 

~

an»... yrnuyunqqguculav-unruu "WW

   

.mo- 9

.
.
n
.

 

  

 

 

 

 

 

 

 

 

”£8I8

ﬂit/i,
5052/“ at

This is to certify that the
dissertation entitled

THE IMPACT OF SEASONAL CHANGES IN REAL INCOMES
AND RELATIVE PRICES ON HOUSEHOLDS' CONSUMPTION
PATTERNS IN BAMAKO, MALI

presented by

Oumou M. Camara

has been accepted towards fulﬁllment
of the requirements for the

 

 

 

Doctoral degree in Agricultural Economics
d" wMajor Professor' 5 @ature

Aug. "Ijgoof/

Date

MSU is an Afﬁrmative Action/Equal Opportunity Institution

..-.-.-._‘-.-.-.-.-.-.-,-.-.-.-.-.-.-,-r-.-.-.-.g.-.-,-.-.-,-.-.-.-,_.-.-‘-,_I-.-.-.-.-.-.-.-t-.-.-.-.-.-.-.-.-.-.-.- -.-~-

 

 

 

 

. LIBRARY l
Michigan State

University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

s/or c:/ClRC/DateDue.p65-p.15

 

 

THE IMPACT OF SEASONAL CHANGES IN REAL INCOMES AND
RELATIVE PRICES ON HOUSEHOLDS’ CONSUMPTION PATTERNS
IN BAMAKO, MALI

By

Oumou M. Camara

A DISSERTATION

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

DOCTOR OF PHILOSHOPHY

Department of Agricultural Economics

2004

ABSTRACT

THE IMPACT OF SEASONAL CHANGES IN REAL INCOMES AND
RELATIVE PRICES ON HOUSEHOLDS’ CONSUMPTION PATTERNS
IN BAMAKO, MALI

By

Oumou M. Camara _..

Mali’s market reforms, initiated in the 19803, have improved the production and
physical availability of cereals in most markets; however, economic accessibility
remains a problem for households partly because the reforms resulted in higher
and more variable food prices.

This study examines the impact of seasonal changes in real incomes,
proxied by real expenditures, and relative prices on households’ consumption
patterns in Bamako, Mali’s capital city, using the complete demand systems
approach and household-level panel data. The panel data used in this study is
from a survey undertaken in Bamako by the Direction Nationale de la Statistique
et de l’Informatique (DNSI).

The study is organized in three essays. The ﬁrst essay (chapter 2) presents
a descriptive analysis of seasonal expenditure patterns and nutrient availability for
households in Bamako. The results show that Bamako households’ real
expenditures vary considerably across seasons and that much of the observed
seasonal variation in expenditures can be attributed to changes in non-food

expenditures as food expenditures remain fairly stable across seasons. In

addition, the results indicate that Bamako households maintain their calorie
consumption during the year by making substantial changes in the consumption of
foods that contain essential micronutrients (i.e., meat, fish, and vegetables).

In the second essay (chapter 3), the Almost Ideal Demand System is
applied to a three-stage demand model for different seasons in order to estimate
the impact of seasonal changes in real incomes and relative prices on households’
consumption patterns in Bamako. The study ﬁnds that Bamako households’
consumption patterns are responsive to changes in real incomes and relative
prices in any given season and that there are seasonal changes in income and price
responsiveness for all the commodities in the three demand models. In addition,
the results indicate that Bamako households engage in food consumption
smoothing from seasonal shocks in real incomes at the expense of non-food
commodities, of non-staple foods, and through signiﬁcant substitutions among
and between broad commodity groups.

The third essay (chapter 4) examines the effects of seasonal changes in
real incomes and relative prices on the effective demand for nutrients for Bamako
households using Engel functions. The results indicate that the effective demand
for nutrients is responsive to changes in real incomes and relative prices and that
there is evidence of seasonal changes in income and price responsiveness. In
addition, the results show that the effective demand for vitamin A and minerals is
more responsive to changes in real incomes and relative prices than is the demand

for calories, more speciﬁcally, calories from staples.

Copyright by
OUMOU M. CAMARA
2004

To my Family

ACKNOWLEDGEMENTS

I would like to thank Almighty God for giving me the dedication and strength to go
through graduate school. I am deeply grateful for the sacriﬁces my family made so I can
be where I am today. In particular, I am thankful to my parents, Kaba Camara and Haby
Sylla, for teaching me that faith, hard work, and discipline are keys to living your life the
way you want to. I would also like to thank my very best friend, big brother Amadou
Camara, for his tough love and guidance throughout ﬁve hard years, and my little angel
“Baba” for his relentless optimism.

I express my profound gratitude to the many individuals who have contributed
their time and expertise to this project. I am heavily indebted to an exceptional teacher,
John Staatz, my major professor, for his invaluable guidance and support from the ﬁrst
day I arrived at MSU. My special thanks to my committee members Eric Crawford,
Kelly Raper, and Jack Meyer for their patience and whose helpful comments and
suggestions improved the quality of this work.

I would also like to express my deepest gratitude to the 40 households in Mali
whose generous cooperation in sharing detailed economic data made this study possible.
I can just hope that some day I will be able to tangibly contribute to improving the life of
at least one individual in my country. My sincere thanks to Arouna Koné and the
Direction Régionale du Plan et de la Statistique (DRPS), Nango Dembelé and the Projet
d'Appui au Systeme d'Information Décentralisé du Marché Agricole (PASIDMA),
Lamine Keita and the Centre d’Analyse et de Formulation de Politiques de
Développement (CAFPD), the ofﬁce of the Provost at MSU, and the United States

Agency for International Development (USAID) for making this study possible.

vi

TABLE OF CONTENTS

List ofTables......... .......................................................................................................... xii

List of Figures. ................................................................................................................ .xvi

CHAPTER 1- INTRODUCTION

1.]. Issues and Background ................................................................................... 1
1.2. Problem Statement and Knowledge Gap ........................................................ 2
1.3. Research Objectives ........................................................................................ 4
1.4. Research Questions and Hypotheses ................................................ 5
1.5. Methodological Framework ............................................................................. 6
1.6. Data ................................................................................................................. 8

1.6.1. Source ......................................................................... 8

1.6.2. Collection Procedure ........................................................ 9
1.7. Speciﬁc Types of Analyses Planned ................................................. 11

1.7.1. Seasonal Changes in Expenditure Patterns and Nutrient Availability
for Households in Bamako, Mali: A Descriptive Analysis .................. 11

1.7.2. Examining the Impact of Seasonal Changes in Real Incomes and
Relative Prices on Households’ Consumption Patterns in Bamako, Mali,
Using the Almost Ideal Demand System Model ................................ 12

1.7.3. Estimating the Effects of Seasonal Changes in Real Incomes
and Relative Prices on Households’ Demand for Nutrients in Bamako,

Mali .................................................................................. 14

1.7.4. Sensitivity Analyses ......................................................... 15

1.8. Conclusion ............................................................................... 15
References ................................................................................ 17
Appendix 1 .............................................................................. 21

vii

CHAPTER 2 - ESSAY l: SEASONAL CHANGES IN EXPENDITURE
PATTERNS AND NUTRIENT AVAILABILITY FOR HOUSEHOLDS
IN BAMAKO, MALI: A DESCRIPTIVE ANALYSIS

2.1 . Introduction ........................................................................... 27
2.2. Methodological Framework ......................................................... 29
2.2.1. The Complete Demand System Approach ................................. 29
2.2.2. The Data ......................................................................... 31
2.2.3. Computation of Variables ..................................................... 32
2.2.3.1. Consumption and Expenditure Aggregates .................. 32
2.2.3.2. Prices ............................................................. 32
2.2.3.3. Nutrient Availability .......................................... 33
2.3. Results ................................................................................ 33
2.3.1. Seasonal Changes in Relative Prices and Real Expenditures ....... 34
2.3.1.1. Seasonal Changes in Relative Prices ......................... 34
2.3.1.3. Seasonal Changes in Real Expenditures ...................... 41
2.3.2. Households’ Seasonal Expenditure Patterns ........................... 43
2.3.2.1. Expenditure Patterns: Food vs. Non-Food .................... 44
2.3.2.2. Food Expenditure Patterns ...................................... 46
2.3.2.3. Non-Food Expenditure Patterns ................................ 53
2.3.3. Seasonal Nutrient Availability .......................................... 59
2.3.3.1. Nutrient Availability ............................................. 59
2.3.3.2. Income and Nutrient Availability ............................. 62
2.3.3.3. Seasonal Fluctuations in Nutrient Availability .............. 62
2.3.3.4. Sources of Nutrients ............................................ 66

viii

2.3.3.5. The Cost of Calories ........................................... 69

2.3.4. Sensitivity Analysis .................................................... 73
2.4. Conclusions .......................................................................... 79
References ................................................................................. 80
Appendix 2 ................................................................................ 83

CHAPTER 3- ESSAY 2: ESTIMATING THE IMPACT OF SEASONAL
CHANGES IN REAL INCOMES AND RELATIVE PRICES ON HOUSEHOLDS’
CONSUMPTION PATTERNS IN BAMAKO, MALI, USING THE ALMOST
IDEAL DEMAND SYSTEM MODEL

3.1 . Introduction ......................................................................... 98
3.2. Methods ............................................................................. 100
3.2.1. Commodity Aggregates and Weak Separability ........................ 100
3.2.2. The Almost Ideal Demand System (AIDS) .................. 101
3.2.3. The Data .................................................................... .104

3 .3 . Empirical Results .................................................................. 105

3.3.1. Coefﬁcients ................................................................. 106

3.3.1.1. Stage I Coefﬁcients ................................................. 106

3.3.1.2. Stage II Coefﬁcients ................................................ 112

3.3.1.3. Stage III Coefﬁcients ................................................ 117

3.3.2. Income Elasticities .......................................................... 121

3.3.2.1. Stage I Income Elasticities: Food vs. Non-Food
Commodities .................................................................... 1 22

3.3.2.2. Stage 11 Income Elasticities: Staple vs. Non-Staple
Commodities ................................................................... 124

3.3.2.3. Stage III Income Elasticities: Rice vs. Other Staple
Commodities ................................................................... 1 26

3.3.3. Own and Cross-Price Elasticities ......................................... 126

3.3.3.1. Stage I Price Elasticities ............................................. 127
3.3.3.1 . 1. Uncompensated and Compensated Own-Price
Elasticities .............................................................. 1 27
3.3.3.1.2. Uncompensated and Compensated Cross-Price
Elasticities .............................................................. 129
3.3.3.2. Stage 11 Price Elasticities ............................................ 132
3.3.3.2. 1. Uncompensated and Compensated Own-Price
Elasticities .............................................................. 1 32
3.3.3.2.2. Uncompensated and Compensated Cross-Price
Elasticities ............................................................... l 33
3.3.3.3. Stage III Price Elasticities ........................................... 137
3.3.3.3.1. Uncompensated and Compensated Own-Price
Elasticities .............................................................. 137
3.3.3.3.2. Uncompensated and Compensated Cross-Price
Elasticities .............................................................. 1 39
3.3.4. Sensitivity Analyses ....................................................... 141
3.3.4.1. Effects of Changes in Households’ Real Incomes
on the Income Elasticity of Food ........................................... 142
3.3.4.2. Effects of Changes in Households’ Real Incomes
on the Food Price Elasticity ................................................. 144
3.4. Conclusions ........................................................................ 145
References ............................................................................... 148
Appendix 3 .............................................................................. 150

CHAPTER 4- ESSAY 3: EXAMINING THE IMPACT OF SEASONAL
CHANGES IN REAL INCOMES AND RELATIVE PRICES ON
HOUSEHOLDS’ DEMAND FOR NUTRIENTS IN BAMAKO, MALI

4.1 . Introduction ........................................................................ 153
4.2. Methods ........................................................ . .................... 1 56
4.2.1. Nutrient Demand Model ................................................... 156
4.2.2. Data ........................................................................... 158
4.3. Empirical Results .................................................................. 159
4.3.1. Nutrient-Income Elasticities .............................................. 160
4.3.1.1. Calories ............................................................... 160
4.3.1.2. Protein, Calcium, Vitamin A, and Iron ........................... 164

4.3.2. Nutrient-Prices Elasticities ................................................ 167
4.3.2.1. Rice Price Effects on Nutrient Availability ...................... 167
4.3.2.2. Millet-Sorghum Price Effects on Nutrient Availability ........ 168
4.3.2.3. Beef Price Effects on Nutrient Availability ...................... 170
4.3.2.4. Dry Fish Price Effects on Nutrient Availability ................. 171
4.3.2.5. Green Leaves Price Effects on Nutrient Availability ........... 172

4.3.3. Sensitivity Analyses ....................................................... 173
4.4. Conclusions ........................................................................ 178
References ............................................................................... 1 82
Appendix 4 .............................................................................. 184

xi

CHAPTER 5: SUMMARY, POLICY IMPLICATIONS, AND DIRECTIONS FOR
FUTURE RESEARCH

5.1 . Introduction ........................................................................... 194
5.2. Summary of Main Findings .......................................................... 195

5.2.1. Seasonal Changes in Expenditure Patterns and Nutrient Availability for
Households in Bamako, Mali: A descriptive Analysis ........................... 195

5.2.2. Estimating the Impact of Seasonal Changes in Real Incomes and Relative
Prices on Households’ Consumption Patterns in Bamako, Mali, Using the

Almost Ideal Demand System Model .............................................. 198

5.2.3. Examining the Effects of Seasonal Changes in Real Incomes and Relative

Prices on Households’ Demand for Nutrients in Bamako, Mali ................ 199
5.3. Policy Implications ................................................................... 200
5.4. Directions for Future Research ..................................................... 204
References .................................................................................. 206

xii

Table 1-1.

Table Al-l.

Table Al-2.

Table A1-3.

Table 2-1.

Table 2-2.

Table 2-3.

Table 2-4.

Table 2-5.

Table 2-6.

Table 2-7.

Table 2-8.

Table 2-9.

Table 2-10.

Table 2-11.

Table 2-12.

Table 2-13.

LIST OF TABLES

Summary of Data Needs and Availability ...................................... 10
Topics Covered by Questionnaires in Each Survey Round .................. 22
Sample Size ......................................................................... 23
Commodity Groups Deﬁnition .................................................. 24
Complete Demand System Approach ........................................... 30
Seasonal Changes in the Relative Prices of Cereals .......................... 37
Seasonal Changes in the Relative Prices of Key Foods ...................... 39

The Consumer Price Index (Year 1996 =100) and Percentage Change
across Seasons ..................................................................... 40

Monthly Mean Nominal and Real Expenditure per Adult Equivalent
(CF A Francs) and Seasonal Changes in Expenditure (%) by Income
Group ............................................................................... 42

Source of Income for the Head of Household by Season .................... 43

Weekly Mean Nominal Food and Non-Food Expenditure per Adult

Equivalent by Season (CFA F ranc/AE), Budget Shares (%), and
Percentage Change in Expenditures across Seasons (%) ..................... 45

Mean Weekly Expenditure (FCFA/AE) and Budget Share (%)
Allocated to Individual Food Commodities .................................... 46

Mean Weekly Expenditure (FCFA/AE) and Budget Share (%)
Allocated to Individual Food Commodities by Income Group ............. 48

Mean Weekly Expenditure (FCFA/AE), Budget Share (%), and
Percentage Change across Seasons ............................................. 49

Mean Weekly Expenditure by Season and by Income Group (FCFA/AE),
and Percentage Change across Seasons ........................................ 51

Mean Budget Shares by Season and by Income Group (%), and Percentage
Change across Seasons (%) ...................................................... 52

Mean Weekly Non-Food Expenditure (FCFA/AE) and Budget

xiii

Table 2-14.

Table 2-15.

Table 2-16.

Table 2-17.

Table 2-18.

Table 2-19.

Table 2-20.

Table 2-21.

Table 2-22.

Table 2-23.

Table 2-24:

Table 2-25.

Table 2-26.

Table 2-27.

Table 2-28.

Share (%) ........................................................................... 53

Mean Weekly Non-Food Expenditure (FCFA/AE) and Budget Share (%)
Allocated to Individual Non-Food Commodity Groups by Income
Group ................................................................................ 54

Mean Weekly Non-Food Expenditure (FCFA/AE) and Budget Share (%)
Allocated to Individual Non-Food Commodity Groups by Season and
Income Group ..................................................................... 55

Mean Weekly Non-Food Expenditure (FCFA/AE) by Season and Income
Group ............................................................................... 57

Mean Budget Share (%) Allocated to Individual Non-Food Commodity
Groups by Season and Income Group .......................................... 58

Daily Nutrient Availability per Adult Equivalent by Income Group and
Nutrient Adequacy Ratios (%) ................................................... 60

Protein Contributed by Major Food Groups (%) by Income Group ........ 61

Nutrient Availability, Nutrient Adequacy Ratios, and Percentage
Change in Nutrients Availability across Seasons ............................. 64

Calories Contributed by Major Food Groups (%) by Season and
Percentage Change Across Seasons ............................................ 66

Contribution of Meat and Fish to Calorie Availability (kcal/AE/day) and
Budget Shares (%) by Season ................................................... 66

Sources of Nutrients (%) by Income Group .................................... 67
Mean Daily Animal Protein Availability in Grams/AE/day and
Contribution of Speciﬁc Types of Meat and Fish to Animal Protein
Availability (%) by Season ...................................................... 69

Average Cost of Calories (FCFA/IOOO kcal) by Season and by
Income Group ..................................................................... 71

Average Price Paid Per 1000 Calories for Staples by Type of Purchase..72

Effects of Including Estimates of Nutrient Availability from Away-From-
Home Foods by Income Group ................................................. 76

Effects of Including Estimates of Nutrient Availability from Away-From-
Home Foods by Season .......................................................... 78

xiv

Table A2-1.

Table A2-2.

Table A2-3.

Table A2-4.

Table A2-5.

Table A2-6.

Table A2-7.

Table A2-8.

Table A2-9.

Table A2-10.

Table A2-11.

Table 3-1.

Table 3-2.

Table 3-3.

Table 3-4.

Table 3-5.

Table 3-6.

Table 3-7.

