in
Table 18. Two points are particularly apparent in Table
18. First, a smaller proportion of rice growers relied on

SAED for their seeds than in large perimeters. Indeed, a

78
Table 17

Average Number of Pe0ple Living in Cereals-Growing
Households of the Small Perimeters by Zone

 

 

 

 

CROPS
Zones Rice Millet Sorghum Maize Cowpeas
Aere Lao 10 9 7.3 * 8.5
(N=362) (N=302) (N=106) (N=130)
Matam 11 8.7 10.6 11 9.8
(N=221) (N=292) (N=102) (N=29) (N=57)

 

*Not reported because of the large number of missing
cases.

79

significant percentage of farmers kept their own seeds from
their previous production, especially in Matam. Second,
the seeds of other cereals came almost exclusively from
production retained or from other farmers. This was also
true in the large perimeters.

Beside their origins, the quantities of inputs used are
relevant to our analysis.

2.2 Seeding Densities and Fertilization Rates

The quantities of seeds and fertilizer used per unit of
area are summarized in Table 18. Important differences
exist on the seeding density and the fertilization rate of
given crops across perimeters and zones. For instance, the
seeding densities of rice are relatively small for both Aere
Lao and Matam, (87.7 kg/ha and 29.6 kg/ha, reSpectively)
versus 120.7 kg/ha for the Delta and 122.9 kg/ha for the
Middle Valley. Those differences could be partially
attributed to differences in the input endowment of the two
types of perimeters. But, they also reflect differences in
production techniques. While the large perimeters use drill
seeding, tranSplanting is performed in the small
perimeters. The nurseries used require smaller quantities
of seeds.

For the other cereals, the relatively low densities
recorded reflect either their extensive methods of

production or scarcity of seeds or, more probably, both. It

80

Table 18

Origins of Seeds in the Small Perimeters. Percentage
of Farmers Hho Received Their Seeds From Various
Sources by Crop and by Zone

 

 

 

 

CATEGORIES
Production *Other

Crops and Zones Retained Farmers SAED Others

Aere Lao 18 1.5 73 7.5
Rice

Matam 41.5 1.8 55.7 1

Aere Lao 84 15 0 I
Millet .

Matam 100 0 0 0

Aere Lao 91 9 0 O
Sorghum

Matam 100 O 0 0

Aere Lao 70 30 0 0
Maize

Matam 96.5 0 3.5 O

Aere Lao 99 O 0 1
Cowpeas

Matam 100 0 0 0

 

81
Table I9

Seeding and Fertilization Rates in Small Perimeters,
by Crap and by Zone (in kilograms per hectare)

 

 

 

 

 

 

ZONES
AERE LAO MATAM
Seeding Fertilization Seeding Fertilization
Crop Density Rate Density Rate
Rice 87.7 230 29.60 426
Millet 7.1 O 9.10 0
Sorghum 9.4 0 3.90 O
Maize 18 O 32.18 124.6
Cowpeas 2.2 O 17.90 15

 

82
is not, however, correct to compare seeding densities
across craps because of differences in their vegetative
characteristics and requirements.

Hith respect to the fertilization rates, the quantities
recorded on rice are larger than those found in the large
perimeters. The highest fertilization rate was found at
Matam, with 426 kilograms per hectare. Incidentally, the
highest yield was not recorded at Matam but in the Middle
Valley zone, where it averaged at 3778 kilograms per
hectare. The second highest yield was found at Aere Lao
with 3633 kilograms per hectare. The fertilization rates in
these two zones were equal to 221 and 230 kilograms per
hectare respectively.

Beside seeds and fertilizers, other categories of
inputs relative to the operation of the irrigation pumps
and other aspects of cereal production could have been used
in the small perimeters. However, no data were collected on
them during our survey.

In summary, the structure of production in small
perimeters displays interesting features in terms of the
allocation of resources among the various craps. In many
instances, significant differences appeared between large
and small perimeters regarding the relative importance of
different cereals and their performance. Differences also
appeared between the two zones which had the small

perimeters. Even though those differences remain to be

83
explained, we suspect that they have to do with Specific
determinants highly correlated with location.

One thing seems to be true for all zones: that is the
interdependency and complementarity of the various
components of the production system. In the limit, we can
say that it is impossible to understand any component of the
system in isolation.

Because of the aggregation involved in the computation
of our performance variables and given their large
variability among households, it seems indicated to
undertake our comparative analysis at more disaggregated
level. Indeed, the figures uSed in the analysis, as
averages, do not reveal the large internal variability
existing within each perimeter of a given zone. For this
reason any causal relationship derived from such aggregated
figures could be misleading.

However, the global figures allow one to understand the
general characteristics of each type of perimeter. The
information provided can then guide further inquiries on the
determinants of production and cereal availability
undertaken at lower levels of aggregation.

The next chapter of the thesis will examine the

determinants of rice products in the Fleuve region.

CHAPTER FIVE

THE DETERMINANTS OF RICE PRODUCTION
IN THE FLEUVE REGION

The analysis undertaken in Chapters Three and Four has
revealed the important contribution of farmers' own
production to the per capita cereal availability both in
large and small perimeters. Indeed, despite the importance
of purchases, the bulk of farmers' consumption requirements
is covered by their own production. Therefore, the
production performance of the Fleuve region rice growers
constitutes a major determinant of their food security.

In addition, Senegalese government officials place a
high value on rice production in the Fleuve region, with the
ultimate goal of reducing the large quantities of this
commodity imported every year. For this reason, important
resources are being injected into the development of
irrigation schemes. Yet, there is a lot to be learned about
the ways resources contribute to policy objectives. In
fact, until recently, a major performance criterion used by
SAED was the number of hectares developed. Moreover,
increases in the size of current small perimeters and/or
their multiplication are perceived by many as the way to
guarantee the food security of the Fleuve region's farmers.
However, because of the hypothesized correspondence between
the scale of operation and farmers' management capacities,
it seems necessary to move with caution.

84

85

From a policy point of view, it is of critical
importance to provide a framework that allows one to assess
the impact of alternative actions on rice production.
However, it is important to mention that despite the fact
that the focus is put here on irrigated rice, the other
components of the production system must be included in an
overall assessment of Senegalese agricultural policies in
the Fleuve region. As we shall see below, the other
components of the Fleuve region's diverse agricultural
system affect rice production.

The objective of this chapter is to present the
theoretical and empirical foundations of the econometric
analysis developed in this study. The chapter provides a
set of hypotheses about which factors determine the
performance of the rice production systems in the Fleuve
region. The hypotheses reflect various opinions held by
many researchers and government officials on the character-
istics of the perimeters and the impact of these on
production. They are also partially inspired by the results
of the descriptive analysis presented in earlier chapters.

For analytical purposes, we have classified the
elements believed to affect production into several
categories: the physical and technical factors, the
economic and policy elements, and the institutional and
sociological determinants. He shall examine each category

in a separate section.

86

A. The Physical and Technical Determinants

The physical and technical elements that influence rice
production in the Fleuve region are numerous and could not
possibly be all mentioned here. On the other hand, it is
possible to isolate some important characteristics of
perimeters that can help to explain variations of production
both between the two types of schemes and within a given

scheme.

1. The Physical Factors

The physical factors include the effects of climatic
and environmental characteristics on rice production. In
this respect, we know that the Fleuve region involves a lot
of diversity, in terms of climate, soil characteristics,
land and water availability and suitability for irrigation.1
Such diversity in the environment is believed to be a
primary determinant of productivity in the region. Indeed,
we expect perimeters with the same dimensions and similar
organization to have different performance because of
differences of location in the Fleuve region. For instance,
the problems of salt water intrusion experienced in the
Delta put some constraints on how much rice can be produced
in this zone, compared to the large perimeters of the Middle

Valley. Similarly, the small perimeters of the Middle

 

1See O.M.V.S.: "Etude Socio-Ecopnomique du Bassin du
Fleuve. Partie A. Presentation Generale du Bassin du
Fleuve.“ April 1980.

87

Valley zone are expected to have different performance than
those Operating in the Matam zone deSpite their similarities
in terms of size and management. Even within a given
perimeter we cannot expect homogeneity with respect to the
quality Of soils and other physical elements. Therefore,
when combined with other factors, those differences help to
explain the dispersion of yields recorded in different plots

of a perimeter.

2. Technical Factors
These include the production techniques used, the size
of irrigated plots, the existence of appropriate inputs and
varieties, the type of irrigation equipment, etc. Tuluy
(1978) has made a classification of the various techniques
Of rice production in Senegal by combining several
criteria: irrigation method, the number Of harvests per
year, and the degree Of labor and capital intensity. These
criteria have allowed him to establish a distinction among
perimeters in the Fleuve region:
1. Pump irrigation, single crop, mechanized
technique, typified by Boundoum in the
Delta.
2. Pump irrigation, two crops, mechanized
technique, corresponding to Nianga of

the Middle Valley.

88

3. Pump irrigation, double cropping and

manual techniques as typified by
Matam (Tuluy, 1978, pp. 7-12).

Such a classification outlines differences in
production techniques. For instance, direct seeding is
widely used in many large perimeters, while transplanting of
young plants from nurseries is practiced in some large
perimeters (Nianga and Guede) and almost all small
perimeters (C.C.C.E., 1982. p. 21). Each technique has its
own requirements in terms of labor. Indeed, drill seeding
allows reducing the labor needed to about 6 to 8 mandays
versus 78 for tranSplanting. However, the seed drilling
technique implies more difficult weeding, as the risk of
infestation by weeds is increased (Bonnefond et. al., 1981,
p. 32). Consequently, unless herbicides are used, the labor
requirement increases for weeding. On the other hand, yield
differences were recorded as a result of seeding versus
transplanting.

The size of individual plots constitutes a major source
of output variations not only between large and small
perimeters, but also among farmers Of a given perimeter.
According to Fresson, three factors are considered by SAED
when fixing the size of family plots:

1. The desire to extend the benefits of

irrigation to a maximum number of

89
families even if this means that each
family will get only a tiny plot.
2. The size Of households which include on
the average 10 peOple, among whom four
are active workers.
3. The time available to farmers, who
combine irrigation with the rainfed and
flood-recession agriculture. In this
respect, Fresson indicates that SAED
does not intend for irrigated rice
production to substitute for the
traditional farming but to complement
it.2
Besides the size Of the area, the availability Of
modern inputs is believed to affect quantities produced. In
the Fleuve region, the major inputs besides land are labor,
fertilizers and seeds. Differences in labor availability
among producing households are hypothesized to partially
explain differences in production. Labor availability
includes both quantitative and qualitative aspects. But
quality distinctions could not be represented in our
analysis because of the unavailability Of the needed data.
For fertilizers, several elements determine their

effectiveness in increasing yield. Indeed, their impact

 

2A critical discussion Of these points is made in
Fresson, Public Participation, p. 35.

 

9O

depends not only on the quantities but also on the
availability of complementary inputs such as fertilizer—
reSponsive varieties, adequate water control, insecticides
and appropriate cultural practices (Zalla, Diamond, and
Mudahar, 1977, p. 4). With respect to seeds, there is a
need to account for differences in the seeding techniques
used in the different perimeters.

The technical elements that have been mentioned in this
section constitute a sample of a more complex set of
technological and economic factors that affect the variation
in rice production both among perimeters and among producers
Operating within any given perimeter. But physical and
technical factors, alone, do not fully explain variations in

production. Some economic reasons need to be considered.

8. Economic and Policy Factors

The economic and policy factors relate to the various
elements that motivate or dissuade different participants to
produce rice in the Fleuve region. From a macro-economic
point of view, a crucial question relates to the comparative
advantage or disadvantage of Senegal in rice production.
Given the local production costs and import prices, the
question is whether it is profitable for the country to grow
rice instead Of importing it. Despite the complexity of the

question and the methodological problems associated with

91
cost-benefit analysis in the Senegalese_ setting3, the
studies undertaken yield similar conclusions. That is, as
Tuluy puts it ". . . except for the animal traction, gray
soil rice production from Casamance, none of the domestic
production techniques is competitive with imports in Dakar.“
(Tuluy, 1978, p. 32)

The same studies found that under the current
technologies and given factor availabilities and prices,
labor intensive and intermediate mechanization were superior
to large scale, capital intensive projects in meeting the
national Objective Of increased rice production. Moreover,
the studies seem to agree on the fact that because of the
protection due to tranSportation and handling costs,
substitution of imports by local production could profitably
take place within producing regions. I

From farmers' point Of view, given the subsidized costs
of irrigation services and inputs, producing rice for their
own consumption constitutes the cheapest way to supply
themselves with this commodity. Indeed, in the 1980 season,
the official price of paddy was equal to 41.5 CFA francs per
kilogram while the private production costs in small
perimeters were estimated at 7.8 CFA per kilogram and the

consumer rice of rice was equal to 80 CFA francs. Given

 

3See Makhone MBaye, “Implementation and Evaluation of
Rural Development Projects Under Uncertainty: With a
Special Reference to the Sedhiou II Project (Casamance-
Senegal)." (Plan B paper, MSU, 1983) for a discussion of
some of these problems.