Mean Weekly At-Home Food Consumption (kg/AE) by Season
and by Income Group ............................................................ 86

Weekly Mean Food Items Consumption (kg/AE) by Phase ................. 87

Mean Budget Shares (%) Allocated to Individual Food Items Consumed
At-Home by Season and by Income Group .................................... 88

Mean Daily Nutrient Availability per Adult Equivalent by Season and by
Income Group ...................................................................... 90

Mean Nutrients Contributed by At-Home Foods (%) by Income Group..91

Nutrients Contributed by Major Food Groups (%) in August by Income

Group ............................................................................... 92
Nutrients Contributed by Major Food Groups (%) in November

by Income Group .................................................................. 93
Nutrients Contributed by Major Food Groups (%) in February by Income
Group ............................................................................... 94
Nutrients Contributed by Major Food Groups (%) in May by Income
Group ............................................................................... 95
Contribution of Food Commodities to Protein Availability in grams/AB
and Shares (%) by Season ........................................................ 96
Mean Budget Shares Allocated to Fruits, Nuts, and Dairy Products

At and Away From Home ....................................................... 97
Parameter Estimates and Chow Test Results for Stage 1 Model ........... 109
Parameter Estimates and Chow Test Results for Stage II Model. . . . . .....1 14
Parameter Estimates and Chow Test Results for Stage 111 Model. . . . .....1 19
Estimated Income Elasticities for Stage I, II, and III Models .............. 121
Stage I Compensated and Uncompensated Own-Price Elasticities ....... 127
Stage I Compensated and Uncompensated Cross-Price Elasticities... . 1 31
Stage II Compensated and Uncompensated Own-Price Elasticities ...... 133

XV

Table 3-8.

Table 3-9.

Table 3-10.

Table 3-11.

Table 3-12.

Table A3-1.

Table A3-2.

Table 4-1.

Table 4-2.

Table 4-3.

Table 4-4.

Table 4-5.

Table 4-6.

Table 4—7.

Table 4-8.

Table 4-9.

Table 4-10.

Table 4-1 1.

Table A4-1.

Table A4-2.

Stage II Compensated and Uncompensated Cross-Price Elasticities.....135

Stage III Compensated and Uncompensated Own-Price Elasticities. . ...138

Stage III Compensated and Uncompensated Price Elasticities..........140
First Scenario Base Parameters and Variables ............................... 142
Second Scenario Base Parameters and Variables ............................ 144
Deﬁnition of Commodities ..................................................... 151
Deﬁnition of Variables and Summary Statistics ............................ 152
Calorie-Income Elasticities by Season and for the Pooled Data ........... 161
Calorie-Income Elasticities by Food Source ................................. 162
Nutrient Income Elasticities by Season and for the Pooled Data... .......165

Elasticity of Demand for Nutrients with respect to the Price of Rice....168

Elasticity of Demand for Nutrients with respect to the Price of Millet-

Sorghum ........................................................................... 169
Elasticity of Demand for Nutrients with respect to the Price of

Beef ................................................................................ 171
Elasticity of Demand for Nutrients with respect to the Price of Dry

Fish ................................................................................ 172
Elasticity of Demand for Nutrients with respect to the Price of Green
Leaves ............................................................................. 173
Baseline Values ................................................................... 174

Effects of a 20 Percent Increase in Real Incomes in Percentage

Changes ............................................................................ 175
Effects of a 20 Percent Increase in Real Incomes on the Amounts of

Nutrients Available by Season and By Income Group ..................... 177
Summary Statistics .............................................................. 185
Nutrient Contributed by Major Food Groups (%)by Season ............... 186

xvi

Table A4-3. Nutrients Contributed by Speciﬁc Food Items (%) by Season ........... 187

Table A4-4. Nutrient Demand Estimates .................................................... 191

xvii

Figure Al-l.

Figure A1-2.

Figure 2-1.

Figure 2-2.

Figure 2-3.

Figure 2-4.

Figure 2-5.
Figure 2-6.
Figure 2-7.

Figure A2-1.

Figure 3-1.

Figure 3-2.

Figure 3-3.

LIST OF FIGURES
Map of Mali ........................................................................ 25
Map of Bamako ................................................................... 26
Average Retail Price of Rice in Bamako (CFA/KG) from August 2000 to
July 2001 ........................................................................... 35

Average Millet-Sorghum and Maize Retail Prices in Bamako (CPA/KG)
from August 2000 to July 2001 ................................................. 36

Seasonal Changes in the Relative Prices of Fish, Vegetables, and Oil. . .38

Weekly Mean Food Expenditure Levels (FCFA/AE) and Food Budget

Shares (%) by Income Group ................................................... 44
Distribution of Calorie Availability across Households by Season ........ 63
Average Cost of Calories (CFA Francs/1000 calories) ....................... 70

Average Price Paid for Rice (CFA Francs/kg) by Area of Residence. .....73

Agricultural Calendar in Mali ..................................................... 85

Three-Stage Budgeting Process for Urban Households in Mali ............. 101

Effect of Changes in Real Incomes on the Income Elasticity of Food
by Season ........................................................................... 143

Effect of Changes in Real Incomes on the Own-Price Elasticity of Food
by Season ........................................................................... 145

xviii

CHAPTER 1

INTRODUCTION
1.1. Issues and Background
A key outcome of the food policy reforms initiated in the 19803 in Mali (i.e., the 19805
Structural Adjustment Programs and the 1994 Franc CF A devaluation) was the
liberalization of the cereals markets. Rice and coarse grains prices were decontrolled in
order to stimulate agricultural production and reduce reliance on imported rice (Dembéle’
et al., 1999). The production and physical availability of cereals in most markets
improved with the reforms as producers responded to the higher prices that resulted from
the reforms. However, economic accessibility remained a problem, especially for low-
income urban households, partly because the higher food prices caused a decline in their
purchasing power as their money income remained ﬁxed (Tefﬁ et al., 1997).1 As a result,
the impact of higher food prices on urban consumption patterns was investigated in Mali
(Rogers and Lowdermilk, 1991, Reardon et al., 1994) in an effort to provide government
officials with information they needed to design food safety net programs to help low-
income urban households.

Furthermore, the market reforms also resulted in more variable food prices, since
prices were now determined by market conditions, while before the reforms, the
government ﬁxed ofﬁcial producer and consumer prices for cereals. Dembelé et a1.
(1999) indicated that coarse grain and rice prices have shown signiﬁcantly more temporal

variation following the devaluation.2 They found that the coefﬁcient of variation of

 

' Separation of consumers and producers is assumed for this study.

2 The Malian government ﬁxed official cereals prices. Following the reforms, cereals’ prices were allowed
to vary not only across time and space but could also depend on the quality of the grain, and both local
production and demand from neighboring countries (Dembelé and Staatz, 1999).

monthly prices for rice, millet and maize, increased from 7, 26, and 23 percent in the
1990-93 period to 12, 30, and 28 percent in the 1994-97 period. They also found that
cereal prices were now following a seasonal trend that reﬂected the agricultural calendar:
sorghum and maize prices begin to drop in November (harvest season) to reach their
lowest in December most-harvest season) and start increasing in January (planting
season) to reach their maximum in August (lean season) (Dembelé et a1. 1999).
Empirical evidence (e.g. Chambers, 1981; Sahn, 1989; Paxon, 1993) suggests that
seasonal variation in food prices largely inﬂuence the effective incomes and consumption
potential of households. However, the implications of seasonality in food prices for
households’ consumption patterns have not been explored in Bamako, Mali.
1.2. Problem Statement and Knowledge Gap
Seasonal changes in households’ real income have two major consequences. First, they
result in changes in the quantity (level) of food consumed in the household from one
season to another. For example, Dostie (2000) found that poor Malagasy households
could eat close to the nutritional threshold only after the harvest season. During the lean
season, poor households’ caloric intake declined by 5 percent (Dostie, 2000). Second,
they affect households’ seasonal consumption choices by altering the set of market
baskets they can afford. For instance, households in Madagascar were found to substitute
tubers for rice during the lean season when their real incomes were low and relative
prices were high and increase their consumption of rice during the post harvest season
when they had more purchasing power and relative prices were lowest (Dostie, 2000).
The stability of households’ real incomes from one season to another is an

important determinant of household food security, as it allows households to smooth their

consumption levels throughout the year. The design of safety-net programs to protect at-
risk households’ food entitlements requires substantial knowledge, both descriptive and
analytical, about households’ annual and seasonal food consumption patterns and on the
forces causing changes in those patterns. For example, knowledge of income elasticities
can help the government in its search for self-targeting mechanisms such as those based
on subsidies on “inferior” goods since the policy option of implementing general food
price subsidies entails high costs (e. g. government budget and producer disincentives)
(Timmer, 1979). 3 Consumer theory indicates that income elasticities are likely to vary
systematically with the income of the consumer and from one price environment (set of
relative prices) to another (Timmer, 1983). Most of the empirical evidence in Mali and
West Africa (Rogers and Lowdermilk, 1981, Savadogo et a1, 1999, and Reardon et a1,
1999) has addressed the question of whether the price elasticity of demand varies with the
level of income and not whether the income elasticity of demand for commodities varies
with the level of relative prices faced by households. This study postulates that the
consumption patterns of households in Bamako are responsive to changes in their real
incomes and that the income response of demand for commodities will change from one
season to another. This implies that the effectiveness of safety-net programs will depend
on the season considered. The impact of seasonal changes in real incomes and relative

prices on households’ consumption patterns has not been investigated in Bamako, Mali.

 

3 Commodities that are self-targeting (i.e., good mechanisms for transferring food to the poor) are those
whose consumption declines with increasing income (also referred to as inferior goods); i.e., they have
negative income elasticities.

1.3. Research Objectives

The general objective of this study is to investigate the impact of seasonal changes in real
incomes and relative prices on households’ consumption patterns in Bamako. The study
is based on the hypothesis that changes in the relative prices of commodities from one
season to another translate into seasonal changes in households’ real incomes, which in
turn cause households to change their consumption patterns. This study’s aim is ﬁrst to
examine households’ consumption patterns at four different periods (seasons) within a
year; second to compare how consumption patterns change across seasons; and third to
identify the factors that cause changes in those patterns. The speciﬁc objectives of the
study are:

1. To describe (i) seasonal changes in relative prices and households’ real
expenditures; (ii) households’ seasonal consumption patterns; and (iii) the sources
of main nutrients available for various socio-economic groups and across
seasons;4

2. To estimate (i) income elasticities of demand for various commodities and
commodity groups for different seasons in order to investigate whether there
exists, in the Malian context, any self-targeting foods; and (ii) own and cross
price elasticities for different seasons, thus under diverse economic conditions
(supply, stocks, relative price levels), in order to identify households’ seasonal

substitution among and between broad commodity groups.

 

4 The nutrient estimates represent nutrients in foods that are available for household consumption and not
actual nutrient intakes by individuals. They are derived from the at-home food consumption data and
exclude nutrients from the inedible or non-servable components of foods (e.g., bones) and losses from
trimming, cooking, plate wastage, and spoilage.

3. To investigate the impact of seasonal changes in households’ real incomes and
relative prices on the effective demand for nutrients in Bamako.
The study will be organized in the form of three essays:

1. Essayl: Seasonal Changes in Expenditure Patterns and Nutrient Availability

for Households in Bamako, Mali: A Descriptive Analysis

2. Essay 2: Examining the Impact of Seasonal Changes in Real Incomes and

Relative Prices on Households’ Consumption Patterns in Bamako, Mali,

Using the Almost Ideal Demand System Model

3. Essay 3: Estimating the Effects of Seasonal Changes in Real Incomes and

Relative Prices on Households’ Demand for Nutrients in Bamako, Mali.
1.4. Research Questions and Hypotheses
The problem and the knowledge gap discussed above raise the following research
questions: What is the impact of seasonal changes in households’ real incomes and
relative prices on their consumption patterns? More precisely:

Question 1. Do households’ consumption patterns differ across season?
Hypothesis. Staples dominate at—home food purchases for all income groups during the
entire year; however, households will increase their spending on non-staple commodities
(e.g., meat and ﬁsh and vegetables), and thus diversify their diets, during the harvest and
post-harvest seasons. The reason for this is that these are periods of greater grain
availability (and lower prices) in urban markets.

Question 2. Does the impact of changes in households’ real incomes and relative

prices on their consumption patterns differ across seasons?

Hypothesis. Households’ consumption patterns are responsive to changes in real incomes
and relative prices, and the income and price elasticity of demand for food and non-food
commodities will change from one season to another.
Question 3. Does the impact of changes in households’ real incomes and relative
prices on the effective demand for nutrients differ across seasons?
Hypothesis. Bamako households’ demand for nutrients is responsive to changes in their
real incomes and relative prices and that the magnitude of the nutrient income and price
elasticities will change from one season to another.
1.5. Methodological Framework
The framework chosen for this study is the Complete Systems Approach (CSA) to
estimate demand equation parameters. The process of using complete demand systems in
policy analysis can be separated into three parts:
a) The choice of the appropriate complete demand system to be used;
b) The adaptation of the estimated demand model, to permit development of an
empirical framework so the policy issue can be addressed; and
c) The use of an elasticity matrix to answer problems from a price and/or quantity
dependent perspective (Raunikar and Huang, 1987).
The complete systems approach entails estimating a set of demand equations that
result from allocating total expenditure among a group of commodities. This approach
involves estimating an entire system of demand equations, one for each commodity, or

commodity grouping, deﬁned in the analysis:

XI! = f(plr,p21,. pm,yr) i=1’2’ °°°°°°°°°° ,0 (1)

................

Complete demand systems generate estimates of own and cross price elasticities
(compensated or uncompensated), income elasticities, and marginal budget shares for all
commodities in the set. The CSA provides information on the degree and nature of inter-
relatedness of commodities and the nature of the utility function (Raunikar and Huang,
1987). The theory of complete demand systems allows (1) the derivation of estimable
functional forms of demand equations from mathematically speciﬁed models of
consumer choice and (2) the imposition of constraints on demand parameters to reduce
the number of independent parameters to be estimated to manageable numbers relative to
the data available (Sadoulet and De Janvry, 1995).

The CSA allows incorporating the inherent simultaneity of consumer purchase
decisions across the spectrum of goods and services into the estimation process (Raunikar
and Huang, 1987). Policies that use prices as the instruments for change (prices are
controlled, and these changes affect the quantities purchased or consumed) are well
suited to being analyzed within a complete demand system framework (Raunikar and
Huang, 1987). The price of any speciﬁc commodity can affect the quantity demanded of
every commodity bought by the consumer and this simultaneity should be reﬂected in
policies that require the manipulation of commodity prices to produce changes in the
quantities demanded (Raunikar and Huang, 1987). The advantage of using this
framework is that these effects can be traced across all demand categories. Complete
demand systems include the translog system, the Rotterdam system, the addilog system,
the constant elasticity of demand system, the linear expenditure system, and the almost
ideal demand system (AIDS).

The AIDS model is chosen to estimate urban consumers’ demand functions in this

study. The AIDS is a demand system that is superior to its predecessors and is
recommended as a vehicle for testing, extending, and improving conventional demand
analysis for numerous reasons. First, the system is linear in the parameters and hence
simple to estimate. Second, the functional form is general and ﬂexible (Deaton and
Muellbauer, 1980b, p74). Third, the model is the most satisfactory in terms of being able
to test the restrictions of adding up, homogeneity and symmetry through linear
restrictions on ﬁxed parameters. Since Deaton and Muellbauer (1980) proposed the AIDS
model, it has been widely applied in many empirical studies of consumer behavior using
both cross-sectional and time series data (Green & Alston, 1990, Chen & Veeman, 1991,
Buse, 1994). Hence, part of the reason for the popularity of this demand system is due to
the considerable ease with which it can be estimated and used for testing the predictions
of consumer demand theory (Chambers and Nowman, 1997).

1.6. Data

1.6.1. Source

The panel data used in this study is from a 2000-2001 survey undertaken in Bamako by
the Direction Régionale du Plan et de la Statistique (DRPS) of the Direction Nationale de
la Statistique et de l’Informatique (DNSI) and the Projet d'Appui au Systeme
d'Information Décentralisé du Marché Agricole (PASIDMA) of Michigan State
University (MSU), the Assemble’e Permanente des Chambres d’Agriculture du Mali
(APCAM), and the Centre d’Analyse et de Formulation de Politiques de Développement
(CAF PD). The survey was conducted during the period August 2000 — May 2001. The
funding for this project was provided by USAID/Mali under the USAID-MSU Food

Security 11 Cooperative Agreement. The sampling frame was adapted from that

developed by the Direction Nationale de la Statistique et du Plan for the 1989 national
Enquéte Budget Consommation (Budget Consumption Survey).

The objective of the survey was to provide a detailed understanding of
procurement of food and non-food items in terms of type, quantities, source and
expenditure. Along with detailed information on consumption and expenditure, the
surveys also collected data on the demographic characteristics of households, their
educational and employment status and ownership of assets. Detailed information on the
data is available in Table Al-l of Appendix 1.

1.6.2. Collection procedure:

The DRPS administered the survey questionnaires to households gathered in
homogeneous functional entities called “unités alimentaires” or Food Consumption Units
(F CU). An F CU is deﬁned as a group of related individuals who share at least one meal
together per day (DNSI, 1991). The FCU could consist of one household that prepares
and consumes its meals alone, many households that prepare a common meal, or many
households that eat together separately prepared meals. An F CU could consist of one
individual, or a single conjugal family, or more than one conjugal family (DNSI, 1991).
Five “Sections d’Enumeration” (SE), geographical units that encompass 1000 to 1500
inhabitants in urban areas, were randomly selected and then 40 FCU were also randomly
chosen. One DNSI cartographer participated in delimiting the boundaries of the SE. A
pre-test was performed in six FCU on June 22nd, 23rd, and 24th of 2000 to check the
questionnaires’ adequacy.

During the month of July 2000, 40 enumerators were chosen and trained. Five

team chiefs were selected per SE to supervise the daily work of all the enumerators in

their SE. Three inspectors and two supervisors, Arouna Kone (the director of the DRPS)
and I, ensured that the questionnaires were properly ﬁlled out. During the data collection
week, each enumerator went to the FCU three times a day, before each meal was
prepared, to weigh food products and collect data on food at and away from home and
non-food expenditures.

The same households were interviewed in four rounds over a period of one-year
starting in August 2000 to May 2001 for the capital city, Bamako, in order to capture the
seasonal variation in consumption. There was no sample attrition. Data were collected for
seven consecutive days during each round. The four surveys covered 40 Food
Consumption Units (F CU), the sample size in each survey round being the same. Table
Al-2 of Appendix 1 contains detailed information on the sample size.

Table 1-1: Summary of Data Needs and Availability

 

 

Types of Analysis Variables Data Required Where are these data
Available?
Descriptive analysis Consumption and expenditure Quantities of food DRPS/PASIDMA/
of households’ Nutrients available Price per kg APCAM survey
consumption patterns Food composition tables OMA price data
DNSI price indices

 

Econometric analysis Share of budget devoted to Quantities of food, price DRPS/PASIDMA/
of demand for food speciﬁc food groups and non per kg, household size. APCAM survey
and non-food food items
commodities
Vector of household
characteristics, prices, price
index, total expenditure

 

Econometric analysis Nutrient availability Quantities of food DRPS/PASIDMA/
of nutrient demand Food prices, total expenditure, consumed converted to APCAM survey
and household size nutrients, food prices,

and number of adult
equivalents in each
household.

 

 

 

 

 

 

1.7. Specific Types of analysis planned

The impact of seasonal changes in real incomes and relative prices on households’
consumption patterns in Bamako will be examined using three complementary methods
of analysis. First, a descriptive analysis will be conducted on the (i) seasonal changes in
relative prices and households’ real expenditures; (ii) households seasonal consumption
patterns; and (iii) main nutrients’ availability from at-home food purchases, major
sources of nutrients, and the prices for kilocalorie across seasons and for various income
groups.5 Second, econometric analyses of the determinants of demand for food and non-
food commodities and nutrients will be performed for each season. Third, sensitivity
analyses will be performed to determine the effect of seasonal changes in real incomes
and relative prices on households’ budget allocation to various commodity groups or
items.

1.7.1 Seasonal Changes in Expenditure Patterns and Nutrient Availability for
Households in Bamako, Mali: A Descriptive Analysis

In this section, seasonal changes in expenditure patterns and nutrient availability for
households in Mali are examined through a descriptive analysis of (i) seasonal changes in
relative prices and real expenditures; (ii) urban households’ seasonal food and non-food
expenditure patterns; and (iii) seasonal availability of nutrients from at-home food
purchases. The descriptive analysis is essential for food policy purposes because it
provides critical information on the composition of households’ basket of goods and
services under different economic conditions (e. g., food supply stocks, relative prices)

and on the adjustments households make between and within food and non-food

 

5 The term “availability of nutrients” refers to nutrients in foods that are available for household
consumption through purchases and own-supply and not availability at the market level.

11

commodities across seasons. Table Al-3, in Appendix 1, shows the commodity groups
and speciﬁc items that will be included in the analysis.

1.7.2. Examining the Impact of Seasonal Changes in Real Incomes and Relative
Prices on Households’ Consumption Patterns in Bamako, Mali, Using the

Almost Ideal Demand System Model

In this essay, the Almost Ideal Demand System is applied to a three-stage demand
model for different seasons in order to estimate the impact of seasonal changes in real
incomes and relative prices on households’ consumption patterns in Bamako. The study
tests the hypothesis that households’ consumption patterns are responsive to changes in
their real incomes and that the relationship between household income and food and non-
food consumption patterns will change from one season to another. The study assumes
that consumers’ preferences are weakly separable in order to allow singling out and
studying only a small group of closely related goods. The reasoning behind the concept
of weak separability is that the optimization problem is intractable for the consumer if the
demand for every commodity is a function of the prices of all other commodities. To
simplify this problem, we may assume that the consumer partitions total consumption
into groups of goods, so that preferences within groups can be described independently of
the other groups.

Under the assumption of weak separability, the consumers’ simultaneous
decision-making process is broken into three steps by adopting a three-stage budgeting
process. In the ﬁrst stage, households allocate their total expenditures among seven broad
groups of commodities: (1) Food, (2) Durable Goods, (3) Semi-Durable Goods, (4)

Health, (5) Energy and Utilities, (6) Other Non-Durables (Hygiene and Tobacco), and (7)

12

Services. 6 In the second stage, households allocate their food expenditure on seven food
groups: (1) Staples, (2) Vegetables, (3) Meat and Fish, (4) Oil, (5) Sugar, (6) Other
Foods, and (7) Food Away From Home. In the third and ﬁnal stage, households allocate
their staple group expenditure to (1) Rice, (2) Millet-Sorghum, (3) Maize, (4) Wheat, and
(5) Roots and Tubers. Hence, it is thus assumed that preferences are weakly inter-
temporally separable, that food is weakly separable from non-food commodities and that
staples are weakly separable from the other food groups. It should be noted that weak
separability between the goods studied and the rest of a consumer’s bundle is generally
assumed before the empirical speciﬁcation, and not tested as a hypothesis. It is possible
to test for weak separability (Eales and Unnevehr, 1988; Salvanes and DeVoretz, 1997),
but it is hard to ﬁnd data sets of sufﬁcient size and richness that will allow this.

The study proceeds in three steps. First, the study estimates demand parameters
using the almost ideal demand system model for each season.7 Second, the analysis
computes income elasticities for different seasons in order to determine if Bamako
households’ consumption patterns are responsive to changes in their real incomes. Third,
the study computes own and cross-price elasticities for different seasons in order to
identify Bamako households’ seasonal substitution among and between broad commodity
groups. Finally, the study performs sensitivity analyses on the estimated income and price.
elasticities using several simulation scenarios.

The Chow Test, which is simply an F test, will be performed to test the hypothesis

of the constancy of the parameters of the demand system across seasons. The aim here is

 

6 Table AI-4 of Appendix 1 presents the deﬁnition of the various commodities and commodity groups.

7 The seasons are deﬁned, based on the pattern of agricultural activity in Mali, as follows: August = lean,
November = harvest, February = post-harvest, and May = planting.

l3

to determine whether the estimated demand parameters are stable over seasons
(corresponding to each survey round). The study will test for the stability of the
coefﬁcients under the null hypothesis that the estimated income elasticities do not vary
across seasons; thus that the impact of changes in urban households’ real income on

consumption patterns is assumed to be constant across seasons.

1.7.3. Examining the Effects of Seasonal Changes in Real Incomes and Relative
Prices on Households’ Demand for Nutrients in Bamako, Mali

This section examines the impact of seasonal changes in real incomes and relative prices
on households’ effective demand for nutrients in Bamako.8 This study is based on the
hypothesis that Bamako households’ demand for nutrients is responsive to changes in
their real incomes and relative prices and that the magnitude of the nutrient income and
price elasticities will change from one season to another. The study will address two
questions. The ﬁrst question pertains to whether the effective demand for nutrients is
responsive to changes in Bamako households’ real incomes and relative prices. The
second question regards whether the magnitude and sign of the nutrient-income and
nutrient-price elasticity of demand differs across seasons. These issues are analyzed using
Engel functions to estimate nutrient income and price parameters and the Chow test to
assess the stability of the estimated income and price coefﬁcients across seasons. The
demand functions are estimated by ordinary least squares for calories, protein, calcium,

vitamin A and iron for the pooled data and for each season separately.

 

8 The present study can not look at individual nutrient intake because the study has no information on the
quantities consumed by each individual in the household. In addition, the consequence of excluding food
away from home consumption is a systematic underestimation of households’ calorie availability.

14

1.7.4. Sensitivity Analysis

The impact of seasonal changes in Bamako households’ real incomes on the allocation of
expenditure among food and non-food commodity groups and on the effective demand
for nutrients is examined in this section by performing sensitivity analyses. Sensitivity
analyses provide simple demonstrations of how the demand model can be used to
simulate conditions in which alternative changes in relative prices and real income could
be evaluated and traced through the system to determine the effect on each food group or
item and on the demand for nutrients. Different scenarios (e. g., a 10 percent decrease or
increase in relative price levels of rice) are simulated by manipulating relative price
levels, budget shares, and real income and their effects on the reallocation of
expenditures and nutrient demand are traced.

1.8. Conclusion

The objectives of this study are: 1) to examine seasonal changes in expenditure patterns
and nutrient availability for households in Bamako, Mali; 2) to estimate the impact of
seasonal changes in real incomes and relative prices on households’ consumption patterns
in Bamako using the almost ideal demand system model; and 3) to identify the effects of
seasonal changes in real incomes and relative prices on households’ demand for nutrients
in Bamako. The study is organized as follows. The ﬁrst essay examines seasonal changes
in expenditure patterns and nutrient availability for Bamako households through a
descriptive analysis of (i) seasonal changes in relative prices and real expenditures; (ii)
households’ seasonal food and non-food expenditure patterns; and (iii) seasonal

availability of nutrients.

The second essay tests the hypothesis that the relationship between Bamako
households’ real incomes and relative prices and their consumption patterns differ across
seasons. This means that understanding the impact of seasonal changes in Bamako
households’ real incomes and relative prices on their households’ consumption patterns is
crucial to informed food policy making. Following the methodological framework
pioneered by Leser (1941, 1963), Stone (1959) and Frisch (1959), this essay uses the
Complete System Approach to examine the impact of seasonal changes in Bamako
households’ real incomes and relative prices on their consumption patterns.

The third essay uses Engel functions to examine the magnitude of the impact of
seasonal changes in households’ real incomes and relative prices on the effective demand
for nutrients. The study tests the hypothesis that households’ demand for nutrients is
responsive to changes in their real incomes and relative prices and that the magnitude of
the nutrient income and price elasticities will change from one season to another. This
means that improvements in households’ real incomes will lead to increases in the
quantity (i.e. calories) and the quality (protein, minerals, and vitamins) of food available
in those households and, thereby will be an effective mechanism in reducing malnutrition
in Mali.

Chapter 5 summarizes the main ﬁndings of each essay and provides a discussion
on the policy implications and the scope for future research on consumption analysis in

Mali.

REFERENCES

Abdulai, A, D.K. Jain and AK. Sharma, 1999. Household food demand analysis in India.
Journal of Agricultural Economics 50: 316-327.

Anderson, G and Blundell, 1983. Testing restrictions in a ﬂexible dynamic demand
system: an application to consumers’ expenditure in Canada. The Review of Economic
Studies , pp. 397-410.

Behrman, Jere R., and Deolalikar, Anil B., 1990. The Intrahousehold Demand for
Nutrients in Rural South India. The Journal of Human Resources, Vol. 25, No. 4 (1990),
pp. 665-696.

Blundell, R, P. Pashardes and G. Weber. 1993. What do we learn about consumer
demand patterns from micro data? The American Economic Review 83: 570-597.

Boughton, D., Reardon, T., 1997. Will the Promotion of Coarse Grain Processing Turn
the Tide for Traditional Cereals In the Sahel. Recent Empirical Evidence ﬁ'om Mali.
Food Policy 22(4), 307-316.

Buse, A. 1994. Evaluating the linearized almost ideal demand system. American Journal
of Agricultural Economics 76: 781-793.

Byme, Pl, 0. Capps, J r and A. Saha. 1996. Analysis of food-away-from-home
expenditure patterns for US. households, 1982-1989. American Journal of Agricultural
Economics 78: 614-627.

Carpentier, A. and H. Guyomard. 2001. Unconditional elasticities in two-stage demand
systems: an approximate solution. American Journal of Agricultural Economics 83: 222-
229.

Chen, P.Y. and M.M. Veeman. 1991. An almost ideal demand system analysis for meats
with habit formation and structural change. Canadian Journal of Agricultural Economics
39: 223-235.

Cox, TL. and MK. Wohlgenant. 1986. Prices and quality effects in cross-sectional
demand analysis. American Journal of Agricultural Economics 68: 908-919.

Dawson, P]. and R. Tifﬁn. 1998. Estimating the demand for calories in India. American
Journal of Agricultural Economics 80: 474-481.

Deaton, A and Muellbauer J. 19803. Economics and consumer behavior. Cambridge
Cambridge University Press

Deaton, A. 1990. Price elasticities from survey data. Journal of Econometrics 44 281—
309

Deaton, A and J. Muellbauer. 1980. An almost ideal demand system. The American
Economic Review 70: 312-326.

Deaton, A. 1997. The Analysis of Household Surveys: A Microeconomic Approach to
Development Policy. Baltimore, London: World Bank and the John Hopkins University
Press.

Deaton, A. “Quality, Quantity, and Spatial Variation of Price.” The American Economic
Review 78 (1988b)3: 418-430.

Deaton, A. 1992. Understanding Consumption. Oxford University Press.

Dong, D, J .S. Shonkwiler and O. Capps, Jr. 1998. Estimation of demand functions using
cross-sectional household data: the problem revisited. American Journal of Agricultural
Economics 80: 466-473.

Ealas, J .S. and L.J. Unnevehr. 1988. Demand for beef and chicken products: separability
and structural change. American Journal of Agricultural Economic 70: 521-532.

Edgerton, D L. 1997. Weak separability and the estimation of elasticities in multistage
demand systems. American Journal of Agricultural Economics 79 1, 62-79.

F afchamps, M. 1992. Cash Crop Production, Food Price Volatility, and Rural Market
Integration in the Third World. American J. of A gricultural Economics, Vol. 74, No. l
(1992), pp. 90-99.

Gabas, J-J. and Giri, J. (1987). The Food Situation in the Sahel: an Assessment. In
Cereals Policies in Sahel Countries. Acts of the Mindelo Conference, eds CILSS/Club du
Sahel/OECD, Paris.

Green R, and Alston J M. 1990. Elasticities in AIDS models. American Journal of
Agricultural Economics May 442—445.

Heckman J J. 1979. Sample selection bias as a speciﬁcation error. Econometrica 47
153—161

Heien, D. and CR. Wessells. 1990. Demand systems estimation with microdata: a
censored regression approach. Journal of Business and Economic Statistics 8: 365-371.

Karagiannis, G. and K. Velentzas. 1997. Explaining food consmnption patterns in
Greece. Journal of Agricultural Economics 48: 83-92.

Meenakshi, J .V. and R. Ray. 1999. Regional differences in India's food expenditure
pattern: a complete demand systems approach. Journal of International Development 11:
47-74.

Moschini, G. 1995. Units of measurement and the stone index in demand system
estimation. American Journal of Agricultural Economics 77: 63-68.

Moschini, G., D. Moro and RD. Green. 1994. Maintaining and testing separability in
demand systems. American Journal of Agricultural Economics 76: 61-73.

Nelson J A. 1991. Quality variation and quantity aggregation in consumer demand for
food. American Journal of Agricultural Economics, November 1204—1212.

Newbery, David M, 1989. The Theory of Food Price Stabilization. Economic Journal,
Vol. 99 (398) pp. 1065-82.

Park, J .L., R.B. Holcomb, K.C. Raper and O. Capps, Jr. 1996. A demand systems
analysis of food commodities by US. households segmented by income. American
Journal of Agricultural Economics 78: 290-300.

Pashardes, P. 1993. Bias in estimating the almost ideal demand system with the stone
index approximation. The Economic Journal 103: 908-915.

Perali, F. and J .P. Chavas. 2000. Estimation of censored demand equations from large
cross-section data. American Journal of A gricultural Economics 82: 1022-1037.

Pitt, M.M. 1983. Food preferences and nutrition in rural Bangladesh. The Review of
Economics and Statistics 65: 105-114.

Pitt, Mark M., Rosenzweig, Mark R., and Hassan, Md. Nazrnul. Productivity, Health,
and Inequality in the Intrahousehold Distribution of Food in Low-Income Countries.
American Economic Review, Vol. 80, No. 5 (December, 1990), pp. 1139-1156.

Reardon, T. Singare, K., Y. Camara, M. Wanzala, B. Teme, O. Sanogo. 1999. Household
consumption responses to the franc CFA devaluation: evidence from urban Mali. Food
Policy, 24(5), October: 517-534.

Rogers, BL. and Lowdermilk, M. 1991. Price Policy and Food Consumption in Urban
Mali. Food Policy_l 6, 461-473.

Sadoulet, Elisabeth, and de Janvry, Alain. 1996. Quantitative Development Policy
Analysis. Chapter 6. Baltimore, MD: The Johns Hopkins University Press, 1995.

Sen, Amartya. 1981. Poverty and Famines: An Essay on Entitlement and Deprivation.
Oxford: Oxford U. P., 1981.

Singh, 1., Squire, L., and Strauss, J. (eds.) Agricultural Household Models. Chapters 1
and 2. Baltimore, MD: The Johns Hopkins University Press, 1986.

Strauss, John and Duncan Thomas. 1995. Empirical Modeling of Household and Family
Decisions. In Handbook of Development Economics 3, ed. Behrman and TN. Srinivasan,
pp. 1 883-2023. North-Holland: Amsterdam.

Strauss, John and Duncan Thomas. 1998. Health, Nutrition and Economic Development.
Journal of Economic Literature. 36/2: 766-817

Tefﬁ, James F. Research Methods in the MSU Food Security in Africa Project:
Conceptualizing and Implementing Policy Relevant Studies. MSU International
Development Papers, Working Paper No. 38, 1990.

Teklu, Tesfaye, 1996. Food Demand Studies in Sub-Saharan Africa: A Survey of
Empirical Evidence. Food Policy, Vol. 21, No. 6, pp. 479-496, 1996.

Tifﬁn, Abigail and R. Tifﬁn. 1999. Estimates of food demand elasticities for Great
Britain: 1972-1994. Journal of Ag Econ 50: 140-147.

Tirnmer, C. P. and H. Alderman. 1979. Estimating consumption parameters for food
policy analysis. American Journal of Agricultural Economics 61: 982-987.

Tobin, James. 1958. Estimation of Relationships for Limited Dependent Variables.
Econometrica. 26: 26-36.

Yen, ST. and TL. Roe. 1989. Estimation of a two-Level demand system with limited
dependent variables. American Journal of Agricultural Economics 71: 85-99.

Zellner, A. 1962. An Efﬁcient Method of Estimating Seemingly Unrelated Regressions
and Tests for Aggregation Bias. Journal of the American Statistical Association 57:348--
368.

Zellner, A. and H. Theil. 1962. Three-Stage Least Squares: Simultaneous Estimation of
Simultaneous Equations. Econometrica 30:54--78. 3.

20

APPENDIX 1

21

Table Al-l: Topics Covered by Questionnaires in Each Survey Round

Topics
Socioeconomic status

Daily food at-home
consumption

Variables
Age
Gender
Education
Marital status
Source of revenue
Class of revenue
Profession
Household composition
Housing status
Ownership of assets
Years in Bamako
Access to basic infrastructure
Quantity (kg)
Price (FCFA)
Source (purchased, gift)
Market of purchase
Processing time
Individual consumption or not of prepared meals

 

Daily food away from home
consumption

Source

Type (street food vendors, restaurants, individual
home consumption)

Purchases

Unit (plate, spoon, kg)

Quantity

 

Daily non-food purchases

Price (F CFA)

Unit

Quantity

Household member incurring the purchase

Household member beneﬁting from the purchase

Type (health, hygiene, education, transportation service
and clothing)

 

Monthly expenditures recall

Payments for services, energy, sacs of grains, etc...
Household member incurring the purchase
Household member beneﬁting from the purchase
Price (F CFA)

Unit

Quantity

22

Table Al-2: Sample Size

 

 

 

Phase
L H PH P Avg Total
Total # AB 509 504 530 537 520 2080
# of individuals 664 660 695 706 681 2725
# of FCU 40 40 40 40 40 160

 

Note: The seasons are deﬁned as follows: L = August = lean season, H = November =
harvest, PH = February = post-harvest and P = May = planting. AB = Adult Equivalents;

F CU = Food Consumption Unit.