92

those prices and considering an extraction conversion rate
Of 67% for paddy, farmers save 30 CFA francs/kg by keeping
their consumption and paying their debts in cash. Besides
the producers' price of rice, farmers' decisions are
influenced by such macro policy variables as the country's
exchange rate and the import price of rice. These macro-
policy variables contribute to shape the relative
attractiveness of several crops. This constitutes an
additional justification for a greater consistency between
agricultural policies and the overall macro-economic
objectives.

Related to the macro-policy variables, the relative
profitability of different types of agriculture constitutes
an important determinant of rice production, mainly in small
perimeters. In this respect, the reduced risk and the
possibility of forecasting that irrigation provides gives it
an advantage over rainfed agriculture when there is a
competition for labor during the winter season. But a
different situation prevails during the off-season, when
there is a competition between irrigation and the flood-
recession agriculture. Indeed, as the flood water starts
to recede, farmers are able to predict the output that they
can get from their flood-recession fields. Production is
the product of the yield achieved times the area
cultivated. Hhile yields are higher in irrigated

perimeters, many farmers give priority to their recession

93

fields whose areas per household are about ten times larger
than those of the irrigated plots (Fresson, 1978,
pp. 35-36).

Besides the competition between different types of
agriculture, there is no agreement between SAED and farmers
about the choice of an Off-season crop in small perimeters.
SAED has attempted to introduce wheat as a second crop in
irrigated perimeters. The choice of this crop was partially
motivated by the desire to meet the Senegalese government's
Objective of reducing the imports Of wheat and ultimately
eliminating them. But wheat also presents advantages as an
Off-season crop. First, wheat. has a relatively short-
growing cycle (90 to 110 days) and can, therefore, reduce
irrigation costs. Growing wheat as a second crop would also
lessen the competition between irrigation and other
agricultural activities. Second, wheat seems to be well
suited to the ecology of the region. But, farmers do not
seem to be attracted by wheat, whose food potential and
prOper growing methods are not well known. In addition, the
selling price of wheat is relatively unrewarding for many
farmers (Fresson, 1978, pp. 32-33).

On the other hand, many farmers favor rice as an
off-season crop. But, SAED opposes this option because two
consecutive crops Of rice during the rainy and dry seasons
would overlap due to the length of the growing cycle. In

addition, growing rice during the hot season would involve a

94
lengthening Of the growing cycle. This would result in an
increased demand for water at a time Of the year when the
river level is very low. Therefore, the pumping costs borne
by farmers would also increase.

It appears that the differences between SAED and the
farmers reflects the fact that each faces a different set Of
implicit or explicit costs. In this respect, we can
hypothesize that farmers' preference for rice reflects the
fact that they pay a highly subsidized price for water.
Therefore, the price of water can be an important policy
variable that policy-makers could manipulate to influence
the volume and mix of production.-

The increased activity of various types of pests during
the dry season and the more extensive leaching of the soil
resulting from a double rice crop constitute additional
reasons invoked by SAED for opposing double cropping of
rice.

These factors, and many others, affect farmers'
motivations and ability to produce rice. Indeed, despite
its policy of non-interference in the village management,
SAED still plays an active role. For instance, SAED's
Objective is to charge each producers' group the real costs
of small perimeters. But in fact, the groups do not
currently bear all the costs generated by irrigation.
Indeed, SAED ‘provides some services at no charge (delivery

of insecticides) and credits farmers for decreases in

95
production due to calamities or resulting from its own
actions. (C.C.C.E., 1982, p. 94).

Related to this point is the question of the
replacement of the irrigation pumps. The first Sets were
provided free to farmers, who are expected to take in charge
their renewal. Suggestions that migrants' remittances be
used in order to finance the pumps were not well accepted by
many farmers. Yet, the financial analysis undertaken by the
”Caisse Centrale de COOperation Economique“ (C.C.C.E.. 1982)
revealed that farmers provide a monetary subsidy to
irrigated rice. In fact, this study found that the system
could not operate without the income from other activities.
By permitting farmers to partially pay for the charges
emanating from irrigation, the migrants' remittances and
other Off-farm incomes allow them to keep most of their
production for their own consumption.

These economic and financial elements, as they
contribute to shape the incentives Of the participants of
the agricultural system, represent major determinants Of
rice production. When they are combined with other
institutional and sociological factors, the same elements
are hypothesized to explain variations in production between
types Of perimeters and among producers within a given

perimeter.

96

In our next section, we shall examine some
institutional and sociological factors believed to affect
the performance of different perimeters.

C. Some Sociologicaland Institutional Determinants
of Rice Production in the Fleuve Region

Case studies of irrigated perimeters have revealed
particular features conducive to the success or failure of
different projects (Patterson, 1984; Casey, 1984; Fresson,
1978; Diemer, et. al., 1983; and CCCE, 1982). In
particular, these studies have isolated some institutional
and sociological factors that are believed to affect
performance.

An important characteristic of the small perimeter is
its integration into the framework of existing social
structures. Such an approach has allowed utilization of
existing village institutions by gearing the perimeters to a
social and economic unit of space: the village or ward.
These features are reinforced by the fact that some
management functions are delegated to farmers' associations,
resulting in a greater participation of farmers. This has
facilitated the development of collective work. Indeed,
volunteer farmers undertake all the work manually and share
the different costs generated by their activities. But the
positive effects of farmers' participation were attributed

to the following:

97

--As local residents, farmers know the rules,
constraints and balance of power that prevail in their
village. Therefore, they are able to devise a stable
organizational system, suitable to their socio-economic
context.

--Because they are appointed by their groups, farmers'
representatives who stand for traditional authority
and/or enjoy the esteem and trust of their fellow
citizens, get the consensus necessary for winning
acceptance Of their decisions and ruling.

These elements justify the critical necessity of

securing farmer' participation in the planning stage Of a

project. For this purpose, some basic conditions must be

met:

--First, the project's aim must be compatible with
those of the socio-economic system. In relation to
this point, it is important to remember that the
introduction of village irrigation systems was
motivated by the desire to lessen the dramatic impacts
Of the food crisis which followed the 1970's droughts.
The rapid multiplication of the small perimeters is
generally credited to the reinforcements provided to
farmers by the climatic calamities.

--Second, the production models of the project must be
flexible enough to allow their assimilation by the

targeted social system.

98

--Finally, delegation Of some decision-making powers

should be achieved through decentralized methods and

noncompulsory Operating procedures (Fresson, 1978,

pp. 5-6).

Even though these conditions contrast sharply with the
centralized management methods used by SAED in large
perimeters, the relationships between the conditions for a
successful Operation and the size Of perimeters need to be
specified. Analyzing the characteristics Of the village
systems conducive for develOpment, Patterson emphasizes,
among other things, their smallness and their flexibility in
terms of innovation (Patterson, 1984, p. 51). In fact,
bigger plots and memberships are hypothesized to multiply
the difficulties Of organization and management, making
conflicts more likely and more frequent. This point implies
that rather than size per se, the homogeneity of the
producer group constitutes the critical variable.

But, the producer groups do not always constitute
homogenous entities with respect to kinship, caste or
political view. In fact, most Of the small perimeters are
located in areas dominated by the highly stratified
Halpulaar ethnic group. While further studies are needed on
the subject, there seem to be some close relationships
between a person's caste and his participation in the

benefits of an irrigation project.

99

Fresson (1978, p. 47) for example, indicates that the
disputes which arose in the early phases of some perimeters
she studied involved farmers of different ethnic groups.
Hhile Patterson (1984, p. 73) could not find a direct
relationship between the noble caste status and the
financial well-being of participants, he did emphasize the
traditional concentration of power and wealth in the hands
of a few notables.

These points reinforce the need to evaluate the
irrigation systems within the socio-cultural and historical
context Of the Valley. A significant diversity
characterizes the different zohes Of our study in this
respect. This diversity, combined with the relative youth
of the village irrigation system, explains the difficulty of
making generalizations about the institutional and
sociological determinants of performance.

DeSpite such difficulties, our econometric analysis
will attempt to isolate some common characteristics of the
different irrigation schemes. The discovered features will
help us to gain additional insights into the design and
operation Of future perimeters. Before starting the
analysis itself, it is useful to briefly present some
elements Of the methodology used in the sixth chapter of

this study.

100
0. Elements of Methodology

In this section, we briefly present the estimation
procedures used in the following analysis before discussing
the use Of dummy variables to represent categorical
variables.

All equations are estimated with the method Of ordinary
least squares (OLSQ), which provides, when its assumptions
are met, the best linear unbiased estimates Of the equation
parameters. He shall initially assume that the assumptions
of the ordinary least squares are met here. Later, we
shall critically analyze what happens when these assumptions
are relaxed to account for some problems in our data.

Our analysis will attempt to capture the impact on
cereal production in the Fleuve region of phenomena like the
type Of perimeter (large or small), location Of farmers in a
particular zone and organizational factors. The standard
way of representing such qualitative elements is to use
dummy variables. This is generally done by "dropping out“
one category from each class of dummy variables and,
therefore, having it represented by the intercept. In fact,
this procedure makes the categories "left out“ the standard
from which the classes represented in the equation are
allowed to deviate.

D.B. Suits (1984, pp. 175-180) showed how, in cases
where a set of dummy variables is used to measure the

variation in behavior among several classes, the

101

presentation of results in appropriate forms and their
interpretation differ from the purely mechanical problem of
fitting the regression equation. He indicated a way to
transform the regression coefficients in order to overcome
these difficulties. The transformation, in essence, derives
from the fact that we can add a constant to each of the
estimated dummy variables' coefficients and subtract the
constant from the estimated intercept; the new set of
coefficients have statistical properties identical to those
initially estimated. In fact, the standard way of leaving
one category out constitutes a particular case of the
principle, as the coefficient Of the variable left out is
equal to zero, which is also subtracted from the constant
term. However, the same principle allows us to select our
constant in a way that facilitates the interpretation and
presentation of our results.

We shall use Suits' transformation procedures in our
analysis. But, as implied by the method itself, we need to
estimate first the parameters of our equations in the usual
way. In the next chapter, we shall present the econometric

model as well as the results of the estimation procedures.

CHAPTER SIX
THE ECONOMETRIC MODEL

A. Introduction

The objective Of this chapter is to develop an
econometric model which approximates the quantitative
relationships involved in the Fleuve region's rice
production systems. More Specifically, the model tries to
identify and isolate the variables that are relevant in
explaining variations Of rice production in the Fleuve
region. The analysis presented in Chapter Five dealt with
various hypotheses regarding the determinants of
production. The econometric model will incorporate some of
these hypotheses and will test them with the cross-sectional
data from the 1982/83 season.

The model involves three equations:

--The first equation describes the area allocated to

paddy rice by producing households of the Fleuve

region. The equation investigates the relationships

between area cultivated and relevant variables, such as

the number of active workers in the household, the

location in a particular zone and the participation in

a specific producer group.

--The second equation describes the distribution of

yields in relation to the variation of the major

inputs: fertilizers, seeds and number of active

102

103

workers. In addition, dummy variables were included

to represent the effects of categorical variables:

location, organizational factors, etc.

--Finally, the production equation was derived as an

identity given by the product of area cultivated times

the yield achieved.

Several specifications were tried for both the area and
yield equations. The most "satisfactory“ equations were
selected to derive the production equation. The following
criteria were used in the selection:

--The various hypotheses discussed in the previous
chapters provide some guidance on the types of
relationships that can be expected. In addition,
theory and empiricism give some indication on the
functional form of the phenomena studied. For
instance, the usual theories Of production functions
suggest nonlinear relationships between output and the
various inputs used. Both the selection Of variables
and the specification of the equations were inspired by
those considerations.

--The value Of the adjusted R2 was another criterion

used to choose among the several equations. Equations

with higher adjusted coefficients of determination
explain more of the dependent variable's variance.

Everything else being equal, we considered the

equations with the highest R2 to be more satisfactory.

104
--Finally, the equations more useful for policy
purposes were preferred. For instance, the intro-
duction Of dummy variables for producer groups would be
a way to make some allowance for our specific ignorance
Of localized factors which affect the yield of paddy
rice. Indeed, changes in the intercept of the yield
equation signifies the existence of those factors
in the concerned group. But this does not give any
Specific information on the nature Of the localized
factors. A more useful Option is to capture the
effects Of the localized factors on the efficiency Of
major inputs. Equations allOwing for changes in the
elasticity of yield with respect to fertilizers and
seeds would be a way to get more useful information
from our ignorance.
The combination Of these criteria allowed us to select
the most satisfactory area and yield equations. In our next
section, we Shall present the economic and statistical

models along with the results of the estimation procedure.

8. Presentation Of the Models

The econometric model postulates the behavioral
relationships that characterize rice production in the
Fleuve region. These postulated behavioral relationships
will be quantified in the statistical model.