23

Table Al-3: Commodity Groups Deﬁnition

Commodity groups
Cereals

Speciﬁc Items
Rice, Millet-Sorghum, Maize
Wheat products, Other Cereals

 

 

 

 

 

 

 

 

 

Roots and tubers Atieké, Cassava, Potato
Sweet Potato
Meat and Fish Beef, Mutton, Poultry, Fish
Oil Peanut oil, Palm oil, Sheanut oil
Sugar
Vegetables Leaves, Okra, Onions, Tomatoes
Other vegetables: fresh
All other vegatbles
Milk, Dairy,and Eggs Butter, Buttermilk, Fresh milk
Condensed sweetened milk
Powdered milk, Eggs
Beverages Coffee, Lipton tea, Green tea
Quinqueliba
Fruits Banana, Lemon, Dates
Raisin, Citronella, Tamarind
Others Nuts, Seasonings, and Spices

 

Food Away From Home (FAF H)

Food purchased from street vendors and
Food purchased at restaurants

 

Durable goods

Household Appliances
Equipment for entertainment,
Education

 

Semi-Durable Goods

Clothing, Footwear, Books,
Newspaper, Magazines, Jewerly,
Watches, Toilet Articles, Cosmetics

 

Non-Durable Goods (excl. food)

Electricity, Brutane Gaz, Other Fuels,
Medical and Health Care,
Gasoline, Oil, Tobacco, and Hygiene

 

Services

Laundry, Domestic Services,
Other Household Services
Purchased Transportation,
Recreational and Cultural Services,
Communications.

24

Figure Al-l: Map of Mali

0 290 4? In

 

ﬁ

 

     
 
  

 

o 2110 460m
Western
Sahara I
ALG E R IA
Taoudennl
MAURITANIA
,krdar
Tombouctou
ENEGAL
,Kayas
Koulikoro NIGER
“WW0 BURKIIIA
FASO
GUINEA
BENIN
,.... roco
_ - core o'rvnlne Gm“ ] ugcertm

 

25

 

Al-2: Map of Bamako

Figure

 

 

—Bamako

 

 

 

26

CHAPTER 2
SEASONAL CHANGES IN EXPENDITURE PATTERNS AND NUTRIENT
AVAILABILITY FOR HOUSEHOLDS IN BAMAKO, MALI:
A DESCRIPTIVE ANALYSIS

2.1. Introduction
Malian agriculture has long been characterized by strong seasonal variations in
production, primarily because the country’s economy relies predominantly on rainfed
crop cultivation (F EWS, 2000).9 Prior to the 19808 market reforms, the Malian
govemment’s price stabilization schemes restrained how the volatility in output translated
into seasonal variation in grain prices (Dembéle’ et al., 1999). '0 However, a quite
different picture has emerged now that cereals markets are liberalized. Today, grain
prices are inﬂuenced not only by the seasonal pattern of production and availability but
also by regional and international supply and demand conditions and by the political
situation in neighboring countries such as Cote d'Ivoire (e. g., Tefﬂ et al., 1997; Dembélé
et al., 1999).

Despite the importance of seasonal grain price variation in Mali, measurements of
its immediate effects on households’ consumption patterns have been relatively scarce.
Thus far, the focus of policy and previous consumption studies (Rogers and Lowdermilk,
1991; Reardon et al. 1999) has been on the long-terrn adjustment of households’

consumption patterns to price and income changes. However, urban households are net

 

9 According to FAOSTAT, in 2001, out of Mali’s total agricultural area, 34,700,000 hectares, only 138,000
hectares (or 0.4%) were irrigated.

’0 Examples of such schemes include the Malian government ﬁxing of official producer and consumer
prices, through the ofﬁcial grain marketing agency (Ofﬁce Malien des Produits Agricoles (OPAM)),

for cereals and restrictions on inter-regional grain shipments (Dembelé and Staatz, 1999). However,
official prices were not available to everyone, as OPAM handled at most 40 percent of the country’s
marketed grain surplus and due to the illegal private trade of grains (Dembelé and Staatz, 1999).

27

food purchasers (Rogers and Lowdermilk, 1991). They earn cash income, allocate 54
percent of their income on food, and spend 40 to 50 percent of their food budget on
cereals (Rogers and Lowdermilk, 1991; Singare et al., 1996). Therefore, seasonal
variation in grain prices is likely to affect urban households’ ability to obtain adequate
food through the effects of grain price ﬂuctuations on consumers’ real incomes. Recent
studies (e.g. Chambers, 1981; Sahn, 1989; Paxon, 1993) have shown that the stability of
urban households’ real income from one season to another constitutes an important
determinant of household food security. According to the Food and Agricultural
Organization (F AO), food security remains a major problem in Mali. The FAO found
that, in 1999, the average annual caloric intake was in the order of 2073 kilocalories per
person per day (compared to the 2200 kcal minimum requirement), and the average
annual per capita consumption of cereals amounted to 155 kilograms (versus the 200
kilograms recommended amount) (FAO, 1999).

This essay is based on the hypothesis that seasonal variation in the relative prices of
commodities translate into seasonal changes in households’ real incomes, which in turn
will affect households’ seasonal consumption choices by altering the set of market
baskets they can afford. This would mean that seasonal changes in real incomes could
affect not only the quantity but also the quality of food consumed in households in any
given season. From a policy perspective, this implies that safety-net programs may be
more or less effective at different periods of the year, depending on the set of relative

prices faced by households and their real incomes at the time of their implementation.ll

 

” Safety nets are formal (e. g., food aid and consumption subsidies) and informal (e.g., extended family)
measures that help improve low-income households’ access to food. The national food security stock
program is an example of a current safety net program that is implemented in Mali. This program utilizes
early warning systems to target food to the food insecure population.

28

This study examines (i) seasonal changes in relative prices and real expenditures, (ii)
households’ seasonal expenditure patterns, (iii) seasonal availability of nutrients from at-
home foods, and (iv) the effects of including estimates of nutrient availability from away-
from-home foods on average daily total nutrient availability using sensitivity analyses.
The results will help close some of the knowledge gap in the food consumption literature
in Mali.

2.2. Methodological Framework

2.2.1. The Complete Demand System Approach (CSA)

The framework chosen for this study is the Complete Systems Approach (CSA). The
CSA describes the household’s budget allocation among a comprehensive set of
consumption categories. This framework takes into account the mutual interdependence
of large number of commodities in the choices made by consumers (Raunikar and Huang,
1987). Thus, the approach provides information on the degree and nature of
interrelatedness of commodities and allows incorporating the inherent simultaneity of
consumer purchase decisions across the spectrum of goods and services into the analysis
of households’ consumption patterns (Raunikar and Huang, 1987). Table 2-1, below,
depicts the commodity deﬁnitions used in this study.12 The groupings were chosen based
on our a priori knowledge about food consumption patterns among Bamako households
and in order to keep the number of non-consuming households for the groups to be very

small.13

 

'2 Housing expenditures are excluded from the analysis because over 90 percent of the sample households
own the dwelling in which they reside and do not pay rent. In this case, rental equivalents are potentially
inaccurate, and the beneﬁts of completeness need to be weighed against the costs of error (Deaton and
Zaidi (1999)).

'3 The method of commodity classiﬁcation in this study is as follows: ﬁrst, the classiﬁcation of goods and
services started with the identiﬁcation of 137 speciﬁc food items and 300 non-food items; second, these
goods and services were then aggregated into 12 commodity groups.

29

Table 2-1: Complete Demand System Approach

 

Major

Component

Commodity

Item

 

lFood

Rice

Rice

 

N

Other Staples

Millet-Sorghum
Maize

Wheat and Fonio
Atieke

Cassava

Potato

Sweet Potato

 

L»)

Meat and Fish

\OOO\IO\kJIJ>-L~)N.—

Beef

Mutton and Poultry
Dry Fish

Fresh F {sh

 

.32.

Vegetables

Leaves

Okra

Onions

Tomato (fresh and concentrate)

Other Vegetables
Beanstﬁssbsaéstiséz .................

 

Ur

Peanut Oil
Palm Oil
_S_heanut Oil

 

Others

 

Butter and Buttermilk

Fresh Milk

Condensed Sweetned Milk

Powdered milk

Eggs

Peanuts

Seeds

Coffee

Tea Lipton

Green Tea

Quinqueliba and other

Banana

Lemon

Tamarind

Other Fruits (Dates, Orange, Raisins)
5§§§9ﬂiﬂg§ﬂﬂﬂ§£§§§ ..................

 

00

Food Away From Home

 

Food Away F r9311 _H9_rp_e_ ________________

 

2 Non-Food

10
ll
12
l3
I4
15

I7

 

 

Education
Housewares
Personal Care
Health

Hygiene

Energy and Utilities
Tobacco
Transportation
Recreation

 

 

2.2.2. The Data

The panel data used in this study is from a 2000-2001 survey undertaken in Bamako by
the Direction Regionale du Plan et de la Statistique (DRPS) of the Direction Nationale de
la Statistique et de l’Informatique (DNSI) and the Projet d'Appui au Systeme
d'Information Décentralisé du Marché Agricole (PASIDMA) of Michigan State
University (MSU), the Assemblée Permanente des Chambres d’Agriculture du Mali
(APCAM), and the Centre d’Analyse et de Formulation de Politiques de Développement
(CAF PD). Data collection took place in Bamako, the capital city of Mali. The survey was
conducted in four rounds, one week in each quarter of the year, during the period of
August 2000 to May 2001. The four surveys covered 40 Food Consumption Units (FCU),
the sample size in each being the same. The same 40 households were tracked over time
and interviewed in all four periods. There was no sample attrition. Data were collected
for seven consecutive days during each survey round. The objective of the survey was to
provide a detailed understanding of urban households’ seasonal food and non-food items
procurements in terms of type, quantities, source and expenditure.

The survey was organized in four months of equal periods. The seasons were deﬁned
based on agricultural activity in Mali. Phase 1 (August) corresponds to the lean season,
Phase 2 (November) to the harvest season, Phase 3 (February) to the post-harvest season,
and Phase 4 (May) to the planting season.14 The harvest season extends from September

through November for millet, sorghum, and maize; from November through December

 

" The data collection week, within each survey month, was randomly selected in order to avoid bias
associated with a speciﬁc week. Furthermore, the distribution of expenditures across households was
closely examined, in the data cleaning process, in order to assess whether expenditures data collected in the
ﬁrst week of the month was higher for salaried households. The data did not provide any supportive
evidence with respect to such bias.

3]

for rainfed rice; and from October through November for irrigated rice. In a typical year,
cereals prices tend to fall during the harvest season, as surplus producers and traders
unload stocks in anticipation of a good harvest. However, if the season has been a poor
one, prices may remain high or even increase as stocks are withheld. The post-harvest
season extends from December through February and also corresponds to the cold
season. Cereals’ prices are generally lowest during the post-harvest season, as granaries
are full during this period and grain availability in urban markets is highest. The planting
season extends from May through July for millet, maize, and sorghum, from June-July
for rainfed rice and from October through December for irrigated rice. Farming activities
such as planting and weeding take place in this period. The hot season extends from
March through May. From this point on, grain stocks begin to gradually decrease and
reach their lowest levels during the lean season, also called the “hungry” season, which
occurs right before the ﬁrst harvest, primarily in August.

2.2.3. Computation of Variables

2.2.3.1. Consumption and Expenditure Aggregates

Following Deaton and Zaidi (1999), the food consumption in kilograms and expenditure
in CFA Francs, the non-food expenditure in CFA Francs, and the total expenditures in
CFA Francs aggregates were computed using the DRPS/MSU data. Detailed information
on the construction of the expenditure aggregates is provided in Appendix A2-l.

2.2.3.2. Prices

This study uses weekly cereals price data observed over the year 2000-2001 for the
capital city, Bamako. The weekly cereal price data for12 markets in Bamako was

obtained from the Mali Market Information System (MIS) called “Observatoire du

32

Marché Agricole” (OMA). The Consumer Price Index used in this analysis is from a
monthly report prepared by the Statistics Bureau, Direction Nationale de la Statistique et
de l’Informatique (DNSI), of Mali. The DNSI consumer price index is based on data
collected from surveys of households residing in the District of Bamako and is computed
using the Laspeyres methodology.

2.2.3.3. Nutrient Availability

The nutrient estimates were derived from the at-home food consumption data on the
quantities of food consumed and data on the nutrient composition of foods. ’5 Nutrient
values exclude nutrients from the inedible or non-servable components of foods (e. g.,
bones). The food quantities were converted into the edible portion using conversion
factors (called “refuse percentage”) computed by the USDA and found in the
“Composition of Foods Raw, Processed, Prepared” (USDA, 2003). Once the edible
portion was computed, the amount of nutrient in the edible portion was calculated using
the food composition table. Losses from trimming, cooking, plate wastage, and spoilage
are not accounted for in these values. The nutrient estimates computed this way represent
nutrients in foods that are available for household consumption and not actual nutrient
intakes by individuals.

2.3. Results

Seasonal changes in expenditure patterns and nutrient availability for Bamako households
are examined in this section through a descriptive analysis of (i) seasonal changes in
relative prices; (ii) households’ seasonal expenditure patterns; and (iii) seasonal

availability of nutrients. The analysis ends with sensitivity analyses on the estimates of

 

'5 The food composition data come from the food composition table for Mali prepared by Sundberg and
Adams (1998) and from the USDA’s Nutrient Data Bank System (2003).

33

nutrient availability in order to assess how the results would change when nutrient
estimates from away-from-home foods are taken into account.

The descriptive analysis is essential for food policy purposes because it provides
critical information on the composition of households’ basket of goods and services
under different economic conditions (e.g., food supply stocks, relative prices) and on the
adjustments households make between and within food and non-food commodities across
seasons.

2.3.1. Seasonal Changes in Relative Prices and Real Expenditures

The aim of this section is to examine seasonal changes in relative prices and Bamako
households’ real expenditures. First, the study uses the Observatoire du Marché Agricole
(OMA) price data to provide descriptive evidence on the seasonality of food prices.
Second, the study describes seasonal changes in the relative prices of all-items, food and
non-food components using the DNSI Consumer Price Index (CPI). Then, the CPI is used
to deﬂate households’ nominal expenditures in order to remove the effect of price
changes.

2.3.1.1. Seasonal Changes in Relative Prices

Figures 2-1 and 2-2, below, present the average price of rice, millet-sorghum, and maize
in Bamako markets from August 2000 to July 2001. First, the graphs show that rice is the
most expensive cereal, selling at an annual average retail price of 272 CF A Francs per
kilogram ($0.382). Millet-sorghum is the second most expensive staple, at 125 CFA
Francs/kg ($0.176). Maize is the least expensive cereal, with an average annual retail

price of 120 CFA Francs/kg ($0.168).

34

Second, Figure 2-1 shows that the price of rice was high during the lean
season, in August, averaging about 275 CF A Francs per kilogram, but reached its highest
level, 279 CF A Francs per kilogram, during the harvest season (November-December). In
a typical year, the price of rice tends to fall during the harvest season as surplus producers
and traders unload stocks in anticipation of a good harvest. However, during the year of
the survey, cereal production was estimated to be 18 percent below that of the previous
year (1999-2000) dueto lower rainfall and an outbreak of desert locusts that began in
October (FEWS, 2001). Thus, surplus producers and grain traders, expecting a bad
harvest, withheld stocks; which led to an increase in cereals’ prices prior to and during
the harvest season (FEWS, 2001). Nevertheless, as the new harvest began to reach urban
markets, the price of rice started to gradually decline and reached its lowest level, about

262 CFA Francs per kg in February, during the post-harvest season (December-

 

 

 

 

February).
Figure 2-1: Average Retail Price of Rice in Bamako (CFA/KG)
From August 2000 to July 2001

A 285 7- —

(D

i 280 -~7 7 7 —7- —---- ---——-——7—7--
_- , , _ I , ., _LL, -_ -_.__.,. Q

g 275

.5 270 +2 ~— 7~ Pee—*7 ‘7 ‘rv ‘ l

O

.5 265 4 — 7 7 77 --77 7 7 ~- —-————————————

260 TL—I—IﬁfT—T—‘T—Tﬁ—F—r—T—T—JlTlITIIIITTIIIjTTTﬁ—TITTjIIIIIIIFIIII

 

 

 

Aug Sept Oct Nov Dec Jan Mar Apr May Jun Jul
Source: Observatoire du Marche Agricole (OMA) data

Third, Figure 2-2, below, indicates that millet-sorghum and maize show similar
price movements across seasons. The price of millet-sorghum averaged around 135 CFA

Francs per kg during the lean season, in August, and began to gradually increase during

35

the harvest season, between September and early December, averaging about 137 CFA
Francs. The price of maize averaged around 135 CF A Francs per kg in August and,
contrary to millet-sorghum, began to gradually decline at the beginning of the harvest.
Maize prices declined sooner than millet-sorghum prices because maize is harvested
earlier, usually beginning in August. The average price of maize during the harvest
season was in the order of 125 CFA Francs per kg. Millet-sorghum and maize prices
dropped sharply in the middle of December, averaging around 110 and 108 CFA Francs
per kg, respectively, and remained low until mid-April. They begin to increase during the
planting season (May-July), averaging around 126 and 122 CF A Francs per kg. This was
partly due to depleted grain stocks, resulting in low food availability in urban markets, as
signiﬁcant coarse grain exports to a number of neighboring countries took place (e.g.,
millet supply in Bamako’s market decreased from 3,229 tons in January to 2,422 tons in
February (OMA, 2001)). In April 2001, 3000 tons of millet was exported to Burkina

F aso, 500 tons to Niger, and 250 tons to Cote d’Ivoire (OMA, 2001).

Figure 2-2: Average Millet-Sorghum and Maize Retail Prices
in Bamako (CFA/KG) from August 2000 to July 2001

 

160

150 +~ _ --_ __.-.-__
1404- , V; - II __________

130 q- ’ e'II‘k, .. m ,-L-._ MM l 1'
120 _g, ilk-(""777-5-77K. 7 ___ ..ﬁ.,H. _ I \Lm M
WI

110 --— ~~ »-~—~v~—-- » .-

100 4 e A- .1“... Lm _l/

90 Will lllIIIIITIIVTIITIIIIIIIIIIIIMIIIIIIIIIIIIIIII

Aug Sept Oct Nov Dec Jan Mar Apr May Jun Jul

 

 

 

 

 

 

Price (in FCFA/KG)

 

 

 

—°— Millet-Sorghum + Maize

Source: Observatoire du Marche Agricole (OMA) data

36

Table 2-2, below, presents the relative prices, as measured by the price ratio,
of rice and millet-sorghum, rice and maize, and millet-sorghum and maize. First, the
results indicate that Bamako households must give up, on average, about 2 kilograms of
millet-sorghum and maize for 1 kilogram of rice. Second, the results indicate that the
relative prices of cereals show substantial variations across seasons. For instance, the
relative price of rice and millet-sorghum decreased by 5 percent between the lean and
harvest seasons, increased by 28 percent between the harvest and post-harvest seasons,
and dropped by 17 percent between the post-harvest and planting seasons.

Table 2-2: Seasonal Changes in the Relative Prices of Cereals
Price Ratio % Change Between

L H PH P Mean H-L PH-H P-PH

Rice/Millet-Sorghum 2.041 1.944 2.483 2.051 2.130 -5 28 -17
Rice/Maize 2.054 2.239 2.527 2.149 2.242 9 13 -15
Millet-Sorghum/Maize 1.006 1.152 1.017 1.048 1.056 14 -12 3
Source: Observatoire du Marche Agricole (OMA) data

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting.

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2-3, below shows the relative prices, as measured by the consumer
price index, of ﬁsh, vegetables, fruits, and oil for Bamako markets from August 2000 to
July 2001. The price of vegetables increased between September and November, as most
horticultural goods, such as green beans and leafy vegetables, are planted during this
period. The growing season for vegetables corresponds to the cool dry season, which
extends ﬁ'om October to January. The price of vegetables started to gradually decline
from November until May, when they began to rise again. This is the period when most
horticultural crops are harvested.

Fish prices increased between August and September, dropped between

September and January, and remained fairly stable until June, when they begin to

37

increase again. Fishing activity in Mali largely depends on two hydrological seasons:
rainfall and river discharge (1RD, 2002). The Niger River usually ﬂoods in July, during
the rainy season. The ﬂood recedes between November and January. The ﬁshing
campaign usually begins then and ends when water levels are low, between March and
June.

Figure 2-3: Seasonal Changes in the Relative Prices
of Fish, Vegetables, Fruits, and Oil

 

 

 

3140*
2
I1120~ 'v/
3 .. “ ;:
2100~
380— ..
a 60*
U

40 ﬂ F l J I I T T 1 I f

 

 

 

Aug Sept Oct Nov Dec Jan Feb Mar Apr May Jun Jul
- * - Fish (Fresh) *Vegetables (Fresh) —‘- Fruits (Fresh) —*—'Oil

Source: Observatoire du Marche Agricole (OMA) data

Table 2-3, below, presents the relative prices, as measured by the price ratio,
of rice and beef, rice and green leaves, and beef and green leaves. 16 First, the results
indicate that Bamako households must give up, on average, over 3.5 kilograms of rice for
1 kilogram of beef, 1.4 kilograms of rice for 1 kilogram of green leaves, and 2.646
kilograms of green leaves for 1 kilogram of beef. Second, the results show that the
relative price of rice and beef is fairly stable across seasons. The biggest change, +3
percent, in the price ratio of rice and beef occurs between the post-harvest and planting

seasons, due to a drop in the price of beef over that period. Beef prices usually increase

 

'6 Unit values are used as proxy for prices.

38

during the cool dry season, as cattle weight drops due to decreased availability of grassy
vegetation and bushy plants in grazing areas. The relative price of rice and green leaves
and beef and green leaves both decreased by 44 percent between the harvest and post-
harvest seasons. This was due to a sharp drop in the price of green leaves, which was
caused by greater availability of these goods in Bamako markets. Between the post-
harvest and planting seasons, the relative price of rice with respect to green leaves and
beef with respect to green leaves increased by 51 and 47 percent, respectively. This was
due to lower beef prices due to the arrival of the rainy season, which usually occurs
between June and September, as animals gain weight. The price of green leaves also
dropped between this period.

Table 2-3: Seasonal Changes in the Relative Prices of Key Foods

 

 

 

 

 

 

 

 

 

 

Price Ratio % Change Between

L H PH P Mean H-L PH-HI P-PH

Rice/Beef 0.274 0.272 0.271 0.279 0.274 -1 0 3
Rice/Green Leaves 0.900 0.896 0.502 0.759 0.726 0 -44 51
Beef/GreenLeaves 3.278 3.293 1.851 2.716 2.646 0 -44 47

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting.

Table 2-4, below, presents estimates of the CPI for all items, food and non-food
major components, and percentage change in the CPI across seasons. The all-Items CPI
measures the average rate of price change for all goods and services purchased by
Bamako households from one point in time to another. The all-items CPI rose by 1
percent between August and November (lean-harvest) due to increases in the price of
food (1 percent), energy and utilities (1 percent), and education (1 percent). The higher
costs of food during this period can be largely attributed to increases in the price of
cereals (1 percent) as the price of ﬁsh, vegetables, and fruits decreased by 4, 5, and 26

percent, respectively. Cereal prices remained unusually high since the season had been a

39

relatively poor one compared to the record production yields of 1998 and 1999 (OMA,
2001). Yet, cereal prices started to gradually decline with the arrival of the new harvest
in urban markets.

Table 2-4: The Consumer Price Index (Year 1996 =100) and Percentage Change
across Seasons] 7

 

 

 

 

 

 

 

 

 

 

Components CPI (%) Percenta e C e Between
L H PH P H-L PH-H P-PH

All-Items 1 03 1 04 102 106 1 -2 4
Food 98 99 94 101 1 -4 8
Cereals (unprocessed) 87 88 86 99 1 -3 15

Fish (Fresh) 123 118 118 125 ' -4 0 5

Vegetables (Fresh) 105 99 69 71 -5 -30 2

Fruits (Fresh) 93 68 79 78 --26 15 0

Oil 114 114 113 112 0 -1 -1

Footwear and Clothing 105 105 108 110 0 2 2
Energy and Utilities 106 107 107 109 1 0 2
Housewares 110 110 106 , 107 .- 0 -3 0
Health 1 02 102 102 103 0 0 1
Transport 114 114 115 115 0 1 0
Recreation 99 99 100 100 0 1 0
Education 104 106 112 112 1 6 0
Other Goods and Services 114 113 115 114 0 1 0

Source: DNSI

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting.

As shown in Table 2-4, the price of cereals, vegetables, and oil fell by 3 percent,
30 percent, and 1 percent, respectively, between the harvest (November) and post-harvest
(February) seasons, resulting in a 5 percent drop in food prices. The lower price of
vegetables can be explained by the fact that the availability of horticultural products

substantially increases in Bamako markets during this period. The price of goods and

 

'7 The rate of change from the previous period is calculated as follows: Rate of change (%) = [(Et - Et-l)/
Et-l], where Et and Et-l are expenditure in the current and previous period, respectively.

40

services decreased by 2 percent over the same period due the lower food prices and a 3
percent decline in the price of housewares. Between February and May (post-harvest-
planting), the all-items CPI increased by 4 percent due to higher prices of food (+7
percent), footwear and clothing (+2 percent), energy and utilities (+2 percent), and health
(+1%). Higher food prices can largely be attributed to increases in the price of cereals
(13 percent), ﬁsh (2 percent), and vegetables (2 percent) over the same period. The
increase in cereal prices was partly due to the low grain availability in urban markets
following a period characterized by substantial coarse grain exports to a number of
neighboring countries.
2.3.1.3. Seasonal Changes in Real Expenditures
Table 2-5, below, reports average monthly and annual nominal and real expenditure per
adult equivalent (AE) by season and seasonal changes in real expenditure for the entire
sample and by income group.18 First, the results provide an indication of the poverty that
prevails in Bamako households, as their average annual real expenditures are in the order
of 280,154 FCFA/AB (US$392).19 Low-income households spend on average 184,495
FCFA/AB (US$258) annually, followed by 249,61 SFCFA/AE (US$349) for the middle
tercile, and 408,701 FCFA/AB (US$572) for the high-income groups.

Second, the results, presented in Table 2-4, indicate that Bamako households’

mean nominal expenditures vary considerably across seasons. Households mean

 

'8 The sample was divided into three income groups. The low-income group’s annual expenditures per
adult equivalent are strictly less than 212,000 FCF A. The middle income group’s annual expenditures per
adult equivalent are between 212,000 and 300,000 FCFA. The high-income group’s annual expenditures
per adult equivalent exceeded 300,000 FCFA. The adult equivalent scales used are: male>14 years=l.0,
female>l4years=0.8 and children=0.5 (Duncan, 1994).

'9 The exchange rates are $1=626 FCFA in August 2000, $1=769 FCFA in November 2000; $l=708
FCFA in February 2001; and $l=760 FCF A in May 2001(OANDA, 2003). The annual average exchange
rate was $1=711FCFA.

41

expenditures per AE are highest in August, the lean season, 33,471 F CFA/AB (US$54),
and lowest in May, the planting season, 18,793 FCFA/AB (US$24). Households’ mean
nominal expenditures decrease by 35 percent between the lean and post-harvest season
(August and November), increase by 2 percent between the harvest and post-harvest
season (November and February), and drop by 15 percent between the post-harvest and
planting season (February and May).

Table 2-5: Monthly Mean Nominal and Real Expenditure per Adult Equivalent

(CFA Francs) and Seasonal Changes in Expenditure (%) by Income Group
Income Phase % Change Between

Group L H PH P Avg Yearly H-L PH-H P-PH

Nominal Expenditure

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Low 25411 12642 12941 12349 15836 190030 -50 2 -5
Middle 33663 20468 17442 14128 21425 257104 -39 -15 -19
High 41323 32311 36425 30261 35080 420962 -22 13 -17
Mean 33471 21774 22149 18793 24047 288558 -35 2 -15
Real Expenditure
Low 24671 12156 12687 11650 15375 184495 -51 4 -8
Middle 32682 19681 17100 13329 20801 249615 -40 -13 -22
High 40120 31068 35711 28548 34058 408701 -23 15 -20
Mean 32496 20936 21714 17729 23346 280154 -36 4 -18
Difference
Low 740 486 254 699 461 5535 1 -2 3
Middle 980 787 342 800 624 7488 1 -2 3
High 1204 1243 714 1713 1022 12261 1 -2 3
Mean 975 837 434 1064 700 8405 1 -2 3

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting; Difference = Nominal Income — Real Income.

The seasonality of households’ expenditures can be partly explained by the fact
that households receive substantial ﬁnancial help from migrants, relatives, and the
extended family during the year. As shown in Table 2-6, below, the proportion of the
sample head of households’ real incomes that comes from remittances ranges from 22

percent in August and to 12 percent during the other seasons.

42

Table 2-6: Source of Income for the Head of Household by Season

 

 

 

 

 

 

 

 

 

 

 

Income Source Phase % Change Between
of Head of Household L H PH P Avg H-L PH-H | P-PH
Salaries 33 40 45 43 40 21 13 -4
Commercial activities 17 22 20 20 20 29 -9 0
Agricultural activities 5 5 5 8 6 0 O 60
Aid 22 12 12 12 15 -45 0 0
Other activities 23 20 17 17 19 -13 -15 O

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting

Third, the effect of seasonal price changes on households’ expenditures is
removed by deﬂating their nominal expenditures by the all-items CPI in order to assess
whether seasonal changes in the relative prices of goods and services are driving the
changes in total expenditures across seasons. The results, presented in Table 24, indicate
that only a small fraction of the observed seasonal variation in expenditures can be
attributed to seasonal changes in the relative prices of goods and services. During the
year of the survey, the all-items CPI rose by 1 percent between August and November
(lean-harvest), decreased by 2 percent between November and February (harvest-post-
harvest), and increased by 4 percent between February and May (post-harvest-planting).
2.3.2. Households’ Seasonal Expenditure Patterns
In this sub-section, a descriptive analysis of Bamako consumers’ expenditure patterns
with special emphasis on the differences observed between seasons and income groups is
performed. The allocation of households’ total nominal expenditures between and within
two major expenditure groups, food versus non-food, is closely examined for each season
and annually in order to uncover the source of the observed seasonal variation in

expenditures. Tables A2-1 through A2-4 of Appendix 2 provide detailed results on

43

Bamako households’ food consumption in kilograms per adult equivalent and food and
non-food expenditures in FCF A per adult equivalent by season and by income group.
2.3.2.1. Expenditure Patterns: Food vs. Non Food

Figure 2-5, below, shows the average weekly expenditure per adult equivalent on food
and non-food commodity groups by income group. Households allocate on average
annually 37 percent of their total budget to food (or 2201FCFA/AE) and 63 percent of
their total budget to non-food commodities (or 3810 FCFA/AB). The food budget share
declines with rising income levels (Engel’s Law): 47 percent, 41 percent, and 29 percent
among the low, middle, and high expenditure groups, respectively.

Figure 2-5: Weekly Mean Food Expenditure Levels (FCFA/AE) and
Food Budget Shares (%) by Income Group

 

 

 

 

%
3000 55
47 /2/- 2562
32:1 2000 4 E37.» ~- 188 r 45 +mdme
D \‘A
a. 41\\
B 1000 ~~ \\\\ l” 35 -+—Budget Share
"‘1 29 (sample avg =39%)
0 . 1 25
Low Middle High

Table 2-7, below, shows the average weekly mean nominal food and non-food
expenditure per adult equivalent by season, budget shares, and percentage change in
expenditures across seasons. The results indicate that much of the observed seasonal
variation in expenditures can be attributed to changes in non-food expenditures, as food
expenditures remain fairly stable across seasons. For instance, between August and
November, urban households’ average weekly expenditures on non-food commodities
decrease by 46 percent (from 5990 F CF A/AE ($9) to 3241 F CF A/AE ($4)), whereas food

expenditures decrease by 7 percent (from 2377 FCFA/AB ($4) to 2202 FCFA/AB ($3)).

44

Table 2-7: Weekly Mean Nominal Food and Non-Food Expenditure per Adult
Equivalent by Season (CFA Franc/AE), Budget Shares (%), and Percentage Change
in Expenditures across Seasons (%)

 

 

 

 

 

 

 

 

 

 

Commodities L H PH P Annual
Mean Expenditure/AB
Food 2375 2204 2101 2127 2202
Non Food 5991 3248 3467 2601 3827
Total 8366 5452 5567 4728 6028
Budget Share
Food 29 44 43 50 39
Non Food 71 56 57 50 61
Total 100 100 100 100 100
Percentage Change From

Nov Feb May
Food -7 -5 1
Non Food -46 7 -25
Total -35 2 -15

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting.

Between November and May, households reduce their non-food expenditures by
25 percent (from 3467 FCFA/AB to 2601 FCFA/AB) while food expenditures increase by
1 percent (from 2101 FCFA/AB to 2127 F CF A/AE). There are two possible explanations,
which are not mutually exclusive, for the observed seasonal variation in non-food
expenditures. The first is that households may attempt to smooth their food consumption
levels across seasons by incurring large changes in their non-food budget. This can be
explained by the fact that these households, especially poor households, consume near
subsistence levels of food, thus are more likely to make large cutbacks in their non-food
expenditures because this is the only way for them to maintain their food consumption
levels. However, the observed seasonal variation in non-food expenditures could also be
due to the seasonality of demand for non-food commodities. For instance, households’

expenditures on clothing and footwear are generally highest in August as they prepare for

45

the school year, which begins in September, and during periods of religious festivities,
such as the Tabaski. Similarly, households’ expenditures on traditional and formal health
services are high during the lean and planting seasons, as often-fatal illnesses such as
malaria and diarrhea are prevalent during those periods. Hence, the issue for households
could either be one of smoothing consumption in the face of variable income and/or one
of meeting seasonally high expenditure requirements in the face of relatively stable
income. One must keep in mind that, given the extreme poverty that prevails in Bamako,
households will have limited scope for discretion with respect to their spending.

2.3.2.2. Food Expenditure Patterns

Table 2-8, below, presents the mean weekly expenditure per AE and food budget share.
The three most important food commodities for Bamako households are staples, food
away from home, and meat and ﬁsh. Staples constitute the dominant part, 32 percent, of
the food budget share. However, rice expenditures alone account for about 21 percent of
the food budget, as Bamako households spend on average weekly 465 FCFA/AB ($0.65)
on this item.

Table 2-8: Mean Weekly Expenditure (FCFA/AE) and Budget Share (%) Allocated
to Individual Food Commodities

 

Commodities Expenditure Budget Share
(FCFA/AE/Week) (%)
Rice 465 21
Other Staples 242 11
Meat and Fish 353 16
Vegetables 272 1 2
Oil 76 3
Sugar 136 6
All Others Food At-I-Iome 238 11
Food Away From Home 419 19
Total 2201 100

 

Note: FAFH: Food Away From Home; Other Staples: Millet-Sorghum, Maize, Wheat,
Fonio, Sweet Potato, Potato, Atieke, Cassava; Others: Fruits, Beverages, Legumes, Nuts,

46

Seeds, Seasonings and Spices.

The second most important expenditure category is food away from home.
Bamako households allocate on average 19 percent (or 419 F CFA/AB ($0.59)) of their
food budget to food away from home. Away-from-home expenditures are those incurred
at restaurants, purchases from street vendors and foods purchased for individual
consumption. Street vendors are the most predominant source in the food away from
home category for all income groups mainly because they provide inexpensive,
accessible service and varied foods. The food away from home data indicates that a
substantial proportion, on average about 86 percent, of food away from home
expenditures, are incurred by household members who are employed, suggesting that
young children are largely excluded from this consumption. Also, the data shows that on
average about 20 percent of food away from home expenditures are made by the head of
household, while, on average, the household head accounts for only 6 percent of the
household population (The average household in Bamako is composed of 17 members).