1. The Economic Model

1. AREA = f (number of active workers, location)

105

2. YIELD = f (Number of active workers, fertilizers,
seeds, location)

3. PRODUCTION = AREA x YIELD (Identity)

Our production equation has been disaggregated into an
area and a yield equation. Such an approach has been
justified in the literature- on the basis of the biological
nature Of the production process in agriculture, the time
lags involved between planting and harvest and the generally
extensive use of land and climate (Houch et. al., 1976;
Askari et. al., 1977; Nerlove, 1956). Indeed, the area
allocated to different crops is thought to be determined by
their relative profitability. 0n the other hand, yields are
generally considered to be determined by technological
factors, weather and other more or less economic
influences. In the case of the Fleuve region, we hypo-
thesize that the distribution Of the area cultivated in
paddy rice reflect the attractiveness Of irrigated rice,
compared to the other forms of agriculture. In addition, we
consider technical and managerial elements to be major
determinants Of yield differences among producing
households.

Equation 1 describes the variations in the area
allocated to paddy rice in the different households as a
function of variations in the households' endowments Of
active workers and the households' location. The:
relationship between area cultivated and number Of active

workers is based on the fact that SAED determines the Size

106

of each household's plot in relation to the availability of
workers in the household. However, because of the physical
limitation in the amount of irrigated land available
relative to the amounts requested by farmers, mainly in the
small perimeters, we do not expect a linear relationship
between area cultivated in paddy rice and number Of active
workers.

The locational variable accounts, among other things,
for the variations in the availability Of land suitable for
irrigation. Indeed, more irrigation land is available in
the Delta and Middle Valley zone, in which the large
perimeters are found, than in the other two zones.
Moreover, similar differences are expected to exist both
among perimeters Of a given type and among producer groups
Operating in a particular perimeter. In the statistical
model presented below, locational variables are introduced
to represent both zones and producer groups. The intro-
duction of locational variables representing the different
producer groups is designed to capture localized factors
related to those differences. Besides differences in the
physical availability of land, the two categories of
locational variables could represent the effects of other
economic, policy and sociological factors, to the extent
that these factors are highly correlated with zones and
producer groups. In the analysis of the determinants of

rice production that we conducted in Chapter Five, we Showed

107
how several economic, policy and sociological factors could
affect rice production. However, we could not explicity
incorporate most of these elements in our econometric
model. Nevertheless, some of these factors which are
location-specific will be picked up by the coefficients of
the locational variables.

Equation 2 postulates a relationship between the yield
of paddy rice and both the use Of different inputs in
producing households and specific factors related to the
location in a given zone. The inputs involved are the
number Of active workers in producing households, the
fertilization rate and the seeding density used in each
household. We expect a positive relationship between the
households' endowments in active workers and the yields
achieved by different producers. Our expectation is based
on the hypothesized positive link between labor-intensive
activities such as transplanting and manual weeding and the
yield of paddy rice achieved by producers. In addition,
most farmers combine irrigation with rainfed and flood-
recession agriculture. As several activities Of the three
subsystems overlap, we expect households with higher
endowments Of active workers, everything else being equal,
to achieve a better performance.

In addition to the number of active workers, the
fertilization rate and the seeding density used in the

different households are believed to affect the yields Of

108

paddy rice. More specifically, everything else being
equal, the households which used the highest fertilization
rates and seeding density are expected to achieve the
highest yields. However, as explained earlier, the
effectiveness of fertilizers depends not only on the
quantities used but also on the availability of adequate
complementary inputs including fertilizer-reSponsive
varieties, adequate water control and appropriate cultural
practices. The different producer groups are hypothesized to
be heterogenous with respect to the availability Of such
complementary inputs. Therefore, in the statistical model,
interaction terms between the fertilization rate and the
household's participation in a given producer group were
introduced into the yield equation.

In addition, our two types Of perimeters present
differences in seeding techniques. Indeed, in large
perimeters direct seeding is the most common technique used,
whereas farmers Operating in small perimeters generally
perform the transplanting Of seedlings initially grown in
nurseries. It seems, therefore, interesting to test whether
significant differences in the responsiveness of yield to
seeds were recorded in the two types of perimeters as a
result Of their different seeding techniques. For this
purpose, an interaction term between the seeding density and
the household's Operation in a large perimeter was included

in the statistical model's yield equation. The statistical

109
significance of our two categories Of interaction terms
will allow us to isolate the effects of localized factors
and a particular seeding technique on the effectiveness Of
fertilizers and seeds in increasing the yield of paddy
rice. Locational variables representing the different zones
of the Fleuve region are also included. Factors related to
the type of soils, the climate and other natural and
agronomic elements are hypothesized to be location
specific. The effects of these factors .On the yield of
paddy rice could be picked up by the coefficients of the

locational variables.

2. The Statistical Model

Area Equation

3 55
1. LSUPi = a0 + alLMODMi + jgicjpzj + EgldeGk + U,

Yield Equation

2. LRDMTi = b0 + blLENGUi + bZLSENU, + b3LMODMi +

3 55 2
3:1“ij + Elgkmieaik + 5181mm“ + v,

Production Equation

3. LPRODi = LSUPi + LRDMTi + N,

where:

LSUPi = Natural logarithm of the area allocated to
rice in household i

LMODMi

DZ

OGk

LRDMTi

LENGU1

LSEMUi

INTG3ik

INTG4ik

110

Natural logarithm of the number of active
workers in household i

1 =19 2’ o 0 0,1366

Dummy variable representing the different
zones

DZj = 1 if the household is located in
zone j
DZj = 0 if the household is located in a

different zone
j = 1, . . . 3

Dummy variable representing the different
producer groups

06k = 1 if the household belongs to
producer group k
06k = 0 if the household belongs to

a different producer group
k = 1, 2. . . . , 55

Natural logarithm of the yield of paddy rice
(in kg/ha) achieved by household i

Natural logarithm of the fertilization rate
(in kg/ha) used by household i

Natural logarithm of the seeding density
(in kg/ha) used by household i

Interaction term between the logarithm Of
the fertilization rate in the household i
and the participation of the household in
producer group k

INTGBik = LENGUi * DGk
i = 1, 2, . . . , 1342
k = 1, 2, . . . , 55

Interaction term between the logarithm of
the seeding density in household i and the
operation Of the household in large peri-

meters

111

INTG4,1 = LSEMU, * DP1
where

DP

1 if the household Operates in a

1 large perimeter

DP1 = 0 if the household Operates in a
‘ small perimeter
LPROD, = Natural logarithm of the production of
paddy rice in household i
PROD, = Production of paddy rice in household i
(in kg)
MODM, = Number of active workers in household i
ENGUNIT, = Fertilization rate used by household 1
(in kg/ha)
SEMUNIT, = Seeding density used by household i

(in kg/ha)

U-, V, and H- represent the error in terms Of the
equations 1, 2 and 3. U-, V. and H- are assumed
to be aormal with means Of 0 and variance equal to

and .

u, v w
Taking the anti-log Of equations 1 and 2 and collecting
and rearranging terms, we can rewrite the production

equations as:

3 55
(c- +o<-) ZIDZ- 21d 06
(a1 + b3) 3 J j=1 J k=1 k k
PROD, = A(MODM,) *e *e *
55 2
(b, +K§IBkDGk> (b2 +1251X,DP1)
(ENGUNIT,) * (SEMUNIT,) + w,

The empirical estimates of parameters are summarized in
Table 20. A more complete presentation Of our most

satisfactory equations is given in Appendix C. In our next

112

section, we shall analyze the empirical results Of our

econometric model.

C. Analysis of the Empirical Results

Our analysis of the empirical results will involve two
parts. In the first part, the results of the estimation
procedure which are summarized in Table 20 are discussed.
In the second part, some examples are presented for
different producer groups in order to illustrate the

findings of the econometric analysis.

1. Analysis Of the Estimated Parameters

Our econometric model was estimated with cross-
sectional data. This has important implications for the
interpretation of the estimated parameters. In contrast to
time-series estimates, which yield Short-run elasticities,
the cross-section estimates represent long-run elasti-
cities. This difference stems from the underlying
assumptions Of these two techniques. While a time-series
analysis implicitly assumes that the various time periods
are homogenous except for factors explicitly appearing in
the function, estimating parameters from a cross section
assumes that all individuals are homogenous except for
differences explicitly introduced in the cross-section
function (Koutsoyonnis, 1983, p. 405). As a result of the
latter assumption, we are saying that if an individual with

a lower resource endowment gets the same quantities of

113

umuapucp mm_nmpem> asszu mmumuwu:_ x

mcopumacm umscoem:~EFw

 

 

  

Amo.v Aec.v ANo.V
No.5m so. Nemfi x x x cm. mm. OH. 8N.e exams .m.~
Aeo.v Aao.v ANo.v ANN.V
me.em om. Nemfi x x cm. em. oH. mm.e ezo¢4 e.N
Aeo.v Aeo.v ANc.V Amm.v
om.mm mm. mama x x «N. mm. c". mN.m page; m.~
Aeo.v Amo.v Amo.v ASN.V
mu.mHH em. Nemfi x «N. cm. m“. Ne." exam; N.N
Amo.v Amo.v ANc.v ANN.V
oa.HeH 8N. Nemfl we. we. NH. m~.m exams H.N
ANo.V
mN.omH mm. ecmfl x x NW. mc.m- 22m; ,m.H
ANo.v Aeo.v
em.-~ we. can“ x «H. 55.“- aamS N.“
Amo.v Amo.v
am.Hm No. memfl em. 58.“- aams H.fi
mowum_umum a mco_um> m mmpampem> memnesz
N umumanu< neomno mmpnapem> acmuco mu=_ acoucmamo cowumzcm
Lo
congaz

 

 

mucwpu_ywwou .mcopmmmeawm mg» mo mmume_umu pmupcwqu

cm mpnmh

114
inputs as another individual who has more, the first
individual will achieve a similar level of production. But
since the adjustment Of the production process requires
time, the estimates derived from cross-section data must be
interpreted as long-run elasticities.

Table 20 represents, sequentially, three area equations
and five yield equations. The different area equations
differ only by the introduction Of one or two categories of
dummy variables. Indeed, equation 1.2 is identical to
equation 1.1 except for the zones' dummy variables.
Similarly, equation 1.2 is identical to equation 1.1 except
for the dummy variables repreSenting the various producer
groups. The five yield equations present similar character-
istics, with the difference that the last two equations-
involve interaction terms. I

Under the particular design implied by Table 20,
changes in the adjusted R2 from equation 1.1 to equation
1.2 and from equation 1.2 to equation 1.3 represent the
specific contribution of the zones' and producer groups'
dummy variables to the explained variance of our dependent
variable. Therefore, the variations in R2 represent the
aggregate effects Of the localized factors which affect
variations in area allocated to paddy rice. In equation
1.1, an R2 of .02 is computed. This means that variations
in the number of active workers explain only 2% of the

variance Of the area cultivated in paddy rice in the

115

different households Of the Fleuve regions. When the
zones' dummy variables are introduced in the equation, the
equation now explains 68% Of the variance in the dependent
variable as Opposed to 2% in the first equation. Finally,
when the dummy variables representing the producer groups
are introduced, R2 increases from .68 to .85. These figures
indicate the critical importance of the factors Specific to
the location in a particular zone and the participation in a
given producer group in explaining differences in area
allocated to rice.

Area equation 1.3 reveals an elasticity Of area
allocated to paddy rice with respect to the number of active
workers equal to .22. The coefficient is statistically
significant even at 1% level Of confidence. This indicates
that differences in households' endowments Of active workers
was a significant determinant Of variations in the area
allocated to rice. Indeed, a household with 1% more active
workers than another one had, everything else being equal,
a plot .22% larger.

The complete area equation presented in Appendix C2
shows that all three dummy variables' coefficients are
significant at 1% level of confidence. This signifies that
each of our first three zones presents Significantly
different locational characteristics when it is compared
with the Matam zone in terms of its impact on variations in

area allocated to rice. Similarly, 41 Out Of the 55

116

coefficients Of the producer groups' dummy variables are
significant at 10% level Of confidence. This indicates that
41 producer groups have localized factors influencing
variations in area allocated to rice that were significantly
different from the characteristics Of the Gaodal producer
group initially used as the reference. In order to
facilitate the interpretation Of the dummy variables'
coefficients, they were transformed according to Suits'
procedure referenced in Chapter Five. The transformation
procedure and the new sets of coefficients are presented in
Appendix C3.

Yield equation 2.1 relates variations in the logarithm
Of the yield achieved in various households to variations in
the logarithm Of the number of active workers, and the
logarithms of the seeding density and the fertilization
rate. Equations 2.2, 2.3, 2.4 and 2.5 include the same
variables plus several categories Of dummy variables. As
was the case for the different area equations, it is
interesting to analyze the effects Of the localized factors
represented by dummy variables before examining the
individual elasticities.

In equation 2.1. 26% of the variance in yield is
explained by the independent variables included in the
equation. The explained variance Of yield is increased to
34% when the zones' dummy variables are introduced in

equation 2.2. This variation in R2 indicates that we

117
explain 8% more of the differences existing among the Fleuve
region's rice producing households with respect to yield
achieved when we add to their differences in input
availability and use, localized factors Specific to the zone
where each household is located. Hhen dummy variables
representing the different producer groups are added to the
independent variables already included in equation 2.2, R2
increases to .59. In contrast, when a variable representing
the interaction between the fertilizer use by a household
and the household's membership in a particular producer
group is included with the independent variables already in
equation 2.2, as is the case in equation 2.4, R2 increases
to .60. However, no change in R2 was recorded between
equations 2.4 and 2.5 as a result of the introduction of the
variable representing the interaction between the seeding
density used in a household and the location of the
household in a large perimeter. The largest increase in the
explanatory power of the yield equations was recorded when
some allowance was made for localized factors affecting the
yield of paddy rice and for other factors Operating at the
producer group level. Such allowance was made either
through the introduction of dummy variables for the producer
groups as in equation 2.3, or through the inclusion of
interaction terms. Because Of its greater usefulness for

policy purposes, equation 2.5 was found more satisfactory.

118
Therefore, it will be used to derive the different
elasticities.
Yield equation 2.5 reveals an elasticity Of yield with
respect to the number of active workers equal to .10. This
figure means that if a household had 1% more active workers,

its yield of paddy rice would, ceteris paribus, increase by

 

.10%. This elasticity was significant at 1% level Of
confidence. This finding confirms the significant role Of
the household's endowment of active workers in explaining
variations of yield in the Fleuve region. In addition to
the number of active workers, seed and fertilizer use had
important effects on yields. 'Indeed, the elasticity of
yield with respect to seeds was equal to .26 and was
significant at 1% level Of confidence. Such an elasticity
indicates that, everything else being equal, the yield of
paddy rice in a household would increase by .26% if the
seeding density was increased by 1%. However, the
coefficient of the variable repre- senting the interaction
between seeding density used in a household and the
household's Operation in large perimeters was not
significant even at a 99% level of confidence. This
indicates that no significant difference in the elasticity
Of yield with respect to seeding density was found between
large and small perimeters. The test was designed to see
whether differences in seeding techniques between the two

types Of perimeters had an impact on yields achieved. Based

119

on our evidence, no significant difference exists between
the two techniques as far as yield is concerned.

With respect to fertilizers, variables representing the
interaction between the fertilizer used by a household and
the household's membership in a particular producer group
were included in the equation. Therefore, a yield response
to fertilization rates can be derived for each of the 55
producer groups included in this study. Since the Gaodal
producer group was used as a reference, the coefficient of
the variable LENGU, presented in Table 20 represents the
elasticity of yield with respect to fertilizers in that
particular producers' group. He can say that if a household
operating in the Gaodal producer group increased the
fertilization rate it used by 1%, everything else being
constant, the yield of paddy rice achieved by that household
would increase by .38%. The elasticity of yield to
fertilizer in any other producer group can be found by
adding the coefficient Of the corresponding interaction term
to the elasticity found in Gaodal. To the extent that the
coefficient of an interaction term is significantly
different from zero, it indicates that important localized
factors (relative to Gaoda) affect the efficiency of
fertilizers in the producer group involved. But when an
interactive is not statistically significant, this signifies
the fact that the yield response to fertilizer in the

involved producer group was not significantly different from

120

the elasticity found at Gaodal. The elasticities Of yield
to fertilizers found for the producer groups ranged from a
low of -.08 in the 14th "groupement" Of Ndiatine to a high
Of .51 in the 37th “groupement“ of Gangal. Moreover, 31 out
of the 55 interactive terms' coefficients were significant
at 10% level of confidence. This finding is important
because it indicates that 32 different elasticities of yield
to fertilizers can be found. It also confirms important
inter-group variations with respect to the efficiency of
fertilizer use.

Finally, the coefficient of the dummy variable
representing our third zone was significant at a 5% level of
confidence. On the other hand, the other two coefficients
were not significantly different from zero, even at a 10%
level of confidence. These results indicate that in
addition to the variables already included in our equation,
other specific factors, highly correlated with the location
in the small perimeters of the Podor zone, had important
effects on yield. These factors were hypothesized to be
related with some physical, environmental and agronomic
elements. However, our results indicate that the Delta and
Middle Valley zone were not significantly different from the
Matam zone in terms of these zone-specific factors. In the
transformed version of our yield equation, the region's
average, rather than the Matam zone, was used as reference
of the zone's variations. But the Gaodal producer group

remained the standard of the variations in the elasticities

121

of yield with respect to fertilizers. He shall illustrate
our major findings with some examples before analyzing their

implications.

2. Some Examples

He shall take three examples corresponding to one
I'typical" situation and two extremes. For the “typical"
situation,_we shall select the Gaodal producer group which
was used as reference for the producer groups' dummy
variables. The Gaodal producer group, which is located in
the small perimeter Of Nguidjilone in the Matam zone, is
typical in the sense that it had the most frequent
elasticity of yield with respect to fertilizers. Indeed, as
we suggested earlier, the elasticities found in 24 producer
groups. were not significantly different from that of
Gaodal. The extreme cases correspond to the Gangal and the
Ndiatine producer groups where the highest and lowest yield
response to fertilizers were found respectively. For each
of these groups, the estimated paddy rice production
equation is presented below. All the coefficients involved
in our examples were significant at a 10% level Of

confidence.

--Typical Case: Production Equation at Gaodal
PROD, , .04(MODMi)(.22 , .10) e-.O7DZ4* e-.5soeso
,(ENOUNIT,)-33 ,(SENUNIi)-25
Since 0Z4 = 1 and 0660 = lewe have

122

PROD, = (.04)(.93)(.52)(MODM,)‘32 (ENOUNIT,)-33
(SEMUNIT,)°25 or
(6.1) PROD, = .02(NDDM,)-32 (ENGDNIT,)-38 (SEMUNIT)°25

--"Extreme“ Case 1: Production Equation at Gangal

(-1.72 + .34)OZ3

PROD, = .O4(MODM,)‘32* e ,9-230637

(ENGDNIT,)I-33"~13DG37)(SENDNIT,)-25
Since DZ3=1 and DG37=1, we have
PROD, = (.04)(.25)(1.25)(MODM,)‘32(ENGUNIT,)'51
(SENUNIT,)°26 or
(5.2) PROD, . .01(NDDN,)-32 (ENGDNIT,)-51 (SEMUNIT,)°26
--“Extreme” Case 2: Production Equation at Ndiatine
pRODi g '04IMODMiI'32* e(.85 - .18)DZl*e-.97DGI4*
(ENGUNIT,)('38 - ~45DGl4I (SEMUNIT,)°26
Since 02, = 1 and 0614 = 1, we have
PROD, = (.04)(2.80)(.37)(MDDN,)-32 (ENGDNIT,)--08
(SENDNIT,)-26 or
(6.3) PROD, = .O4(MODM,)°32 (ENGUNIT,)'°08 (SEMUNIT,)'26
The three production equations differ by the scaling factors
represented by the value of the constant terms and, more
importantly, by the extent to which yields are responsive to
fertilizers. Such a difference is shown by the different
elasticities Of yield with respect to fertilizer computed in
the three equations (.38 for Gaodal, .51 for Gangal and -.08
for Ndiatine).
It is interesting to analyze, based on these figures
and the information presented in earlier chapters, the

private profitability Of fertilizers at the margin. The

123

point is to see whether, given the prices Of fertilizer and
paddy rice and considering the production equations in our
three producer groups, it paid for the involved farmers to
use additional quantities of fertilizer. The same analysis
will be conducted considering the price changes that took
place in the last two years.

During the 1981/82 season, the average price paid for
farmers for one kilogram of fertilizer was equal to 25 CFA
francs. This price was increased by 100% in 1983/84 before
being set at about 90 CFA francs per kilogram in 1984/85.
During the same time, the producers' price of paddy rice had
the following evolution: 51.5 CFA francs per kilogram in
1981/82, 60 CFA francs per kilogram in 1983/84 and 66 CFA
francs per kilogram in 1984/85. The net incremental return
of using one additional kilogram of fertilizer was calcu-
lated for each of our selected producer groups at the
1981/82, 1983/84 and 1984/85 prices.

For Goadal, given an average fertilization rate of 426
kg/ha, an average paddy rice production of 1203 kg and an
elasticity of production to fertilizer equal to .38, the
ratio of marginal increase of paddy rice production per
kilogram of additional fertilizer use in 1981/82 can be
computed as follows:

Marginal product of fertilizer (in kg of rice per kg of

fertilizer = .0038 X 1203 kg rice 3 4.57 g 1.07
.Ol X 426 kg fertilizer 4.26

 

rice/kg additional fertilizer

124

This ratio indicates how many kilograms of rice could
be obtained at Gaodal by using one additional kilogram of
fertilizer. The net incremental return of fertilizer use is
obtained by multiplying the marginal product of fertilizer
by the producer's price Of paddy rice (to get the marginal
value product Of-fertilizer) and by subtracting the price of
one kilogram of fertilizer (the marginal cost) from the
result. Therefore, in 1981/82, we had the following net
return:

1.07 X 51.5 CFAF - 25 CFAF a 51.10 - 25 = 30.10 FCFA/kg
Of additional fertilizer. For 1983/84, the producer's price
Of rice and the price of fertilizers were respectively 60
and 50 CFA francs per kilogram. Therefore, assuming the
same marginal product of fertilizer as in 1981/82, the net
return of fertilizer used at Gaodal in 1983/84 was equal to:

1.07 X 60 CFAF - 50 CFAF = 64.2 - 50 = 14.2 CFAF/kg of
additional fertilizer.

For 1984/85, given a producers' price of paddy equal to
66 CFA francs per kilogram and considering that the price Of
fertilizer was equal to CFA 90 francs, the net return Of
additional fertilizer use at Gaodal, assuming that the
marginal product of fertilizerremained constant, was equal
to:

1.07 X 66 CFAF - 9O CFAF = 70.62 CFAF - 90 CFAF

= -1938CFAF.

The same computations were done for the Gangal and

Ndiatine producer groups and the results are summarized in

125
Table 21. In order to understand how the recent evolution
in relative prices have affected the private attractiveness
of fertilizer use in the Fleuve region, we can compute the
break-even price of paddy in the three types of producer
groups. The break-even price is calculated as follows:

Break-even price Of paddy = Price Of 1 kg Of fertilizer
_ Marginal Prodflct of fertilizer

 

At Gaodal in 1984/85 the break-even price of paddy was equal

to:
90 CFAF/kg of fertilizer = 81.8 CFAF/kg of paddy
1.1O kg rice/kg of fertilizer

 

At Gangal in 1984/85, we have a break-even price of paddy

equal to:
90 CFAF/kg of fertilizer = 84.1 CFAF/kg of paddy
1.077kg rice/kg of fértilizer

 

At Ndiatine further fertilizer use did not pay at any
positive price Of fertilizer. Moreover, additional
fertilizer use should not be encouraged in the Ndiatine
producer group where the marginal product of fertilizer is
negative.

The following comments can be made on the figures
presented in Table 21.

--First, the highest returns on fertilizer use at the

margin were recorded at the Gangal producer group

during the two seasons, 1981/82 and 1983/84. Indeed,
the households in that group could gain 32 CFA francs
for each additional kilogram of fertilizer purchased in

1981/82. Despite the doubling of the fertilizer

126

 

 

 

N.mom- cm N.m~H- oN.N- “Hoe we“ mm\ema~
o.NHN- om o.~c~- oN.N- “Hoe m¢_ em\mmmH a=.ua_uz
c.58H- mm c.8mfi- ON.N- “_me we“ Nm\Hmafl
¢.NH- ca e.- o“._ mac omm mm\¢mm~
c.8H om c.88 o~._ 88¢ emu ew\mmmH _ameao
o.Nm mN o.Nm o5.“ mac omm Nm\Hmm~
5.8“- cm w.o~ No.“ mom“ ewe mm\¢mo_
N.¢H om N.¢o No.~ moN_ owe ¢m\mmaH _auo~u
H.om mm “.mm No.5 mom" owe Nm\Hm¢H

ea~__eeeme 1N.u.zv Nm>zv Aea~.__ueae (me e. ~;\m¥ c. meomaam ace maaoew

_a=o_p_uua mucaee <L8 c. «agate «L8 =. m¥\au.e mew auuaa co magma

No ax eon EONPppuemm :o*uu=cceg em~.__uemu :owuuzuoLa copum~v__pcmm

m<mu cw :ezumm No on: paucms —uu:msaeu=~ mo «esteem ommem>< mameo><
buucmsmeocfi umzi umeucfi Lo “mop, (LO oabw> pmcpaeaz,

 

muwmowmh<o

 

 

m:¢mo¢m.mwesh mc—cso mqaoew emuauoea vmuum_om mouse
=~ =_meaz meg an mem~__.ueaa co Nae—ena8_eoea aua>_ea

HN mpamh

127

purchase price in 1983/84, fertilizer use remained
profitable at Gaodal even though the incremental net
return decreased by 50% as a result of the modification
in the relative price of paddy rice with respect to
fertilizer. A similar phenomenon took place during the
same periods in the Gaodal group where the incremental
net return drOpped from 30.10 to 14.2 CFA francs.
However, fertilizer use remained profitable at the
margin for households Operating in the Gaodal and
Gangal producer groups.