Table 2-9, below, presents the mean weekly expenditure per AE and food budget
share by income group. Staples account for the largest share of the food budget for all
terciles, ranging from 36 percent for the lowest to 30 percent for the highest. The
proportion of the food budget devoted to staples in fact declines with rising expenditure
levels: 36 percent, 31 percent, and 30 percent for the low, middle, and high income
groups, respectively. In contrast, the proportion of the food budget devoted to meat and
ﬁsh, vegetables, and oil tend to increase with rising income levels. This is an illustration
of Bennett’s law, which holds that expensive sources of calories (i.e., meat and ﬁsh) are

substituted for cheaper ones (i.e., staples) with rising income levels. These results

47

suggest that Bamako households tend to diversify their diets as their income increase.

Table 2-9: Mean Weekly Expenditure (FCFA/AE) and Budget Share (%) Allocated
to Individual Food Commodities by Income Group

 

 

 

 

 

 

 

 

 

Expenditure Budget Share
Commodities (FCFA/AE/Week) (%)
Low Middle High Low Middle High—
Rice 454 462 477 24 21 19
Other Staples 226 210 293 12 10 11
Meat and Fish 250 317 495 13 14 19
Vegetables 216 278 324 12 1 3 13
Oil 65 72 92 3 3 4
Sugar 109 139 158 6 6 6
All Others Food At-Home 200 223 291 11 10 11
Food Away From Home 336 486 431 18 22 17

Total 1855 2188 2562 100 100 100
Note: FAF H: Food Away From Home; Other Staples: Millet-Sorghum, Maize, Wheat,

Fonio, Sweet Potato, Potato, Atieke, Cassava; Others: Fruits, Beverages, Legumes, Nuts,
Seeds, Seasonings and Spices.

Table 2-10, below, presents the mean weekly expenditure per AE, mean food
budget share, and percentage change in expenditures and the budget shares across
seasons. The results indicate that Bamako households make sizable changes in the
composition of their food basket across seasons. Between August and November (lean-
harvest), households’ increase their expenditures on rice and other staples by 6 percent
and 1 percent, respectively, while they reduce their expenditures on all other foods. In
terms of budget shares, the results show that households increase the proportion of their
food budget devoted to rice (+15 percent), other staples (+9 percent), and meat and ﬁsh
(+6 percent) while reducing the proportion allocated to all other foods over the same
period. The reduction in the consumption of other foods can be attributed to higher

vegetable prices that prevail during this period, as the availability of leafy vegetables and

many horticultural goods is low in Bamako markets during their growing season.

48

Table 2-10: Mean Weekly Expenditure (FCFA/AE), Budget Share (%), and
Percentage Change across Seasons

 

 

 

 

 

 

 

 

 

 

Phase % Change Between
Commodities L H PH P H-L | PH-H P-PH
Expenditure (FCFA/AE/Week)

Rice 457 485 476 440 6 -2 -8
Other Staples 242 243 222 262 1 —9 18
Meat and Fish 381 374 365 293 -2 -2 -20
Vegetables 320 266 261 242 -17 -2 -7
Oil 103 69 76 57 -33 9 -25
Sugar 146 132 125 140 -10 -5 11
Others 284 226 223 218 -20 -1 -2
FAFH 444 407 356 471 -9 -12 32

Total 2377 2202 2103 2123 -7 -5 1

Budget Share (%)

Rice 19 22 23 21 15 3 -8
Other Staples 10 11 11 12 9 -5 17
Meat and Fish 16 17 17 14 6 2 -2l
Vegetables 13 12 12 11 -10 3 -8
Oil 4 3 4 3 -28 14 -25
Sugar 6 6 6 7 -2 0 10
Others 12 10 11 10 -14 3 -3
FAFH 19 18 17 22 -l -8 31

Total 100 100 100 100 0 0 0

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting; F AF H: Food Away From Home; Other Staples: Millet-
Sorghum, Maize, Wheat, Fonio, Sweet Potato, Potato, Atieke, Cassava; Others: Fruits,
Beverages, Legumes, Nuts, Seeds, Seasonings and Spices.

Between November and February (harvest-post-harvest), with the exception of
oil, households reduce their expenditures on all food commodity groups. During this
period, households increase the proportion of the food budget allocated to rice (+3
percent), meat and ﬁsh (+2 percent), vegetables (+3 percent), oil (+14 percent), and other
foods (+3 percent) while they decrease the proportion spent on other staples (-5 percent)

and food away from home (-8 percent). Thus, Bamako households increase the

proportion of their food budget devoted to non-staple commodities, and diversify their

49

diets, only during periods characterized by low grain prices.

Between February and May (post-harvest-planting), households decrease the
proportion of the food budget allocated to rice (-8 percent) for the ﬁrst time during the
entire year and that of meat and ﬁsh (-30 percent), vegetables (-9 percent), oil (-22
percent), and other foods (~10 percent). At the same time, they increase the budget share
for other staples (+14 percent), sugar (+10 percent), and food away from home (+24
percent). These changes can be attributed to the high prices of grains, vegetables, and
ﬁsh, as shown in Table 2-4, that prevail during the planting season due to low food
availability. A possible explanation for the boost in the food share devoted to other
staples and sugar is an attempt by households to maintain their calorie levels by preparing
meals such as porridge, usually made with millet, sorghum, or maize ﬂour, that are
consumed in the morning and evening. The increase in the budget share of food away
from home (i.e. street vendors) may reﬂect the head of household “individualizing”
consumption in this period of high food costs by consuming foods that are too expensive
to provide to the entire household (Reardon et al., 1999).

Tables 2-1 1 and 2-12, below, present the mean weekly expenditure per AE, mean
food budget share allocated to food commodity groups, and percentage change in
expenditures and the budget shares across seasons by income group. The results also
indicate that the income groups show great similarities in their allocation of the food
budget among food commodities as the relative importance of foods in their diets remains
uniform. However, they do exhibit strikingly different adjustments in the proportion of
their food budget allocated to individual food commodities in any given season. For

instance, between August and November (lean-harvest), the biggest increase in the

50

proportion of the food budget devoted to individual food commodities is other staples (31

percent) for the low income group and food away from home (12 percent) for the high

income group.

Table 2-11: Mean Weekly Expenditure by Season and by Income Group
(FCFA/AE), and Percentage Change across Seasons
Expenditure (FCFA/AE/Week)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Phase % Change Between
Commodities L H PH P H-L PH-H P-PH
ce
Low 458 477 440 443 4 -8 1
Middle 465 488 490 405 5 0 -17
High 448 491 496 474 9 l -4
""Other Staples
Low 182 236 218 268 29 -8 23
Middle 25 1 192 176 222 -23 -9 27
High 292 306 275 300 5 -10 9
'"Meat and Fish
Low 268 272 227 232 2 ~17 2
Middle 350 270 388 259 -23 44 -33
High 527 587 477 389 11 -19 -18
Vegetables
Low 256 202 205 199 -21 l -3
Middle 314 275 288 234 -12 -19
ﬂiéh 391 320 288 295 -18 -10 2
Oil
Low 99 51 57 52 ~48 11 -9
Middle 85 67 82 55 -21 22 -34
High 127 89 88 65 -29 -2 -26
Sugar
Low 125 104 93 114 -17 -10 22
Middle 152 129 135 140 ~15 4 4
High 160 162 147 164 2 -9 11
""‘Xli Others
Low 239 197 175 189 -18 -ll 8
Middle 299 2 l 7 I97 180 -27 -9 -8
High 312 266 299 289 -15 12 -3
""""" F X15131
Low 395 285 253 409 -28 -ll 61
Middle 546 487 443 468 -l l -9 6
High 384 441 365 536 15 -17 47

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest and P = May =
planting; FAF H: Food Away F rom Home; Other Staples: Millet-Sorghum, Maize, Wheat, Fonio, Sweet
Potato, Potato, Atieke, Cassava; Others: Fruits, Beverages, Legumes, Nuts, Seeds, Seasonings and Spices.

51

Table 2-12: Mean Budget Shares by Season and by Income Group (%), and
Percentage Change across Seasons (%)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Budget Share (%)
Phase % Change Between
Commodities L H PH P H-L PH-H | P-PH
Rice
Low 23 26 26 23 16 1 ~12
Middle 1 9 23 22 21 22 ~3 ~8
High 17 18 20 19 9 10 ~7
Other Staples
Low 9 13 13 14 43 l 8
Middle 10 9 8 11 ~11 ~12 42
High 11 12 11 12 4 -2 6
Meat and Fish
Low 13 15 14 12 13 -9 ~10
Middle 14 13 18 13 ~l 1 39 ~25
High 20 22 20 15 10 ~11 ~21
Vegetables
Low 13 11 12 10 ~12 11 ~15
Middle 13 13 13 12 1 1 ~9
High 15 12 12 12 ~19 ~1 ~1
Oil
Low 5 3 3 3 ~43 21 ~20
Middle 3 3 4 3 -8 18 ~26
High 5 3 4 3 ~30 7 ~28
Sugar
Low 6 6 6 6 -8 ~2 7
Middle 6 6 6 7 ~1 1 16
High 6 6 6 7 1 -1 8
All Others
Low 12 11 11 10 -9 -3 -6
Middle 12 10 9 9 ~16 ~12 3
High 12 10 12 11 ~16 23 ~6
FAFH
Low 20 16 15 21 ~20 ~3 41
Middle 22 23 20 24 3 ~12 18
High 15 17 15 . 21 14 ~10 42

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting; FAF H: Food Away From Home; Other Staples: Millet-
Sorghum, Maize, Wheat, F onio, Sweet Potato, Potato, Atieke, Cassava; Others: Fruits,
Beverages, Legumes, Nuts, Seeds, Seasonings and Spices.

52

2.3.2.3. Non-Food Expenditure Patterns

Table 2-13, below, presents the mean weekly non-food expenditure per AE and non-food
budget share allocated to individual non-food items. The three most important non-food
commodities for Bamako households are personal care, health, and housewares.
Households allocate on average 20 percent (or 761 FCFA/AE/week ($1.07)) of their non-
food budget to personal care (i.e., clothing and footwear). Health expenditures include
expenditures on service items (i.e., formal public or private doctors, traditional healers
and pharmacists) and medicine. from both formal and informal sources and occupy on
average aboutl 8 percent (or 698 FCFA/AE/week ($0.98)) of the non-food budget.
Households allocate on average during the year 17 percent (or 642 FCFA/AE/week
($0.90)) of their non-food budget to housewares and to energy and utilities. Education
expenditures (i.e., school fees and supplies) occupy on average only 3 percent (or 129

F CF A/AE/week ($0.18)) of households’ non-food budget.

Table 2-13: Mean Weekly N on-F ood Expenditure (F CFA/AE) and Budget Share
(%)

 

Commodities Expenditure Budget Share
(F CFA/AE/Week) (%)
Education 129 3
Housewares 642 17
Personal Care 761 20
Health 698 18
Hygiene 169 4
Energy and Utilities 638 17
Tobacco 92 2
Transportation 55 7 1 5
Recreation 1 24 3
Total 3810 100

 

Note: Education (fees, school supplies); Housewares (cooking items, housing maintenance and repairs,
household appliances); Personal Care (clothing and footwear); Energy and Utilities (electricity, gas, wood,
charcoal), Health (medical and health care); Hygiene (soaps, cleaning supplies), Transportation (purchased
and private transportation, maintenance, repairs, insurance).

53

Table 2-14, below, presents the mean weekly non-food expenditure per AE and
non-food budget share by income group. The proportion of the non-food budget devoted
to education (2 percent for the low versus 5 percent for the high-income group),
housewares (14 versus 21 percent), health (11 versus 17 percent), and recreation (2
versus 4 percent) tends to increase with rising income levels. In contrast, the non-food
budget share allocated to personal care (24 for the low versus 16 percent for the high-
income group), energy and utilities (23 versus 16 percent), and transportation (18 versus
15 percent) decreases with higher income levels.

Table 2-14: Mean Weekly N on-F ood Expenditure (FCFA/AE) and Budget Share
(%) Allocated to Individual Non-Food Commodity Groups by Income Group

 

 

 

 

 

 

 

 

 

Expenditure Budget Share
Commodities (FCFA/AE/Week) (%)
Low Middle High Low Middle High
Education 33 58 302 2 2 5
Housewares 292 287 1374 14 9 22
Personal Care 592 685 1011 28 22 16
Health 218 897 964 10 28 16
Hygiene 98 151 261 5 5 4
Energy and Utilities 397 541 984 19 17 16
Tobacco 1 7 147 1 07 1 5 2
Transportation 394 362 930 19 1 1 15
Recreation 63 40 276 3 1 4
Total 2104 3168 6208 100 100 100

 

Note: Education (fees, school supplies); Housewares (cooking items, housing
maintenance and repairs, household appliances); Personal Care (clothing and footwear);
Energy and Utilities (electricity, gas, wood, charcoal), Health (medical and health care);
Hygiene (soaps, cleaning supplies), Transportation (purchased and private transportation,
maintenance, repairs, insurance).

Table 2-15, below, shows the mean weekly expenditure and budget share by
season and percentage changes in both across seasons. The results suggest that Bamako

households’ expenditures on many non-food goods and services tend to be highly

seasonal. For instance, households’ expenditures on traditional and formal health services

54

are high during the lean and planting seasons, as often-fatal illnesses such as malaria and
diarrhea are prevalent during those periods.

Table 2-15: Mean Weekly Non-Food Expenditure (FCFA/AE) and Budget Share
(%) Allocated to Individual Non-Food Commodity Groups by Season and Income
Group

 

 

 

 

 

 

 

 

 

 

 

Phase % Change Between
Commodities L H PH P H-L PH-H P-PH
Expenditure (FCFA/AE/Week)
Education 244 180 42 51 ~26 ~77 22
Housewares 1 1 59 687 442 279 ~41 ~36 ~37
Personal Care 1522 438 901 182 ~71 106 ~80
Health 1061 409 520 803 ~61 27 54
Hygiene 21 1 168 155 143 ~20 ~8 ~7
Energy and Utilities 827 546 641 538 ~34 17 ~16
Tobacco 79 120 74 93 51 ~38 25
Transportation 722 565 530 412 ~22 ~6 ~22
Recreation 166 127 129 75 ~23 2 ~42
Total 5990 3241 3434 2575 ~46 6 ~25
Budget Share (%)

Education 4 6 l 2 37 ~78 62
Housewares 19 21 13 11 9 ~39 ~16
Personal Care 25 14 26 7 ~47 94 ~73
Health 18 13 15 31 ~29 20 106
Hygiene 4 5 5 6 48 ~13 24
Energy and Utilities 14 17 19 21 22 11 12
Tobacco 1 4 2 4 1 80 ~42 67
Transportation 12 17 15 16 45 ~12 4
Recreation 3 4 4 3 41 ~4 ~23
Total 1 00 100 100 100

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting.

Similarly, mean weekly transportation expenditures are highest, 722 F CFA/AB
(US$1.16), during the lean season due to the fact that males oﬁen migrate to rural areas
because of the increased agricultural labor requirements during the harvest season.

Finally, households mean weekly expenditures on personal care are highest during the

55

lean (August), 1,522 FCFA/AB (US$2.44), and post-harvest season (February), 901
FCFA/AB (US$1.44). Expenditures on clothing and footwear are generally highest
towards the beginning of the school year, in August, and during periods of religious
festivities, such as the Tabaski, which occurred in February during the survey year.

Furthermore, the results also indicate that Bamako households’ expenditures on
education, health, and personal care vary considerably across seasons. For instance,
households’ expenditures on education decrease by 26 percent between August and
November, drop again by 77 percent between November and February, and increase by
22 percent between February and May. Health expenditures decrease by 61 percent
between August and November, increase by 27 percent between November and February,
and rise again by 54 percent between February and May. These results suggest that the
demand for these non-food commodities is highly seasonal.

Table 2-16 and 2-17, below, present mean weekly expenditure on non-food
commodities and budget shares by season and by income group and percentage changes
in both across season. The results indicate the income groups have different adjustment
patterns in the proportion of the non-food budget devoted to individual non-food
commodity groups in any given season. For instance, between August and November,
low-income households reduce the non-food budget share allocated to personal care (~56
percent), housewares (~54 percent), and recreation (~45 percent). In contrast, high-
income households incur the largest reductions in the non-food budget share devoted to
education (~39 percent) and hygiene (~23 percent). These results also suggest that low~
income households may use the timing of their purchases of non-food items as a

mechanism to minimize ﬂuctuations in their food consumption levels across seasons.

56

Table 2-16: Mean Weekly Non-Food Expenditure (FCFA/AB) by Season and

Income Group

 

 

 

 

 

 

 

I Phase I % Change Between
Income Groups | L H PH | P | H-L PH-H [ P-PH
Low 7 9
Education 5 94 6 28 1 879 ~93 3 5 1
Housewares 655 92 161 258 ~86 75 60
Personal Care 1584 216 358 21 1 ~86 66 ~41
Health 392 115 251 115 ~71 118 ~54
Hygiene 129 96 87 80 ~25 ~10 ~8
Energy and Utilities 476 356 462 293 ~25 30 ~37
Tobacco 30 23 3 1 1 ~23 ~89 303
Transportation 877 312 223 163 ~64 ~29 ~27
Recreation 182 31 17 23 ~83 ~45 34
Total 4330 1336 1567 1181 ~69 17 ~25
Middle

Education 68 161 3 1 138 ~98 ~49
Housewares 437 449 80 182 3 ~82 128
Personal Care 1710 314 620 95 ~82 97 ~85
Health 2364 725 269 229 ~69 ~63 ~15
Hygiene 127 200 133 143 57 ~33 7
Energy and Utilities 647 485 540 491 ~25 1 l ~9
Tobacco 142 193 121 132 36 ~37 9
Transportation 408 410 356 275 0 ~13 ~23
Recreation 50 54 38 20 7 ~29 ~47

Total 5954 2991 2161 1568 ~50 ~28 ~27

High

Education 672 288 120 127 ~57 ~58 6
Housewares 2442 1538 1 l 12 404 ~37 ~28 ~64
Personal Care 1256 793 1747 246 ~37 120 ~86
Health 326 362 1059 2109 1 1 193 99
Hygiene 382 206 246 208 ~46 19 ~15
Energy and Utilities 1372 803 930 832 ~42 16 ~11
Tobacco 61 139 95 133 128 ~32 40
Transportation 905 985 1023 808 9 4 ~21
Recreation 275 302 340 185 10 12 ~46

Total 7690 5416 6672 5054 ~30 23 ~24

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting.