--A completely different situation existed at the
Ndiatine producer group, where an increase in the
fertilization rate used by 1 kg resulted in a loss Of
154 CFA francs in 1981/82. The situation was even
worse in 1983/84 since a net incremental loss of 212
was experienced. These figures suggest that it did not
pay to encourage further fertilizer use at Ndiatine.
--Given a producer's price Of paddy equal to 66 CFA
francs and given that the purchase price of fertilizer
was equal to 90 CFA in 1984/85, fertilizer use is not
privately profitable at the margin in any producer
group of the Fleuve region. Indeed, even at the Gangal
producer group where fertilizer use was the most
efficient, a net loss of 17.4 FCFA is recorded per
kilogram of additional fertilizer. The net loss per

additional kilogram of fertilizer is equal to 19.4 CFA

128
francs at Gaodal and 268.2 CFA francs at Ndiatine

These findings have important policy implications that

will be analyzed in our next section.

0. Implications of the Econometric Model

The results of our econometric analysis provide some
additional insights into the characteristics of the
different perimeters Operating in the Fleuve region. In
addition, they have important implications both for the
evaluation of the different perimeters and for the various
strategies to increase paddy rice production. In fact, a
major policy issue deals with understanding the respon-
siveness of yields to fertilizers and seeds. The knowledge
of the magnitudes and directions of these responses provides
an important policy tool to government Officials for their
investment strategies.

In relation to those issues, our analysis has shown the
importance of fertilizers and seeds in explaining variations
in rice production in the Fleuve region. At the same time,
the analysis has established that despite the importance of
modern inputs there is no guarantee that increasing their
quantities will result, everywhere, in higher output.
Indeed, several complementary elements need to be con-
sidered. They include the characteristics of the soils,
the adequacy of the plot's grading, the incentive system and
other technical, managerial and organizational factors which

are important in irrigation. Because of the heterogeneity

129

of the Fleuve region's producer group with respect to those
technical, social and environmental factors, the
effectiveness of fertilizers, for instance, is very
different in many perimeters. In fact, as illustrated by
the case of Ndiatine, negative responses are possible.
Indeed, it is conceivable that the different inputs to rice
production are in the special case Of complementarity that
Glenn Johnson characterizes as the "airplane tail" type, as
opposed to the more general one which allows for some free
disposal.1 Under the ”airplane tail“ type Of complemen-
tarity, deviations from the approximate mix of inputs will
result in decreases of output. I

The negative response of yield to fertilizers recorded
at Ndiatine could also happen if farmers who declared that
they used the largest quantities of fertilizer did not
actually put them on their rice plots whereas those who
received the smallest quantities utilized them in their
fields. Such a situation would be consistent with a
negative elasticity of yields to fertilization rates at
Ndiatine. It is difficult to tell what really happened at
Ndiatine but our second hypothesis cannot be completely
ruled out because we know that at the time of the survey
fertilizers were highly subsidized by SAED and the proximity
of Mauritania provided some Opportunity to smuggle goods

into that country.

 

1Glenn Johnson's AEC 805 couse notes, 1984.

130

This question brings up the more general issue of the
incentive system and its impact on farmers' willingness to
use modern inputs. The examples we provided on the private
profitability of fertilizer use can illustrate the point.
Indeed, for most households producing rice in the Fleuve
region, it paid at the margin to use additional quantities
of fertilizer in 1981/82, given the characteristics of their
production functions and the structure Of relative prices
during that year. But for other farmers with different
production potentials, using more fertilizer resulted in
losses. This indicates that, because of the large degree of
heterogeneity of producer groups with respect to agronomic,
social and environmental factors, no Single strategy to
increase rice production is likely to be successfully in all
instances.

In addition, the recent evolution in the price Of paddy
rice with respect to the price of fertilizer has consider-
ably reduced the private attractiveness Of fertilizer use,
even for the most successful producer group. In 1984/85,
fertilizer use remained privately profitable at the margin
for most producer groups, despite the important decline in
the purchasing price Of paddy rice. But in 1984/85
fertilizer use was not profitable at the margin in any
producer group at a producer price of paddy equal to 66 CFA
francs. In fact, given a price of fertilizer equal to 90

CFA francs, additional fertilizer use was unprofitable in

131

the Fleuve region at any producer price of paddy less than
81.8 CFA francs per kilogram. This explains why SAED was
having difficulties in finding paddy sellers in 1984/85.
whereas farmers were selling paddy rice in the parallel
market at 82 CFA francs per kg.2 These findings illustrate
the critical need for Senegalese government Officials to set
up relative prices consistent with the various components Of
their agricultural policy.

Finally, there is some justification for adjusting the
evaluation methods and the various investment strategies to
the diversity of the production environment. More specifi-
cally, more attention needs to be paid to localized factors
and their compatibility with a favorable production

environment.

E. Limitations of the Study

Before closing this chapter, it is important to analyze
some of the problems associated with this study and the
ways in which they may have affected our econometric
results. The major problems have to do with our data
collection methods. Indeed, the data collection, done by
one-shot surveys, relied to a very large extent on farmers'
memory. This probably constituted an important source of

inaccuracy in the data. Senegalese farmers do not generally

 

2This information on the 1984/85 season was provided by
Michael Morris, and ISRA/BAME researcher studying rice
marketing in the Fleuve.

132

keep written records on their activities and memory recall
remains the most common source of information. In addition
to problems associated with recall, conversions from local
units, generally based on volume, to standard measurements
Of weight can be hazardous because of possible variations in
the volume/weight relationship between and within localities
in addition to its variability over time. It is important,
however, to mention that the same enumerator interviewed all
the farmers of the same perimeter. Therefore, it is
possible that biases related to the quality of the
enumerator did not affect the information collected with a
given perimeter, although some of the inter-perimeter
variations could have resulted from differences among
enumerators.

Because of the probable measurement errors that exist
in the data, it is important to understand how they can
affect our econometric analysis. Several cases of errors
are discussed in the literature (Koutsoyiannis, 1983;
Pindyck and Rubinfeld, 1981; Kennedy, 1979; and Foote,
1958). Let us examine these cases briefly.

--The dependent variable is measured with error: the

effect is to increase the error variance. But, if the

independent variables are uncorrelated with the error
term, the estimated lepe parameters are still unbiased

and consistent.

133

--The independent variables are measured with error.

In this case, the least squares estimates of the

regression parameters will be biased toward zero and

inconsistent. The degree of bias and inconsistency is

a function of the variance of the measurement errors.

--The dependent and independent variables are both

measured with errors, with no other error in the

equation. In this case, the least squares technique
can lead to a bias toward zero of the estimated
regression parameter and to inconsistent estimates.

But, if the variance in the measurement error of the

independent variable is known, it is possible to obtain

consistent parameter estimates.

In addition to these problems, the question of the
representativeness of the 1981/82 season can be raised.
However, given the relative independence Of irrigation
activities on rainfall, it seems reasonable to think that
the 1981/82 season represented a typical year. Nonetheless,
it is important to emphasize the experimental nature of this
study and the need to interpret the results of the econo-
metric analysis with caution. DeSpite its problems,
however, the analysis provides a first approximation of the
relationships involved in the Fleuve region's rice
production systems. More importantly, the hypotheses made
here can be tested and improved in other studies. Sugges-
tions for such further studies are presented in the next

chapter.

CHAPTER SEVEN

SUMMARY, CONCLUSIONS AND RECOMMENDATIONS

A. Introduction

The Objectives Of this final chapter are to
recapitulate briefly the issues dealt with in this study, to
summarize the major findings and to draw some
recommendations.

The series Of droughts in the 1970's and the consequent
food crisis experienced by the sub-Saharan countries have
inspired ambitious programs aimed at achieving food
security. The exploitation of the Senegal River's resources
constitutes a major component of such programs carried out
by the countries along the river. The development of
irrigation schemes, which allows reduction in the weather-
related fluctuations of production, is perceived as a major
means to achieve the stated Objective. But because of the
scarcity of local resources and the need for their efficient
allocation, it is important to analyze how alternative
schemes compare in promoting food security. Food security
is, however, a complex notion and there is no easy and
simple way to measure it. Yet, the way food security is
perceived by government officials determines the policies
implemented.

In this study, we have selected the per capita

availability of cereal in producing households as a major

134

135
index of food security. However, the per capita
availability of cereal in the Fleuve region is determined by
farmers' production and several categories of transactions.
Inasmuch as farmers' production remains the major
determinant of cereal availability, our analysis has focused
on the various elements that affect its performance in the
Fleuve region. Finally, because Of the important role
played by irrigated rice in the Fleuve region's production
system and because of the high value placed by Senegalese
government officials on rice production, this study has put
a considerable emphasis on assessing the performance of the
different irrigation schemes in the region. Our major

findings and conclusions are summarized in the next section.

B. Summary and Conclusions

‘Our brief historical analysis has shown that attempts
to develop irrigation in the Senegal River Valley are not
new. In fact, as early as the nineteenth century, the
French colonists introduced irrigation into the Fleuve
region. The different projects implemented throughout the
past two centuries have had various degrees Of success.
These projects have generally provided the technical
foundations for the more recent irrigation programs
developed in the region. Indeed, SAED'S experience with the
large perimeters in the 19605 and 19705 represented the
latest forms of the colonial technological legacy. In fact,

SAED'S operations have gone through several changes in the

 

136
irrigation techniques used, the conception of the schemes,
as well as in the modes of management.

The major technological shifts have been towards a
better control of water and an adaptation of the irrigation
techniques to the particular physical characteristics of the
Senegal River Valley. —In this respect, the physical
environment, along with the decision of the Senegalese
authorities to develOp a rice industry able to substitute
for the increasingly large quantities of this commodity
imported annually provide the rationale for the earlier
model Of large scale and highly mechanized perimeters. But
the performance of large perimeters has not been
particularly impressive, as they required large subsidies
from the government. Meanwhile, farmers who were left with
little initiative did not seem to benefit very much from the
experience. On the other hand, the smaller perimeters,
which are more recent and involve more participation from
farmers, thanks to their labor-intensive techniques and to
the more decentralized management, seemed to perform better.

Our descriptive analysis of the two types of perimeters
during the 1981/82 season has allowed us to discover several
interesting features of cereal production in the different
zones of the Fleuve region. The most noticeable feature has
to do with diversity of the crOpping structure in the
different zones. Indeed, such a diversity is apparent in

the evolution from the quasi specialization in rice

137
production by the Delta zone's farmers to the combination Of
irrigated farming with rainfed and flood-recession
agriculture in the Matam zone.

Yields were found to be, on the average, higher in
small perimeters. But, thanks to the larger size of plots,
tOtal production of rice per household was more important in
large perimeters, resulting in higher output per capita.
However, rice production constituted only one component of
cereals availability in the Fleuve region. As we have seen
in Chapters 3 and 4, some farmers practiced multiple
cropping, mainly in small perimeters where they combined
irrigation with flood-recession and rainfed agriculture.
Therefore, the quantities produced in the three types of
farming should be included in a more precise assessment of
the overall availability of cereals. But, as indicated by
the number of households involved with the different crOps,
the proportion of farmers who raised all the cereal crops
was relatively small.

Moreover, a large variability existed among individual
households in terms of paddy rice production and yield.
Such a variability was hypothesized to be related to the
heterogeneity of the Fleuveregion's households with respect
to physical, technical, economic and sociological factors
The econometric analysis was undertaken to provide an
approximation of the quantitative relationships involved in

paddy rice production in the Fleuve region. The analysis

138

has allowed us to determine how differences in resource
endowment, along with other localized factors, explain
variations in production. More specifically, it has been
possible to isolate the effects of several inputs on
production. In this respect, the importance of labor
availability, fertilizer and seed use in increasing rice
production were Shown through the derivation Of the
elasticities Of output with reSpect to these inputs.

As anticipated, the efficiency of modern input use
varied significantly among several producer groups. The
variations were apparent not only through absolute
differences in production, but also via the changes in the
elasticities of output with respect to various inputs in
the different producer groups. These findings confirm the
high degree Of heterogeneity in terms of technical,
environmental and social circumstances. They also have
important implications in terms of the methods used to
evaluate the performance Of the irrigated perimeters.

Indeed, almost all evaluation methods require an
approximation Of the production function involved. In the
context of our study, the conceptual problem is to discover
the relationships between several characteristics Of the
irrigated perimeters and cereal availability. Yet, most of
the cost-benefit analysis conducted on the perimeters is
based on the simplistic assumption of an average yield for

a given type of perimeter. Moreover, our analysis suggests

 

139

that even the sensitivity analysis generally undertaken to
allow more realism does not fully account for the hetero-
geneity of any given perimeter. This point is extremely
important for the investment strategies designed to increase
rice production in the Fleuve region. Indeed, the large
degree of heterogeneity of the different perimeters implies
that there is no unique strategy to increase production in
the Fleuve region. Rather, it seems more appropriate to
design and implement policies relevant to the particular
circumstances of producers. In our next section, we shall

provide more Specific recommendations.

C. Recommendations

He Shall divide our recommendations into two parts. In
the first part, we will deal ‘with some policy recommenda-
tions. In the second part, we will suggest some issues that

need to be investigated by future studies.

1. Policy Recommendations

Our policy recommendations are suggested by the
findings Of our study and the constraints imposed by
Senegal's overall macro-economic Situation. The recommenda-
tions are relative to the conditions required in order to
increase cereal availability in the Fleuve region. Some
suggestions are also made with reSpect to the institutional

reforms that are currently being introduced in Senegal.

140

The current strategies designed by Senegalese
government officials in order to increase cereal production
rely on plans to increase irrigable lands by 5,000 hectares
a year. Given the very large costs involved in the
development of new perimeters and considering the financial
problems being experienced by the state, the realism and
feasibility of those strategies are highly questionable.
Indeed, the cost issue is a crucial one and it involves not
only the expenses resulting from the construction and the
maintenance of new perimeters but also the opportunity cost
of giving up more effective options. One option would be to
increase the productivity of already existing perimeters.
In this reSpect, the results of our investigations can
suggest some recommendations.

We have seen in Chapter 2 how SAED was able to bring
noticeable progress in the performance of large perimeters
thanks to the elimination of financial, organizational, and
technical bottlenecks. Furthermore, higher yields were
consistently recorded in small perimeters where farmers have
more responsibilities and are more motivated. Given the
size of individual plots in large perimeters, significant
increases in production could be obtained by achieving in
the large perimeters yields of paddy rice comparable to
those recorded in small perimeters. Our analysis shows that
in most producer groups, labor, fertilizers and seeds have

significant impacts on production. Furthermore, the

 

141

technical efficiency of those inputs might be improved in
most producer groups by eliminating the technical, environ-
mental and social constraints which operate at the local
level. An identification of the localized factors should
allow the design of selective strategies, based on the
concrete conditions and needs Of each perimeter and producer
group, in order to increase yields in currently existing
perimeters.

Another aSpect of our alternative strategy would be to
modify the current approach based on the double speciali-
zation in rice production and irrigated farming. Our
descriptive analysis has shown how SAED's activities were
focused on irrigated paddy rice. Yet, the analysis of the
structure of cereals' availability in the Fleuve region
reveals the important contribution of millet, sorghum, maize
and cowpeas grown in the rainfed and flood-recession
agriculture. This militates in favor of seeking a
simultaneous development of the different types of
agriculture. We have seen that in most zones of the Fleuve
region, farmers combine the different types of farming and
this results in seasonal labor bottlenecks during the rainy
and winter seasons. In this reSpect, the introduction of
animal traction, which is almost nonexistent in the whole
region as _we discovered in our descriptive analysis, might

offer an effective and less expensive means to overcome the

 

 

142
current conflicts existing among the different forms Of
agriculture in terms of labor allocation.

However, other conditions must be met in order to
guarantee the simultaneous 'development Of irrigation,
flood-recession and rainfed agriculture. Among these
conditions, we can mention the need to improve the input
delivery system and encourage a market for cereals other
than rice, the development of high-yielding varieties
adapted to the region and a favorable structure of relative
prices. The question of relative prices deserves special
attention as they relate to the social and private
profitabilities of cereals production in the Fleuve region.

From society's point of view, the question is whether
it pays to produce rice in Senegal given the possibilities
to import the commodity at a lower cost. Indeed, to the
extent that Senegal does not have a comparative advantage in
rice production, Senegalese society may be better off by
using resources saved to produce more goods and services
other than rice. Such an option would contribute more
effectively to the country's food security. However, the
issue is not entirely technical, as welfare and political
considerations could justify options that would not prevail
on purely economic grounds. Moreover, because of the lack
Of homogeneity between the 100 percent broken rice imported
in Senegal and the long-grain rice produced locally, there

could be some economic justification for continuing to

143

produce the commodity in the Fleuve region. Indeed, given
the fact that it has a higher price in international
markets, the locally produced long-grain rice could be
exported while Senegal continues to import the 100 percent
broken rice in order to satisfy domestic food requirements.
This scenario would allow the country to pursue its
objectives of food security while improving Senegal's
balance of payments position. But, because of the large
quantities of cereals imported annually, domestic production
would have to be increased. This brings up the need to use
modern inputs in order to increase production and the
question of the private profitability of such actions.

In this respect, our analysis has shown that the use of
additional quantities Of inputs could contribute Signifi-
cantly to produCtion increases. However, whether farmers
use additional quantities of inputs or not depends largely
on whether it pays to do so. In fact, it seems reasonable
to assume that they would not use additional inputs unless
they could expect to get an incremental value in their
production at least equal to the cost of acquiring
additional inputs. Yet, as we have shown it in Chapter 6,
given the characteristics of the production equations in the
Fleuve region, and given the marginal product of fertilizer,
in 1984/85 it did not pay to use additional fertilizer at
the margin in rice production. This is why, in a recent

study, 60 percent of the Fleuve region's farmers cited high

144

prices as the reason (why they did not use more fertilizer
this year.1 This has some important implications for the
strategies used to increase production. In the short run,
the prospects of increasing farmers' income through
significant shifts in the production function are relatively
limited. Therefore, the Social profitability of the
increased use of modern inputs depends on the relative price
of output to inputs. The issue involves some fiscal and
welfare dimensions which need to be considered along the
more global issue of the social profitability of producing
rice in Senegal. But some basic conditions are required in
order to have a coherent policy. Indeed, it is not
realistic to expect farmers to buy fertilizers at full cost
if they do not get prices for paddy rice which make it
privately profitable to use fertilizers in order to increase
production. The argument calls for a greater coherence in
the Senegalese officials' pricing of inputs and outputs.

In the long run, as technological innovations made
possible by research on new varieties and production
techniques allow increases in production at lower costs,
more policy options will be available to government
officials. Indeed, farmers will be able to increase their
total revenues through larger quantities produced and sold.

In addition, to the extent that they involve cultural

 

1See Eric Crawford and Valery Kelly, “A Field Study of
Fertilizer Distribution in Senegal, 1984: Summary Report,“
BAME, November 1984, p. 22.

 

145
design, as Shaffer argues, policies could be aimed at
reconciling the national demand of cereals with the
production potential of the country. Meanwhile, the
conditions of coherence and consistency apply also to the
different policy and institutional reforms planned by
Senegalese government Officials.

In this reSpect, even though the need to cut costs can
be a legitimate reason to seek a smaller involvement of the
state in production and marketing activities, it is
important to provide farmers the training programs they need
in order to increase their managerial skills and their
technical know-how. In fact, the movement toward a greater
autonomy of farmers can only be successful if they are
prepared to be viable economic actors. This highlights the
critical need for more research oriented toward farmers'
needs. Even if the state does not get directly involved in
production and marketing activities, it will still have to
provide farmers' organizations and the private sector the
incentives to carry out these important functions. Several
important issues are involved and more research is needed on
them. In our next subsection, we suggest some issues that

deserve further investigation.

2. Needed Research

The urgency to tackle the problems resulting from the
current food crisis and the complexity of the questions
involved require that more research be done on several

issues in order to successfully guide future policies. We

146

will outline here three general themes that require some
Special attention.

The first research issue deals with the question of
“food security“. A better definition Of the notion,
including more macro-dynamic perspectives, seems to be
particularly needed. In this respect, we can mention the
work initiated by Frederic Martin, whose results should
provide some useful insights on the issue.2 But, as
indicated earlier, the question is neither simple nor is it
entirely a technical issue. In fact, the way in which this
issue is settled will largely determine whose interests
count. In addition, an improved understanding of the issue
could facilitate the determination of strategies more
appropriate to Senegal's situation.

The second issue has to do with a better assessment of
the production capacities of different zones Of the Fleuve
region. An identification of the localized factors which
affect rice production in the Fleuve region would be
extremely useful. Indeed, a better knowledge of these
factors would allow SAED and the other participants of the
Fleuve region's production system to increase the efficiency

of resource use. In this respect, the findings of the

 

2Frederic Mratin, “Analyse des Avantages Compares
Dynamiques Sous Incertitude du Senegal Dans le Cadre
d'Une Strategie de Securite Alimentaire. Proposition de
Recherche,” Dakar: BAME, ISRA, Janvier 1985.

147
farming system research team working in the Fleuve region
could be very useful.

Finally, in the perspective of ongoing reforms toward
greater involvement of the private sector and farmers'
organizations, the 4 role of training and. education is
critical. Indeed, the large increases in production
required to meet the country's needs can only take place if
participants are trained and motivated to adopt and
assimilate new technologies. This raises the issue of
institutional linkages between training, research, extension
and the political process. Further studies are needed on
the functions and interactions of these institutions.

Those studies can benefit from the lessons of the
errors and problems that we had in this one because of the
lack of a clear conception of its objectives and analytical
procedures. In fact, most of the problems could have been
avoided, had a complete planning of the whole research
process taken place before any data collection was done.
Such a process would have allowed adjusting the data
collection to clearly defined Objectives Of the study, the
method of data analysis and to the conditions under which
farmers Operate. In this respect, a greater familiarity of
researchers with the milieu of the study, the pretesting of
the questionnaire and a greater consideration for analytical
and processing capacities would have allowed us to avoid

collecting unnecessary data.

 

A P P E N D I X A

THE QUESTIONNAIRE USED

 

149 - - .. - ..._..

 

‘1. I I _I Régia: -
:3. I. ‘1‘}
‘5. ' _I__'_I -
7. 1 _I..__I , , $
9. an vide “
.13;1_.__.1m1m:ibiznsteauverso . .
II. I _‘_I_:_I_ - I _I_ : . I Superficie 61 hectares (1e point tepcesaite la margae décirrale) .
45. cu- vide . -_
11m - - '
17.1_1._.'_1___1o.me1t.~:saewenxe
21. I_I_I migine : Vbir les codes an verso
23. I__I__I_I Pris: 61.1 “kg: pan: 1 et 2. ptendte 1e prix de vaite mitaire de la
atcductionr
' 26. I__I__I_I ___I thite's d'engrais utilisées (kg)
Mettte des zeros s'il n 'y a pas d'engzais
I). I__I_I Psi: du Kg
32. !___I Praduit White pm: W I 51. 1e produit eat
33. I__I " " ' " allure I utilisé, mettre I
34. I__I ' ' ' . " témlte I Skmmettre O
35.. C230 vide ’
mmfiaamwmms
35. 1 G._.,_1_"_._._.__1___1 Ptoducticn en kg
:41. _I__I___I_I mantite's vendtm aux cooperatives en kg ' ' .
45.! I I I I WdtésvmdeamScciétésdedéveloppamtmkgv'
I

e'mantités vendues aux particuliezs 61 kg -

I I I I V‘ihleix‘totale des’ ventes 'aux partimliets a: centains de ,
trans. Par ample ppm: 500 _frs, mettre 005

1 1 1 1 1 i’omtite's achetéesdanslevillagemkg.

I I I I ‘,.‘.'ale11t totale enoentaines de francs.

1 1 1 1 1 "mattitésachetées botsduvillageenkg

' m1: mule en caninesde francs.

I
I I I I 1 mm lee ms d'achat de :12, indiquer les giantités de ti:
Paddy (non de'ooztiqué) sur 1:: total achete' e1 Kg.

I I I_I' I'mautizPeddymcuitainesdetrm.
I_I_I___I mutual. domées came cadeau en dizaines de mg
couple pour 40 kg, écrire (1)4 ’
31 cas de bacin arrondir a 1' unité le plus prome
Btu-pie : pour 48 kg écrize CD5, pour 21 kg écrire 002
I 101113 Siledimestdamée,mettreisixmmettxeo

1
3
5
7
9

10..

12
14

16

17‘
18
19

20:
21”

I\) [\I
*J (h \n 4:.

I‘) N

N
(‘3)

150

III. 1.1.110 "VILLAGE"

¢—=-=-=—=—=-=—=-

Inzt'rzngglczi

!.A;J___J Région - Zone Vbir codes au verso
!___I_I Périmétres ou Secteur ' n n n n
I... _! Grouvements de Producteurs. " n .. n
!___1___! Ndieul ou Foyré n n

Case vide

HAIK-D'OEUVRB

3___!__! Population du :‘ldieul ou Foyré

! i i Kain-d'oeuvre disponible

I ' 1 1301111119 de saisormiers

Cale vidd

'"NIVZAU TECHNOLOGIQUE

Kettre 1 dans 1e (3) can concerné (s)
Kettre 0 dons les,gu§re3.c§s., ,

! ! Culture munuelle

3 I Traction bovine

! I Traction equine at on asine

1 I Culture mécanisée

Cece vide
STRUCTURE DE L'EUDSTTELQEQ

iettre 1 cans 12(5) cas concerné(s)
Eattre 0 flags-les autres cos.

2 .2 Bette sur le travail du sol

! ! Bette sur l'irrigation
I i Détte sur ln'livréiécn d'en _.is
!__"}~ Detteosur la livraison des semences

.3 1 Dette sur la livraison de produits phytosanitaires

I 1 Dette sur le battage

Case vide

! I ' ' Remboursement on nature ! tot re 1 :i

I E E I ! ! Quantités remboursées en kg

1 _! Hemboursement en numéruire lKettre 1
fire 0 I
1‘ I. I ! E I Iontant du remboursement on francs.

c'est la cas.

oi c'est 19 cos.