57

Table 2-17: Mean Budget Share (%) Allocated to Individual Non-Food Commodity
Groups by Season and Income Group

 

 

 

 

 

 

 

 

 

Phase I % Change Between
Income Groups L H PH [ P | H-L PH-H | P-PH
Low 7 7
Education 0 7 0 2 N/A N/A N/A
Housewares 15 7 10 22 ~54 49 113
Personal Care 37 16 23 18 ~56 41 ~22
Health 9 9 16 10 ~5 86 ~39
Hygiene 3 7 6 7 142 ~23 21
Energy and Utilities 11 27 29 25 143 l 1 ~16
Tobacco 1 2 0 1 1 50 N/A N/A
Transportation 20 23 14 14 15 ~39 ~3
Recreation 4 2 1 2 ~45 ~53 78
Total 100 100 100 100
Middle
Education 1 5 0 O 373 N/A N/A
Housewares 7 15 4 12 105 ~75 214
Personal Care 29 1 1 29 6 ~63 173 ~79
Health 40 24 12 15 ~39 ~49 17
Hygiene 2 7 6 9 212 ~8 48
Energy and Utilities 11 16 25 31 49 54 25
Tobacco 2 6 6 8 170 ~13 51
Transportation 7 14 16 18 100 20 6
Recreation 1 2 2 1 l 13 ~2 ~26
Total 100 100 100 100
High

Education 9 5 2 3 ~39 ~66 41
Housewares 32 28 17 8 ~1 1 ~41 ~52
Personal Care 16 15 26 5 ~10 79 ~81
Health 4 7 16 42 57 138 163
Hygiene 5 4 4 4 ~23 ~3 12
Energy and Utilities 1 8 1 5 14 16 ~17 ~6 18
Tobacco 1 3 1 3 223 ~45 85
Transportation 12 18 15 16 55 ~16 4
Recreation 4 6 5 4 56 ~9 ~28

Total 100 100 100 100

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting.

58

2.3.3. Seasonal Nutrient Availability

The purpose of this section is to examine nutrient availability from at-home foods by
season and by income group. 2° The nutrients included in the analysis are calories,
protein, carbohydrate, vitamin A, vitamin C, iron, and calcium. The results will show
primary food sources for each type of nutrient and the prices per Kcal paid by households
in various socio-economic groups and by season. Tables A2~5 through A2~10 of
Appendix 2 provide detailed results on nutrient availability and main sources of nutrients
by season and by income group.

The main question addressed in this analysis is whether the consumption choices
Bamako households make in any given season translate into changes in the quantity, as
measured by calorie availability, and quality, as measured by protein and micronutrient
availability (vitamins and minerals), of food consumed in the households.21 The study
also investigates whether nutrient availability at the household level improves with rising
income levels. This is important for policy analysis, as it would mean that policies that
aim at increasing households’ real incomes might also improve their nutrition.
2.3.3.1. Nutrient Availability
Table 2-18, below, presents the average daily availability per adult equivalent of calories,
carbohydrates, protein, vitamin A, vitamin C, iron, and calcium and the nutrient adequacy
ratios by income group. The results show that the only nutrient that is consumed in
adequate amounts by all income groups of the sample population during the entire year is

carbohydrates. Average annual carbohydrate availability is on the order of 408 grams per

 

2° The results presented are estimates of nutrient availability from at-home foods and not actual nutrient
intake. The FAO estimates that about 10 percent of the edible portion of food is wasted by the household
before ingestion. This means that the present ﬁgures underestimate actual nutrient intakes.

59

adult equivalent per day, which is higher than the FAO recommended dietary allowance
(RDA) of 300 grams per adult equivalent per day.

Table 2-18: Daily Nutrient Availability per Adult Equivalent by Income Group and
Nutrient Adequacy Ratios (%)

Income Food Carbo— | Vitamins | Minerals

Group Energy hydrate Protein I Vit A I Vit C I Calcium I Iron

I Kcal Grams Micro grams Milligrams Milligﬂ
4 27

 

 

 

 

 

 

 

 

Low 2082 391 55 23 390 22
Middle 2051 382 51 315 32 354 20
High 2495 452 67 532 40 510 25
Mean 2209 408 57 360 33 418 23
Nutrient Adequacy Ratios (%)
Low 95 130 87 39 59 39 38
Middle 93 127 81 52 72 35 34
High 1 13 151 106 89 89 51 43
Mean 100 136 91 60 73 42 38

 

Note: The Nutrient Adequacy Ratio (N AR) measures the extent an adult equivalent is
satisfying the recommended daily allowance (RDA). It is computed as a ratio of nutrient
availability per adult equivalent to RDA. The FAO’s RDA for an adult equivalent are
2200 kilocalories for energy, 300 grams for carbohydrates, 63 grams for protein, 600
micrograms for vitamin A, 45 milligrams for vitamin C, 1000 milligrams for calcium,
and 59 milligrams for iron (FAO, 1998).

The results indicate that there are some signiﬁcant nutrient and micronutrient
(vitamin A, vitamin C, iron, and calcium) deﬁciencies persisting in Bamako. Average
annual calorie availability in Bamako households is on the order of 2,209 calories per day
per adult equivalent. Although this amount slightly exceeds the FAO’s minimum daily
energy requirement of 2,200 kcal per adult equivalent, it conceals the fact that only
households in the high-income group attain this availability level. The low and middle-

income groups’ calorie availability levels never exceed the recommended levels during

the entire year.

 

2‘ Micronutrients are vitamins are minerals needed in small amount by the body for optimal human growth,
development, and healthy maintenance of the body (F A0, 1999).

60

Vitamin A availability amounts to 360 micrograms per adult equivalent per day
compared to the recommended daily allowance (RDA) of 600 micrograms; thus Bamako
households can satisfy about 60 percent of the RDA. Concerning vitamin C, urban
households are only able to meet 73 percent of the daily recommended intake level; with
an average availability in the order of 33 micrograms per adult equivalent per day
compared to the RDA of 45 micrograms. Average iron and calcium availability is about
23 and 418 milligrams, respectively, (or 38 and 42%, respectively, of the RDA) per adult
equivalent per day.

The protein content of the average Bamako household diet is 57 grams per adult
equivalent per day, which is close to the recommended daily protein allowance of 63
grams per adult equivalent per day. However, as shown in Table 2-19 below, only 18
percent of the total protein available for consumption comes from animal sources. In
general, animal proteins tend to be of higher quality than vegetable and grain proteins
because they are easily digestible and more “complete” as they contain all essential
amino acids. Moreover, the results indicate that the proportion of protein obtained from
animal sources tends to increase with rising income levels: 16 percent, 17 percent, and 21
percent for the low, middle, and high-income groups, respectively.

Table 2-19: Protein Contributed by Major Food Groups (%) by Income Group
Protein Contributed by Major

 

 

 

 

Food Groups (%)
Commodities I Low I Middle High I Mean
Rice * 3o 33 26 ’ 30
Other Staples 33 28 30 30
Meat and Fish l6 17 21 18
Vegetables 8 6 8 7
Oil 0 0 O 0
Sugar 0 0 O 0
All others 14 15 15 14

 

61

2.3.3.2. Income and Nutrient Availability

The results, presented in Table 2-18, also indicate that higher income levels are
associated with greater availability of nutrients in Bamako households. For instance, the
average daily availability of calories, protein, vitamin C, and iron per adult equivalent
increase by 17 percent, 18 percent, 32 percent, and 13 percent, respectively, as household
income increases from the lowest to the highest income tercile. However, signiﬁcant
micronutrient deﬁciencies persist even at high-income levels. For instance, the high-
income group satisﬁes only about 51 percent of the recommended daily allowance for
calcium and 41 percent of the RDA for iron (Table 2-18). These ﬁndings, consistent with
those of Rogers and Lowderrnilk (1981), point to the fact that as households’ income
increase, the immediate concern is to increase the quantity of food consumed. This
underscores the fact that in Mali the consumption patterns of the poor and rich are very
similar. High-income households tend to consume more of the same type of foods that
poor households eat, even if some diversiﬁcation of the diet is evidenced at higher
income levels.

2.3.3.3. Seasonal Fluctuations in Nutrient Availability

Figure 2-5, below, shows the distribution of calorie availability across households and
across seasons. The results are quite alarming, as 48 percent of Bamako households are
unable to meet the 2200 minimum daily calorie requirement during the lean season.
During the planting season, the results indicate that about 68 percent of Bamako

households can’t achieve the minimum calorie availability levels.

62

Figure 2~5: Distribution of Calorie Availability across Households by Season

 

 

 

 

 

 

 

 

4500 -

4000 r

3500 r

3000 - August
'3 250° ‘ T111352?
a 2000 r - - .. May

1500 r standard

1000 r

500 -

0 l l 1 w 1 t r t t t i t

 

 

Percentage of Sample
3 1 5 28 40 53 65 78 90

Table 2-20, below, shows the mean daily nutrient availability, nutrient adequacy
ratios, and percentage change in nutrient availability by season. Bamako households’
mean calorie availability is highest (2,263 kilocalorie/day/AE) during the lean season
(August) and lowest (2087 kcal/AE/day) at the beginning of the rainy season (May),
when it falls well below the recommended intake levels of 2,200 kcal/day/AE. Calorie
availability, thus, remains fairly stable from the lean (August) to the post-harvest
(February) season: decreases by 1 percent between August and November and then
increases by 1 percent between November and February. However, the greatest
percentage change (~8 percent), in calorie availability levels is observed between the post
harvest (February) and the planting season (May).

The results, presented in Table 2-20, also indicate that the availability of nutrients,
as manifested in the nutrient adequacy ratios, is greatest in Bamako households during
the lean season (August): 103 percent for calories, 136 percent for carbohydrates, 96

percent for proteins, 71 percent for Vitamin A, 84 percent for vitamin C, 40 percent for

63

iron, and 49 percent for calcium. This ﬁnding can be explained by the fact that Bamako
households receive substantial ﬁnancial help from migrants, relatives, and the extended
family during the month of August.

Table 2-20: Nutrient Availability, Nutrient Adequacy Ratios, and Percentage
Change in Nutrients Availability across Seasons

 

 

 

 

 

 

 

 

 

 

Food Carbo- I Vitamins I Minerals
Phase Energy hydrate Protein I Vit A I Vit C I Calcium I Iron
r Kcal Grams Micro gram mg mg
L 2263 409 61 428 38 490 23
H 2236 413 59 338 32 396 23
PH 2251 414 S8 392 36 431 22
P 2087 398 52 284 26 355 22
Nutrient Adequacy Ratios (%)
L 103 136 96 71 84 49 40
H 102 138 93 56 70 40 39
PH 102 138 92 65 80 43 37
P 95 133 83 47 58 36 37
Percentage Change (%)
H-L ~1 1 ~3 ~21 ~17 ~19 ~1
PH-H 1 0 ~l 16 13 9 ~5
P-PH ~7 ~4 ~10 ~28 ~27 ~18 ~l

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting.

The results, presented in Table 2-20, indicate that average daily nutrient
availability is lowest during the planting season (May): 95 percent for calories, 133
percent for carbohydrates, 83 percent for proteins, 47 percent for vitamin A, 58 percent
for vitamin C, 37 percent for iron, and 36 percent for calcium. The greatest variation in
nutrients availability is observed between the post-harvest and planting seasons when, a
general decline is registered: ~7 percent for calories, ~4 percent for carbohydrates, ~10

percent for protein, ~28 percent for vitamin A, ~27 percent for vitamin C and -21 percent

for calcium. The smallest variations are registered between the harvest and post-harvest
seasons.

The results in Table 2~20 also reveal that seasonal variations in the availability of
micronutrients (vitamin A, vitamin C, and calcium) are much more pronounced than that
of calories. For instance, calorie availability decreases by 1 percent between the lean and
post-harvest season, increases by 1 percent between the harvest and post-harvest season,
and drops by 7 percent between the post-harvest and planting season. However, vitamin
A availability decreased by 21 percent, increased by 16 percent, and decreased by 28
percent over the same period.

A possible explanation for this pattern is that households are aware of shortfalls in
their calorie intake (they feel hungry). Thus, they attempt to maintain the amounts of
calories available for consumption somewhat constant during the year by reducing the
consumption of foods that contain essential micronutrients but few calories (e. g., meat,
ﬁsh, and vegetables). The results in Tables 2~21 and 2~22, below, support this
explanation. For instance, the 1 percent drop in calorie availability between the lean and
harvest seasons is achieved through a 25 percent decrease in the contribution vegetables
and oil, as shown in Table 2~21, and through substitutions of beef for dry ﬁsh within the

meat and ﬁsh commodity group category (Table 2~22).

65

 

 

r-v

Table 2~21: Calories Contributed by Major Food Groups (%) by Season and
Percenta e Chan e across Seasons

 

 

 

 

 

Share (%) % Change Between

Food Groups L H PH | P JH-L | PH-H | P-PHJ
Rice 39 41 42 42 5 2 0
Other Staples 29 29 27 31 0 ~7 15
Meat and Fish 5 5 5 4 0 0 ~20
Vegetables 4 3 4 3 ~25 33 ~25
Oil 8 6 7 5 ~25 17 ~29
Sugar 7 7 7 8 0 0 14
All others 7 8 8 7 14 0 ~13

 

 

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest

and P = May = planting.

Table 2~22: Contribution of Meat and Fish to Calorie Availability (kcal/AE/day)
and Budget Shares (%) by Season

 

 

 

 

 

 

 

 

Items Mean Dail Caloric Availability (kcal/AB) Nutrient Source (%)
L H I PH_] P L _ H | PH | P _
Beef 68 75 67 58 61 71 65 70
Mutton 3 0 8 1 2 0 8 1
Poultry 2 0 1 0 2 0 1 1
Dry Fish 29 22 25 19 26 21 24 23
Fresh Fish 10 8 2 5 9 8 2 6
Total 1 1 1 105 103 82 100 100 100 100

 

 

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest

andP=May=

planting.

2.3.3.4. Sources of Nutrients

Table 2~23, below, shows the sources of nutrients by income group and for the entire

sample. Staples are by far the leading source of calories, providing on average annually

70 percent of the total calories available for consumption. They are followed by all other

foods (dairy products, fruits, seasonings and spices, and beverages) at 8 percent; oil and

sugar, at 7 percent each; and meat and ﬁsh and vegetables, at 4 percent each.

66

Table 2~23: Sources of Nutrients (%) by Income Group
Nutrients Contributed by Major Food Groups (%)
At-Home Foods Calories I Carbs I Proteinj Vit A I Vit C I Calcium T Iron

 

 

 

 

 

 

 

 

Rice

Low 42 50 30 0 0 16 19

Middle 44 53 33 O 0 19 23

High 37 46 26 0 0 14 18

Mean 41 50 3O 0 0 16 20

Other Staples

Low 31 35 33 6 12 14 49

Middle 26 30 28 4 1 l 12 43

High 29 34 30 6 16 13 45

Mean 29 33 30 5 13 13 46

Meat and Fish

Low 4 0 16 4 0 22 5

Middle 4 O 17 3 0 21 6

High 5 0 21 3 0 22 8

Mean 4 0 18 3 0 22 7

Vegetables

Low 4 4 8 59 85 35 17

Middle 3 4 6 58 85 36 18
High 4 5 8 65 79 35 20
Mean 4 4 7 61 83 35 19

Oil

Low 6 0 0 26 0 0 0

Middle 7 0 0 30 0 0 0

High 8 0 0 21 0 0 0

Mean 7 0 0 26 0 O 0

Sugar

Low 6 8 0 0 0 0 0

Middle 8 l 1 0 0 0 0 0

High 8 l2 0 0 0 O 0

Mean 7 11 0 0 0 0 0

All others

Low 7 2 14 6 3 l3 9
Middle 8 2 15 5 4 13 10

High 8 2 15 5 4 16 9

Mean 8 2 14 5 4 14 9

 

The results, in Table 2~23, indicate that the contribution of rice to calorie
availability decreases from 42 to 37 percent while that of other staples increases from 31
percent to 29 percent, as households’ income increases. The results in Table A2~l of
Appendix 2 show that households reduce their consumption of millet-sorghum and

increase that of wheat, maize, and sweet potato, as their income increases.

67

Moreover, the results in Table 2~23, reveal that the share of calories derived from
meat and ﬁsh (4 percent for the low and 5 percent for the high-income group), sugar (6
percent and 8 percent), and other foods (7 percent and 8 percent) tend to be higher for
high-income households. Thus, as Bamako households’ income increase, the proportion
of the calories they obtain from more expensive sources increases while that of cheaper
sources such as staples decreases (Bennett’s Law). These ﬁndings suggest that
households tend to diversify their diets as they attain higher income levels.

The three most important sources of carbohydrates are rice (50 percent), other
staples (33 percent), and sugar (11 percent). Vegetables provide on average 35 percent of
the total calcium available in urban households. They are followed by meat and ﬁsh (22
percent), rice (16 percent), all others (14 percent) and other staples (13 percent). The two
most important sources of Vitamin C for Bamako households are vegetables (83 percent)
and other staples (13 percent), combining to provide on average 96 percent of the total
vitamin C available in households. The leading source of Vitamin A for all income
groups across all seasons are vegetables (61 percent), followed by oil (26 percent), and
other staples (5 percent) and other foods (i.e., fruits, beverages, and nuts) (5 percent). The
main sources of protein in urban households are staples (60 percent), followed by meat
and ﬁsh (1 8 percent), all others foods (dairy products, fruits, seasonings and spices, and
beverages) (14%), and vegetables (7%).

Table 2~24, below, reports the mean daily protein availability from meat and ﬁsh
sources and the contribution of speciﬁc types of meat and ﬁsh to animal protein
availability by season. The results show that on average 49 percent of the total animal

protein available for consumption comes from beef, followed by dry ﬁsh, at 39 percent,

68

and fresh ﬁsh, at 9 percent. The contribution of mutton to protein availability is greatest,
5 percent, during the post-harvest, which corresponded to the period when the Tabaski, a
religious festivity, occurred during the survey year. Hence, the period of heaviest mutton
consumption is likely to shiﬁ from year to year as the date of the Tabaski shifts in
accordance with the lunar calendar.

Table 2~24: Mean Daily Animal Protein Availability in Grams/AE/day and

Contribution of Speciﬁc Types of Meat and Fish to Animal Protein Availability (%)
by Season

 

 

 

 

 

Items Mean Daily Protein Availability (g/AE) Nutrient Source (%)
L HIPHIPIAvg L|H|PH|PAvg_
Beef 5 6 7 5 4 5 43 52 7 49 52 49
Mutton o 0 1 o o 1 o 5 1 2
Poultry 0 o o o o 2 o 1 1 1
Dry Fish 5 4 4 3 4 41 36 41 38 39
Fresh Fish 2 1 o 1 1 12 11 4 9 9

 

 

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting.