Sinon met

Sinon met-

oouoam-wai—A

CONOSW

LISTE DES CULTURES

 

Riz
Mais
Sorgho
Mil
Fonio
Niebe
Ble
Autres

CAP-VERT
CASAMANCE

DIOURBEL

FLEUVE
SENEGAL-ORIENTAL
SINE-SALOURA

THIES
LOUGA

oowmu-i

151
LISTE DES ORIGINES

 

1 - Personnel

2 - Autres paysans
3 - I.S.R.A.

4 - Cooperative
5 - SONAR

6 - P.R.S.

7 - Ble PIDAC

8 - N.A.C.

9 - SO.DE.VA

10 - S.A.E.D.

11 - SODEFITEX
12 - Autres

- Grands Perimetres du Delta

- Grands Periretres de la Moyenne
Valley.

- Petits Perimetres de la zone de
Aere Lao

- Petits Perimetres de la zone de
Matam

- Zone de cultures traditionelles

A P P E N D I X B

CODES REPRESENTING ZOMES, PERIMETERS AND
PRODUCER GROUPS

153

 

 

 

Zones Perimeters Producer Groups
4.1 Delta 01 Lampsar
02 Ndelle
01 Lampsar 03 Thilene

O4 Boundoum Peulh
05 Diagambal

06 Boundoum Barrage 1

O7 Boundoum Barrage 2
02 Boundoum 08 Boundoum Est 3

O9 Boundoum Nord 4

10 Boundoum Nord 5

11 Thiager
12 Mbagam
03 Richard Toll 13 Khor
14 Ndiatine
15 Thiagar CT

4.2 Large Perimeters of the Middle Valley

16 Gae
17 Gae
01 Dagana 18 Gae
19 Gae
20 Gae

U'Ibwmi-I

21 Niadane 3

22 Niandane 13
02 Nianga 23 Guia 5

24 Guia 7

25 Cuma 2

26 Gurde Chantier
27 Diama
03 Guede 28 Fresbe
29 Guede Village
30 Agnam
4.3 Small Perimeters of Aere Lao

31 Nianga Ndioum
32 Boade

01 Ndioum 33 Lidoube
34 Thioulbe Galle
35 Nienga Edy

36 Pete
37 Gangal
02 Pete 38 Toufne Gande
39 Tikite
4O Mboumba

154

Zones Perimeters Producer Groups

 

 

03 Demete 43

4.4 Small Perimeters of Matam

01 Matam 43

02 Boki Diawe 53

03 Nguidjilone. 58

NOTE:

The producer groups numbers 12,
excluded from the analysis because of
large number of missing values in these

Demete

Data Afaybe
Cascas
Thioubalel
Hassetake

Matam 5
Matam 1
Matam 2
Diamal 1
Tiguire Cire

Ndouloumadji F.2
Nabadjimbal
Dounga 0.1
Mbakhna

Kobilo

Sedel 2
Nguidjilone 1
Nguidjilone 3
Doundous
Gaodal

15, 47 and 48 were
the presence of a
groups.

A P P E N D I X C

0.1 Results of the Estimation Procedure
C.2 Presentation of our Most Satisfactory Equations

C.3 Transportation of the Dummy Variables' Coefficients

156

.c.vn
ac.vm
cN.¢__
=O.nn
co._c.
=M.N
=~.c
am.vm
m...3_
om..-
qw.-n
-.~m
co.N:v
c..cm
mm.e_r

wu.am¢u¢
.. n...

Ii,l III '02.: ..

=c .
..c.
en.

am.

0:“.

«a

mem—
Nep—
Nym—
Nem—
men—

I.I.l|al' I.-- I-

cem-

men-
men—
can.
cen—
cen—
cen-
cem—
cc»—

cen—

.§:.:~£. 32:5. .mmfiwmufifi. whet...
2.2.2... 3 .35...

II 'PIII i

I'I‘IIII. uI.-I|I.II

 

.III OIIIIID

.2... .x...
e... .3.
.2... .x...
2.. a.
.2... .2...
z. 2.
.2... .2...
2. .3.
.2... .2...
2.. a...

:2...

.8... .N...
2... E.
.3... .N...
5.. 3.

.eo..e.e.
eo..~::c as. e. can: o.ao..o> utlae .~.:u...oa as. n...eoe. managed g..3 a..ouN
.mang.eo.oa e. eo,.a o.- n.eo.u...oou emu-l..ao as. .o nee..a..o: stoeea.m 03—.

 

.2...
a
3......
8....
5...:
8.3

a...

S... .5...

.-..

2... 5...:

.2...

2.... ea...

2....

2.. =3...

.2...

2.. is:

.33.

9:3 5....

3.69..

50.82:...

2.3..

8.32;...

.....

3..- .52

3....

3... 23..

E...

3..- ea.

3...

8. ea

.2...

2. a...

8....

m_.. a=m

.2...

8.1 33m
mg.5n.cu>

.w...e_z.....r§ E. 2.2.... -..EFI.E LEEP..E.._E..: £3. ii.
mafia—g) ziluzui—

0,. 3.2.1.3....

IIIII II... II VII-‘OI'IIIIII1".-.501‘0'1901lilt-

iII-o- '1..Iv‘u.IvIIqu.

I'D .iI.II I-OI III'I'IIIOO III- 1.

=o..:l..a. as. .o m..=msz ..u

u.- ! IIIOIIInI .

a=oec0=o=

III I..- 1 I II!

a.
e—
n.
N-
_—

39.1....
eo..ozsu

Ii III'O...‘

where:

INTG1,,

1N182,,

The other

157

Interaction between the fertilization rate
used by household, and the presence Of the
household in producer group j

1N181,, = ENGUNIT, * De,

1 = 1, 2, 3 . . . , 1366
3.31, 2, 3 e o e g 56

Interaction between the seeding density used
by the household, and the presence of the

household in producer group j.

variables are defined as previously.

158

C.2 Presentation of our Most Satisfactory Equations
(Standard errors are shown in parentheses)

I. Estimated Area Equation

 

LSUP = -1.68 + .221 MODM + 1.580Z, + .620Z2 - 1.020Z3
(.11) (.02) (.12) (.12) (.12)

+ .69061 + .25062 + .92DG3 + .25D64 + .67D65 + 1.08066
(.11) (.11) (.11) (.11) (.13) (.11)

+ .68DG7 + .67D68 + .75D69 + .86D610 + .97DGll
(.11) (.11) (.11) (.11) (.11)

+ .140613 .32DGl4 + .59D616 + .680617 + .73DGl8 + .29DGl9
(.11) (.11) (.11) (.11) (.18) (.11)

+ .50D620 + .48D621 + .730622 + 1.200623 + .62D624
(.11) (.12) (.12) (.20) (.14)

+ .96D625 + .430626 - .71D627 - 1.11D628 + .780629 - .18D630
(.14) (.10) (.11) (.10) (.10) (.10)

- .110631 + .O7DG32 - .410633 - .21D634 + .05D635 + .77DG36
(.12) (.12) (.12) (.12) (.12) (.12)

+ .88D637 + .96DG38 + 2.540639 + 1.53D64O + .23D641 +
(.12) (.12) (.12) (.12) (.12)

+ .080642 + .48D643 + .640644 + .48D645 - .62DG46 - .29D649
(.12) (.12) (.12) (.13) (.49) (.21)

+ .ochso + .260651 + .480652 - .090653 + .160654
(.15) (.15) (.15) (.15) (.15)

+ .41D655 + .06D656 + .24D657 - .O9D658 + .130659
(.15) (.15) (.16) (.15) (.15)

N . 1366 R2 = .85 F = 130.75
II. Estimated Yield Equation

 

LRDMT, a 4.95 + .38LEN6U + .26LSEMU, + .1OLMODM, - .46DZl

(.04) (.06) (.02) (.39)
- .29022 - .29DZ3 - .011NTG31 - .0021NTG32 - .06INTG33
(.39) (.15) (.02) (.02)
- .17INT634 - .O3INT635 - .06INT637 - .251NT638 - .121NTG39
(.02) (.02) (.02) (.02) (.02)

- .09INTG310 - .O4INT6311 - .19INT6313 - .46INT6314
(.02) (.02) (.02) (.03)

and

(which is

.O4INT6316
(.02)

.211NT6320
(.02)

.OBINTG324
(.03)

.02INT6328
(.02)

.111NT6332
(.03)

.13INT6337
(.03)

.23INT6341
(.02)

.061NT6345
(.02)

.OZINT6351
(.03)

.13INT6355
(.02)

.121NTG359
(.02)

= 1342

The variables were defined in the text

yield

producer

Therefore the

recorded by

R2:

equations,

group

located in

other

.O9INITG317 -

(.02)

.O7INT6321
(.02)

.17INT6325
(.02)

.O7INT6329
(.02)

.OOlINT6334 +

(.02)

.OO7INT6338 -

(.02)

.OZINTG342
(.02)

.08INT6346
.021NT6352
(.02)

.02INT6356
(.02)

+

+

+

+

.OOOO9INT651

(.08)

.60

and

the
the
area allocated
achieved by farmers Operating in

the 4th

farmers

159

.11INTG318 - .IOINTG319
(.02)
.02INT6322 + .04INTG323
(.04)
.OZINT6326 + .02INT6327
(.02) (.02)
.O7INTG33O + .121NT6331
(.02) (.03) -
.006INTG335 + .01INTG336
(.02) (.02)
.381NT6339 - .061NTG34O
(.03) (.03)
.07INTG343 + .0031NTG344
(.02) (.02)
.111NTG349 + .OSINTG350

.161NT6353
(.02)

.O4INTG357
(.03)

F =

(.03)

.OBINT6354
(.02)

+ .OIINT6358
(.02)

34.07

dummy variables for the 60th
4th zone were “left
to paddy rice and the yields

the

from which the area devoted to the same

are

problems of interpretation resulting

allowed

from

60th

zone) represent the standard

crop and

this

in both our area

producer group

the yield
to deviate.

choice and

160
the arbitrary character Of selecting one producer group as

a standard instead of another justify the transformation of
the two equations. The transformation will make the
region's average in area cultivated and yield be the
reference of the groups' and zones' variations. The

transformation is presented in Appendix C.3

161

C.3 Transformation Of the Dummy Variables' Coefficients

The transformation procedure takes advantage of the
notion that the average of indexes is equal to onel' In our
case, we ewant the region's average to be the base;
therefore, we restrict the sum of each coefficient and a
constant to be equal to 1. The value of the constant can be
derived as a linear combination of the initial coefficients.

The procedure is used here to transform the dummy

variables Of our area and yield equations.

C.31 Area Equation.

The area equation involveS' two categories of dummy
variables.

-The first category represents the different zones Of
our study. Since we have four zones and the last one DZ4
was used as reference, it does not appear in the original
equation.

-The second category of dummy variables represents the
55 producer groups covered by our study.

The transformation consists of deriving a constant for
each category of dummy variables and using these constants

to calculate the new coefficients.

 

1Suits uses seasonal indexes to illustrate the
application of his transformation method to exponential
demand functions. He follow the same rationale here. See
D.G. Suits, “Dummy Variables'.

162
C.311 Derivation Of the new coefficients of the dummy
variables representing the different zones.
We want a set of coefficients b*, = b, + h determined
SO that '3b*j = ebj + h = 1 or
erj T h) = 4 where

bj the estimated coefficient Of DZj

h

a constant

J = 1, 2, 3, 4.

Since ebj + h = ebJeh, the appropriate value for
h is given by

h = Ln [4/Zeb11

we have

4
Zkbj = [4.8549 + 1.8589 + .3606 + 1] = 8.0744
1

Therefore

h = Ln [4/8.0744] = Ln(.4954) = -.7024 -.70

The value of h is used to generate the new set Of
coefficients for the dummy variables representing the

different zones:

b1=b1+h

6*, - 1.55 - .70 = .83

b*2 = .60 - .70 = -.10

6*3 = -1.02 - .70 = -1.72

*

b 3 -. =...
4 0 70 70

Each of these new coefficients represents the amount by
which the Specific effect of the corresponding zone deviates

from the average effect Of the region.

163
C.312 Derivation of the new coefficients of the dummy
variables representing the different producer groups.

The procedure is similar to the one used to get the
transformed coefficients Of the DZ,'s. The new set Of
coefficients is defined as:

Ck* = Ck + l and satisfies

eIck T I) = 56 where

Ck

the estimated coefficient of the group 06,
l

a constant
k = 1, 2, 3 . . . 55.
The appropriate value for l is given by
1 = ln [56/2'561']
we have:
56eCk = 106.8785
k=1
It follows that:
l = ln [56/106.8785] = ln(.5239) = -.6463 2 -.65
This computed value Of the constant l is used to get

the transformed coefficients for the dummy variables

representing the producer group.

C.313 The Transformed Area Equation

 

LSUP = -3.03 + .22 IMODM + .85DZl - .10DZ2 - 1.72D23
- .7ODZ4 + .O4D62 + 27063 - .04064 + .02065 + .43066
+ .05D67 + .02068 + .10069 + .210610 + .32D611

- .510613 - .970614 - .06D616 + .030617 + .080618

- .36D619 - .150620 - .17D621 + .080622 + .550623

- .03D624 + .310625 - .220626 - 1.360627 - 1.76D628

164
+ .080629 - .830630 - .76D631 - .58DG32 - 1.06D633
- .860634 - .700635 + .120636 + .230637 + .310638

+ 1.890639 + .88DG40 - .420644 - .570042 - .17DG43
- .OlDG44 - .17DG45 - .030646 - .94DG49 - .630650

- .350651 - .170652 - .150G53 - .480654 - .240655

- .590656 - .410057 - .740658 - .520659 - .650660.
R2 = .85

The sum of h and l is subtracted from the estimated
intercept of the area equation. The new intercept is given
by: -1.68 - 1.35 = -3.03.

0.32 Transformation of the yield Equatoin's dummy
variables.

The yield equation includes dummy variables
representing the four zones of the study. These variables
are transformed below according to the same principles used
for the area equation.

He ‘want a new set of coefficients b*i = b-

,+k

determined so that:

ekéebj = 4. It follows that:

k gln (4/Zebj)

He have

ZIeb1 = [ .6376 + .7483 + .7483 + 1] = 2.1342

The transformed values of the DZ,'s coefficients are
then computed

‘i = 045 + .63 = 018
b 2 = -929 + 063 3 034

165

-.29 + .63 .34

‘k
b 3
*

b4=0 +063

.63

The value Of k is subtracted from the estimated
constant term to get the new intercept-standard
correSponding to the region's average.

6* = 4.89 - .63 = 4.26

Finally, Since the new coefficients were derived as a
combination of the initially estimated parameters, their
standard errors can be calculated from the variance-
covariance matrix of the estimated coefficients. But,
because Of the large number of computations required, we did

not calculate the standard errors of the transformed

coefficients.

SELECTED BIBLIOGRAPHY

Afrique Agriculture. "L' Autosuffisance Ahimentaine: Le cas
du Perimetre Irrigue de Mbane au Senegal.“ NO. 100
December, 1983.

Askari, H. and Cumming, J.T. ”Estimating Supply Response
with the Nerlove Model: A Survey.“ International
Economic Review. Vol. 18, No. 2. June 1977.

 

 

Bonnefood, P., and Caneil, J. "Elements Pris en
Consideration Por Caracteriser les Systemes de
Production et leur EDVironnement dans la Vallee du
FTeuveJ1r DepartementHrEconomie et de Sociologie
Rurales de L'ISRA. No. 1, Octobre 1978.

 

 

Bonnefond, P; Caneil, J.; Auriol, O; Ndiaye, M; Menvielle,
J; Clement, A. “Systemes De Culture Irriguee et
Systemes de Production Paysons sur Ta Rive Gauche du
Fleuve SenegalTT Dakar, ORSTOM, 1981.

 

 

 

Boutillier, J.L.; Cantrelle, J; Causse, J.; Ndoye, L.T. “L3
Moyenne Vallee du Senegal. Etude SociO-Economique."
INSEE, Paris, 1982. '

Casey, Clayde Franklin. “The Soninke Experience with
Irrigation Development: Its Transferability to the
Senegal River Basin.“ M.S. Thesis, Cornell Univeristy,
1983.

Center for Research on Economic Development.“ Les Effets des

 

Politiqyes Aguicoles sur la Consommation Alimentaire:
Cameroun et enegal.T' Ahn Arbor, The UhTVersity of
Michigan,*l982.

 

deJanvry, Alain. “The Political Economy of Rural
Development in Latin America: An Interpretation“ in
Agricultural Development in the Third Horld. Carl
K. Eicker and J.M. Staatz, eds., 1984.

Diallo, Mamadou Dian. "Comparative Analysis Of Capital and
Labor Intensive Rice Irrigated Perimeters in the
Senegal River Valley.“ Plan 8 Paper, Michigan State
Univeristy, 1980.

Dielman,Terry E .“Pooled Cross-Sectional and Time series
Data: A Survey Of Current Statistical Methodology."
The American Statistician.VOl.37,NO.2.May
l983.pp111-122i

166

167

Diemer, G.; van der Laan, Ellen, C.H. "Small-scale
Irrigation Along the Senegal River“ Overseas
Development Institute. Network Paper 88. October
1983.

 

 

Crawford, Eric and Kelly, Valerie. A Field Study of
Fertilizer Distribution and Use In Senegal, 1984:

Simmary Report. B.A.M.E2 November, 1984. Preliminary
DTaft.

Dumont, R. False Start in Africa. New York: Praeger 1966.

 

 

 

---------4. Le Defi Senegalais: Reconstruire Les Terriors,
Libeger Eés Paysans. Series Etudes et Recherches.
No. 74, 1982.

Dumont, R., and Mottin, M.F. L'Afrique Etranglee: Zambie,
Tanzanie, Senegal, Cote D‘Ivoire, Buinee—Bissau, Cap
Vért. Paris: *Editibns du seiul, 1980.

 

Eicher, Carl K. Faire Face a la Crise Alimentaire de
L'Afrique. MSU Ihternational DeveTOpment Papers, No. 8,
1983.

 

 

Eicher, Carl K. and Baker, Doyle C. Research on Agricultural
Deyelopmentln Sub-Saharan AFrica. InternationaT
Development Paper. No. 1, 1982.

 

Eicher, Cark K. and Staatz, J.M., eds. Agricultural
Development in the Third Horld. Baltimore, John
HOpkins UniveriSty Press, 1984.

 

Faye, Jacques, and Niang, Madicke. “An Experiment in
Agragian Restructuration and Senegalese Rural Space
Planning.“ African Environment. 2(4) and 3(1), 1977,
pp. 143-53.

 

Food and Agriculture Organization Of the United Nations
(F.A.O.). Food Balance Sheets 1975-1977 Average and
Per Caput Food Supplies 1961-65. Average 1967 to
1977. Rome, 1980.

Fresson, Sylvianne. “Public Participation on Village Level
Irrigation Perimeters in the Matam Region of Senegal.“
Experiences in Rural Development, Occassional Paper
No. 4, Paris: OECD Development Centre, April 1978.

Foote, Richard. “Analytical Tools for Studying Demand and
Price Structures.“ Agricultural Handbook No. 16.
United States Department of Agriculture, Hashington,
0.6., August 1958.

168

Founou, Tchigoua. "Strategie de L'autosuffisance
Alimentaire et Choix d'une Cereale Prioutaire au
Senegal“ Paper preesnted at the CODESERIA - CAFRAD
meeting on Integrated development in Africa. June,

1980.

Gaye, Matar. “The Food Challenge in the Senegalese Rural
Economy. Analysis of the Domestic Cereals Production
Policy.“ Plan 8 Paper, Michigan State University, 1983.

Houck, James P. and Gallagher, Paul H. "The Price
Responsiveness Of U.S. Corn Yields.“ American Journal
Of A ricultural Economics Vol. 58 Pt. 2. 1978,
pp. 31-34.

Johnson, G.L. “Agricultural Production Economics" 805 course
notes, 1984.

Kennedy, Peter. A Guide to Econometrics. The MIT Press.
Cambridge, Massachusetts. 1979.

Koutsoyiannis, A. Theory of Econometrics. An Introductory
Exposition of Econometric Methods. Second Edition.
The MacMillan Press LTD., 1983.

Lam, Aboubacry Moussa. "Etre Paysan Aujourd'hui dans la
Region de Fleuve.“ le Monde Diplomatique, March 1983.

Leamer, Edward E. “Let's Take the Cons Out of
Econometrics." The American Economic Review. Vol. 73,

N00 1. MaFCh Iggy; pp. 30-130

Lele, Uma, and Candler Hilfred. "Food Security in
Developing Countries.“ ed. Carl K. Eicher and
J.M. Staatz.

Maina, Sanda. "On Food Security in Niger Republic: An
Economic Analysis of Millet and Sorghum Yield and
Acreage ReSponses.“ M.S. Thesis, Michigan State

Univeristy, 1982.

MBaye, Makhona. "Implementation and Evaluation of Rural
Development Projects Under Uncertainty: Hith a Spcial
Reference to the Sedhiou II Project (Casamance -
Senegal).“ Plan B paper, Michigan State University,

1983.

Morris, Michael. “The Fleuve Cereals Marketing Study: A
Summary Description.“ Title III Seminar paper Michigan
State University, 1985.

169

Mellor, John H. "Food Price Policy and Income Distribution
in Low Income Countries." ed. Carl, K. Eicher, and
J.M. Staatz, 1984.

N'Doye Ousseynou. “An Econometric Study of Supply and
Demand For Peanuts and Peanut Oil: Implications for
Senegalese AGricultural Polciy.“ M.S. Thesis, Kansas
State University, 1984.

Nerlove, Marc. “Estimates of Elasticities of Supply of
Selected AGricultural Commodities.“ Journal of Farm
Economics. Vol. XXXVII. May 1956, pp. 496-508.

 

Newman, Mark and Crawford, Eric. “Macro-Economic Research
on Senegalese Agriculture: Key Issues and Propsed Plan
of Hork.“ BAME, 1984.

Newman, Mark; Crawford, Eric; Faye, Jacques. "Orientations
et Programmes de Recherches Macro-Economiques sur le
Systeme Agra-Alimentaire Senegalais.“ Draft, BAME,

July 1984.

OMVS. ”Etude Socio-Economique du Bassin du Bleuve

Senegal." VOI.TA}E, DAKAR, 1980.

Patterson, Hilliam McClinton. “Village Irrigation Systems
Along the Senegal River: A Comparative Analysis."
Ph.D. Dissertation, Cornell University, 1984.

Pindyck, Robet S., and Rubinfeld, Daniel L. Econometric
Models and Economic Forcasts. Second e6: McGrawTHill,

1981.

Ross, Clark. A Modeling of the Demand and Supply of Food
Grains in Senegal. Center for Research onTEconOmic
Development, DiScussion Paper NO. 85. The University
of Michigan. Ann ARbor, June 1980.

 

SAED. “Elements de Reponse au Memorandum 'Note de Synthese
Etablie a la suite a lo suite de la Reunion de
Concertation CCCE/FAC/IDA/US AID les 9 et 10 Mai 1983 a
Hashington'.“

Schmid, Alan A. Property, Power and Public Choice: An
Inquigy Into Law and Economics. New York: Praeger,
1978.

.......... . Political Economy of Public Choice, upcoming

Schmitz, Jean. “Le Fleuve (des) Discours.“ le Monde
Diplomatique, Mars 1983.

170

SENEGAL. Ministre du Developpement Rural. DGPA Direction
des Methodes et du Plan. Plan d'Investissement
Alimentaire 1977-1985. Dakar, February 1977.

 

 

Shaffer, James 0. “Food Systems Organization and
Performance: Toward a Conceptual Framework.“ AJAE,

Sow Alassane. "Government Food Import Policy: The Case of
Rice in Senegal.“ M.S. Thesis, Michigan State
University, 1983.

Suits, Daniel B. “Dummy Variables: Mechanics
V. Interpretation.“ Review of Economics and
Statistics. Vol. LXVI, No. 1 February 1984,
p. 175-186.

 

Sy Alioune, and Dembele Seydou. "Amenagements
Hydro-Agricoles Dans le Bassin Du Fleuve Senegal (Bilan
de la decennie 1970-1980 et Perspectives). O.M.V.S.,
1981.

Tiffen, Mary. “Using Indigenous Skills and Institutions in
Small-Scale Irrigation: An Example from Senegal:
Introduction and Project Background.“ Overseas
Development Institute. October, 1983.

 

Timmer, Peter C. "DevelOping a Food Strategy.“ ed. Carl
K. Eicher, and J.M. Staatz, 1984.

Timmer, Peter C.; Falcon, Halter P.; Pearson, Scott. Food
Policy Analysis. Baltimore and London: The John
HOpkins University Press. 1983.

 

Tuluy, Hasan Ahmet. "Comparative Resource Costs and
Incentives in Senegalese Rice Production." Food
Research Institute. Stanford Univeristy. 1978.

--------- . “River Basin Development Alternatives: The
Economics of Small-Scale Irrigation Along the Senegal
River. Ph.D. Dissertation. Fletcher School of Law and
Diplomacy, 1980.

U.S. AID/OMVS. Integrated Development Project.
No. 625-0621. DAKAR, October 1982.

Vincent, Harren H., ed. Economics and Management in

Agriculture. Prentice-Hall, Inc. Englewood Cliffs,

 

171

Heiler, Edward M. and Tyner, Hallace E. “Social
Cost-Benefit Analysis of the Ninga Irrigation Pilot
Project, Senegal." The Journal of Developing
Areas. N0. 15. July 1981. pp. 655-669.

 

Horld Bank Staff. “Methodologies For Measuring Agricultural
Price Intervention Effects." Horking Paper No. 394.
June 1980.

nxcwmnw STATE UNIV. LIBRARIES
IIIIIIII III IIII III IIIII III II III IIII II III II IIII IIIIII
31293106513132