2.3.3.5. The Cost of Calories

The results, presented in Figure 2-8 below, indicate that coarse grains (millet-sorghum
and maize), rice, and other staples are by far the cheapest sources of calories: 35, 65, and
74 CF A Franc per 1000 calories, respectively.22 In contrast, vegetables, other foods (i.e.,
fruits, nuts), and meat and ﬁsh are the most expensive sources of calories: 759, 758 and
517 CFA Franc per 1000 calories, respectively. Rice dominates Bamako households’
diets despite the fact that it constitutes a more expensive source of calories than other
staples. Bamako households’ preference for rice is largely attributed to taste factors and

to the fact that rice takes less time, fuel, and labor to prepare. Hence, the price per calorie

 

22 The price of calories is computed as a ratio of households’ total expenditures and total calorie
availability.

69

for rice may actually be lower than that of the other staples when the preparation costs are
taken into account.

Figure 2~6: Average Cost of Calories (CFA Francs/1000 calories)

 

 

 

fcfa/kcal
900 n
800 j 758
700 A
600 -
500 -
400 -
300 1

J

iii 7" 35 m

o I , I ,

Rice Coarse Other Meat and Vegetables Oil Sugar Others

Grains Staples Fish

The price paid per 1000 calories by different income groups can provide an
indication of whether quality upgrading occurs as households’ income increases.
Generally, poor households tend to consume food items of lower quality and as a result
the price per kilocalorie paid by these households is lower than that paid by high-income
households. Reardon et a1. (1999) have argued that rich households consume the better
quality locally produced rice while poor households eat more of the cheap imported
Asian rice.23 Table 2~5, below, presents the average cost of calories for at-home foods by
season and by income group. The results provide no supportive evidence of quality
upgrading with respect to rice as households’ income increase. In fact, the price paid per
1000 calories of rice decreases slightly with rising income levels: 74 F CFA per 1000

calories for the low versus 73 FCFA per 1000 calories for the high income group.

 

23 High-quality rice has a low percentage of broken grains (less than 10%) whereas low quality rice has
more than 10% of broken kernels.

70

Table 2~25: Average Cost of Calories (FCFA/1000 kcal) for At-Home Foods by

Season and by Income Grog
Price per 1000 calories

 

 

 

Food Group I 1_. 1 H 1 PH P 1 Avg.
Rice

Low 75 76 73 74 74
Middle 76 76 71 72 74
High 72 76 72 73 73
M646 ............................ 747672 ............ 7. 3 ............ 74 ......

Coarse Grains
Low 39 39 37 48 41
Middle 35 35 31 40 35
High 28 29 27 38 31
M949 ............................ 343432 ............ 4 3 ............ 33 ......

Other Staples
Low 49 52 34 38 43
Middle 62 87 66 60 69
High 80 75 83 93 83
M93}! ............................ 647261 ............ 64 ............ 6 3 ......

Meat and Fish
Low 401 501 535 492 482
Middle 623 436 547 529 534
High 519 554 516 553 536
MeanSM ........... 4 3.7 ........... 33.3. .......... 3.2.5. .......... 3.1.7 .....

Vegetables
Low 878 641 634 787 735
Middle 1062 708 704 653 782
High 951 833 594 666 761
M46963 ........... 7. 3.7 ........... 64.4. .......... 7 9.2. .......... .733 .....
Oil

Low 75 102 98 74 87
Middle 94 79 76 75 81
High 81 87 83 78 82
Mean 83 90 86 76 84
"""""" s' ugar
Low 128 120 131 132 128
Middle 147 1 12 104 126 122
High 109 128 104 113 114
Mean 128 120 113 124 121
"""""" Others
Low 780 607 600 727 679
Middle 972 681 652 597 725
High 1059 877 872 672 870
Man 937 722 708 665 758

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting.

Similarly, the results reveal that the price paid for 1000 calories of coarse grains

(millet-sorghum and maize) tends to decrease with households’ income: 41, 35, and 31

71

FCFA per 1000 calories for the low, middle, and high-income groups, respectively. Thus,
the evidence here suggests that poor households pay slightly more for 1000 calories of
cereals than rich households, and this is consistent across all seasons. The results,
however, show that the price paid per 1000 calories for other staples (wheat and roots and
tubers) tends to increase with income (43, 69, and 83 FCFA per 1000 calories for the low,
middle and high income group, respectively). Therefore, suggesting that higher income
households perhaps consume better quality of other staples.

However, these ﬁndings may well be due to the fact that rich households often
purchase rice in bulk and consequently, pay lower per unit costs than poor households.
The income groups were divided into groups that reported having purchased rice in bulk
versus those that didn’t, in order to assess to what extent the differences in the prices paid
per 1000 calories reﬂect quality upgrading. Table 2~26, below, shows the average price
paid per 1000 calories for rice by type of purchase, season, and income group.

Table 2~26: Average Price Paid Per 1000 Calories for Staples by Type of Purchase

 

 

 

 

 

 

Phase
Bulk Purchase L l H | PH I P | Avg
Low | High] Low HighI Low] High | Low | Highl Low | High-
No 74 72 77 75 73 7o 74 72 75 72
Yes 82 73 73 77 73 73 74 74 75 74

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting. NA: Not Available.

The results, in Table 2~26, indicate that on average, even among households that
reported having purchased rice in bulk, low-income households still tend to pay more (75
FCFA) per 1000 calories than high-income households (74 FCFA). The harvest season is
the only time during the entire year when high-income households that purchased rice in

bulk paid more per 1000 calories than low-income households (73 for the low versus 77

72

for the high tercile). Thus, only during that period do the results provide supportive
evidence of rich households buying rice of higher quality.

It is also possible that the observed grain price variation reﬂects locational differences
(i.e., the area of residence), as households residing in the city periphery may pay more
per unit costs than those in city center. Figure 2~7, below, shows the average price paid
for rice by area of residence. The results show that households who live close, or who
have transportation means, to the Niamakoro market, which is the largest wholesale
market in Bamako and is located in the center of the city, pay the lowest price for rice,
254 CFA Francs per kilogram. In contrast, households living in Boulkassoumbougo,
which is located in the city periphery, and who are unable to incur search costs (e.g.,
information costs and transportation costs), pay the highest unit price, 274 CFA Francs,
for rice.

Figure 2~7: Average Price Paid for Rice (CFA Francs/kg) by Area of Residence

FCFA/KG
280

 

 

 

270 i

260 4

 

250 4

 

 

 

2.3.4. Sensitivity Analysis
The estimates of nutrient availability presented in the previous section were solely based

on the at-home food consumption data. In this section, rough estimates of nutrient

73

availability from away-from-home foods were computed to provide an idea of how the
results would change when these items are taken into account. Following Subramanian
and Deaton (1996), the away-from-home nutrient estimates were derived based on the
assumption that the average cost per unit of nutrient of away-from-home foods is 50
percent more than at-home foods. The 50 percent premium is assumed to reﬂect
processing margins (Subramanian and Deaton (1996)). Then, the average away-from-
home nutrient availability is computed using the following formula:

NFAFH=(P*W) * NFAH, (1)
Where N is nutrient availability, FAF H is food away ﬁ'om home (FAF H), FAH is food at
home (FAH), P is the premium (e. g., 50 percent), and W is the average budget share
allocated to food away from home. P*W represents the percentage change in nutrient
availability.

The problems with this approach are the implicit assumptions that (1) foods
consumed at-home are of the same quality as those consumed away-from-home and that
(2) nutrient availability from away-from home foods is distributed equally within the
household. However, foods consumed away from home generally include meals that are
very labor intensive and time consuming to prepare (i.e., fonio and atieke) and foods that
provide some diversity to households’ diets (i.e., dairy products, fruits, and nuts).
Mangoes, for instance, are an important source of vitamin A. It would be too costly for
the household to provide these foods to every single one of its members. Table A2-11,
presented in appendix 2, shows that households allocate on average 0.7 percent of their
at-home food budget to fruits whereas fruits occupy 5.4 percent of the away-from-home

food budget. Similarly, dairy products take up on average 2.2 percent and 5.9 percent of

74

food at home and food away ﬁom home budgets, respectively. It is hard to assess
whether households allocate a greater proportion of their away-from-home food budget to
meat and ﬁsh commodities, since meals taken away from home are oﬁen in form of
dishes. The food away from home data indicates that a substantial proportion, on average
about 86 percent, of food away from home expenditures, are incurred by household
members who are employed, suggesting that young children are largely excluded from
this consumption. Also, the data shows that on average about 20 percent of food away
from home expenditures are made by the head of household, while, on average, the
household head accounts for only 6 percent of the household population (The average
household in Bamako is composed of 17 members). Hence, the beneﬁts of away-from-
home consumption, in terms of nutrient content, would be skewed in the households,
beneﬁting mainly the head of household and household members who have a source of
income.

Table 2~27, below, shows the baseline values (e.g., nutrient availability from at-
home foods and nutrient adequacy ratios) and the effect of including estimates of nutrient
availability from away-from-home foods on average daily nutrient availability per adult
equivalent and the nutrient adequacy ratios by income group. The results, in Table 2~27,
show that if away-from-home foods were taken into account, average nutrient availability
in Bamako households would increase by 9.5 percent. The results also indicate that the
availability of nutrients, as manifested in the nutrient adequacy ratios, would increase to:
110 percent for calories, 149 percent for carbohydrates, 100 percent for proteins, 66

percent for Vitamin A, 80 percent for vitamin C, 46 percent for iron, and 42 percent for

75

calcium. One should note that these results are upper-end estimates, as they assume zero

wastage of both at-home and away-from-home nutrients.

Table 2~27: Effects of Including Estimates of Nutrient Availability from Away-
From-Home Foods on Total Nutrient Availability by Income Group

 

 

 

 

 

 

 

 

Income Food Carbo- I Vitamins I Minerals
Group Energy hydrate Protein I Vit A I Vit C I Calcium Iron
I Kcal I m m mg
9 Baseline Values:
Nutrients from at-home foods
Low 2082 391 55 234 27 390 22
Middle 2051 382 51 315 32 354 20
High 2495 452 67 532 40 510 25
Mean 2209 408 57 360 33 418 23
Nutrient Adequacy Ratios (%)
Low 95 130 87 39 59 39 38
Middle 93 127 81 52 72 35 34
High 113 151 106 89 89 51 43
Mean 100 136 91 60 73 42 38
Simulation:
% Change in Nutrient Availability
Low 9.0 9.0 9.0 9.0 9.0 9.0 9.0
Middle 11.1 11.1 11.1 11.1 11.1 11.1 11.1
High 8.4 8.4 8.4 8.4 8.4 8.4 8.4
Mean 9.5 9.5 9.5 9.5 9.5 9.5 9.5
Total Amounts of Nutrients Available (FAH+FAF H)
Low 2269 427 60 255 29 425 24
Middle 2279 425 57 350 36 394 22
High 2705 490 72 577 43 553 28
Mean 2419 447 63 395 36 458 25
Nutrient Adequacy Ratios (%)
Low 103 142 95 43 64 43 41
Middle 104 142 90 58 80 39 38
High 123 163 1 15 96 96 55 47
Mean 1 10 149 1 00 66 80 46 42

 

Once the effects are disaggregated by income group, the results indicate that the

average daily availability of nutrients per adult equivalent increase by 9 percent, 11.1

percent, and 8.4 percent respectively, as household income increases from the lowest to

76

the highest income tercile. Hence, middle-income households would experience the
greatest increase in nutrient availability since, as shown in section 2.3.2.2., they allocate
the greatest percentage (22 percent versus 18 and 17 percent, respectively, for the low
and high-income groups) of their food budget to food away from home.

The results also show that all income groups would now be able to meet minimum
daily calorie requirements but only the high-income group would be able to satisfy the
recommended dietary allowance (RDA) for protein. Moreover, the increase in the
amounts of vitamin A, vitamin C, calcium, and iron would not be enough for households
in all income groups to meet the RDA for these nutrients.

Table 2~28, below, shows the baseline values (e.g., nutrient availability from at-
home foods and nutrient adequacy ratios) and the effect of including estimates of nutrient
availability from away-from-home foods on average daily nutrient availability per adult
equivalent and the nutrient adequacy ratios by season. The results, in Table 2~28, show
that if away-from-home foods are taken into account, average nutrient availability in
Bamako households would increase by 9.4, 9.2, 8.5, and 11.1 percent during the lean,
harvest, post-harvest, and planting seasons, respectively. The results also indicate that
Bamako households would now be able to meet minimum daily calorie requirements
during all seasons; however, they would still not be able to satisfy the recommended
dietary allowance (RDA) for vitamin A, vitamin C, calcium, and iron in all seasons

considered.

77

Table 2~28: Effects of Including Estimates of Nutrient Availability from Away-
From-Home Foods on Total Nutrient Availability by Season

 

 

 

 

 

 

 

 

 

Income Food Carbo~ I Vitamins I Minerals
Group Energy hydrate Protein I Vit A I Vit C I Calcium Iron
I Kcal I g I g I Eg I mg I mg mg
7 Baseline Values:
Nutrients from at-home foods
L 2263 409 61 428 38 490 23
H 2236 413 59 338 32 396 23
PH 2251 414 58 392 36 431 22
P 2087 398 52 284 26 355 22
Avg 2209 408 57 360 33 41 8 23
Nutrient Adequacy Ratios (%)
L 103 136 96 71 84 49 40
H 102 138 93 56 70 40 39
PH 102 138 92 65 80 43 37
P 95 133 83 47 58 36 37
Avg 100 136 91 60 73 42 38
Simulation:
% Change in Nutrient Availability
L 9.3 9.3 9.3 9.3 9.3 9.3 9.3
H 9.2 9.2 9.2 9.2 9.2 9.2 9.2
PH 8.5 8.5 8.5 8.5 8.5 8.5 8.5
P 11.1 11.1 11.1 11.1 11.1 11.1 11.1
Avg 9.5 9.5 9.5 9.5 9.5 9.5 9.5
Total Amounts of Nutrients Available (FAH+FAF H)
L 2474 447 66 468 42 536 26
H 2442 451 64 369 35 433 25
PH 2442 449 63 425 39 468 24
P 2318 442 58 316 29 395 24
Avg 2420 447 63 395 36 45 8 25
Nutrient Adequacy Ratios (%)
L 112 149 105 78 92 54 43
H 111 150 102 61 77 43 43
PH 111 150 100 71 87 47 40
P 105 147 92 53 64 39 41
ﬁg 1 10 149 100 66 80 46 42

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest and P = May =
planting.

78

2.4. Conclusions

This essay has examined Bamako households’ seasonal consumption patterns through a
descriptive analysis of seasonal changes in expenditure patterns and seasonal availability
of nutrients. The results show that seasonal changes in the price of cereals induce
households to incur substantial adjustments in their budget allocation pattern among and
within major food and non-food components in any given season. The ﬁndings suggest
that households are willing to allocate the marginal increase in their income to
diversifying their diets and to acquiring more non food commodities only during periods
of greater food availability in urban markets, thus when lower food prices prevailed, such
as during the post harvest season.

The results also show households’ incur signiﬁcant substitutions among and
within food commodity groups in order to attempt to smooth their calorie availability
across seasons. Such adjustments often result in large variations in the quality, as
measured by protein, carbohydrate, and micro-nutrients’ availability, of food available in
the household. Evidence of signiﬁcant nutrient and micronutrient deﬁciencies persisting

in the households surveyed was indeed brought to light.

79

REFERENCES

Chambers, R. (1981). Introduction. In Seasonal Dimensions to Rural Poverty, pp. 1-8,
ed. R. Chambers, R. Longhurst, and A. Pacey. London: Frances Pinter.

Deaton, A and Muellbauer J. Economics and consumer behavior. Cambridge,
Cambridge University Press: 1980a.

Deaton, A. “Price elasticities from survey data.” Journal of Econometrics 44 (1990)
281—309.

Deaton, A. The Analysis of Household Surveys: A Microeconomic Approach to
Development Policy. World Bank. The John Hopkins University Press. Baltimore,

London: 1997.

Deaton, A. Quality, Quantity, and Spatial Variation of Price. AER. 78 (1988b) 3: 418~
430.

Deaton, A. Understanding Consumption. Oxford University Press, 1992.

Deaton, A. and A. Tarozzi. 2000. Prices and Poverty in India. Research Program in
Development Studies. Princeton University.

Deaton, A. & Zaidi, S., 1999. Guidelines for Constructing Consumption Aggregates for
Welfare Analysis. Papers 192, Princeton, Woodrow Wilson School ~ Development
Studies.

Dembelé, N. N. and J .M. Staatz, June 1999. The Impact of Market Reform on
Agricultural Transformation in Mali. MSU Agricultural Economics Staff Paper No. 99-

29.

Direction Nationale De La Statistique Et De L'informatique (DNSI), Avril ~ Juillet 1996.
Enquéte Sur Les Dépenses Des Ménages De Bamako.

F AOSTAT. 2003. Available on the intemet at: http://apps.fao.org/page/collections

FAO.2002. Food Energy: Methods of Analysis and Conversion Factors ~ Report of a
technical workshop. FAO FOOD AND NUTRITION PAPER 77

FAQ. 1999. Apercus Nutritionnels par Pays ~ MALI 26 janvier 1999.

F A0. 1998. Human Vitamin and Mineral Requirements ~ Report of a Joint FAO/WHO
Expert Consultation held in Bangkok, Thailand, September 1998.

80

F A0. 1997. Carbohydrates in Human Nutrition ~ Report of the Joint F AO/WHO Expert
Consultation 14-18 April 1997

FAO. 1997. Calcium throughout life Food, Nutrition and Agriculture, No. 20, 1997.

FAO. 1993. Fats and oils in human nutrition ~ Report of a joint FAO/WHO expert
consultation 19-26 October 1993.

FEWS 2000. Mali Evaluation de La Vulnérabilité Courante: F évrier 2000. Agence
Amérieaine Pour Le Développement International, Associates In Rural Development,
Inc., and Famine Early Warning System.

F EWS 2000. Document available online at:
http://www.afrol.com/News2000/ma1001_grain_prod.htm

FEWS 2000. Document available online at: http://www.reliefweb.int/w/rwbnsf

Gabas, J ~J . and Girl, J. (1987).The Food Situation in the Sahel: an Assessment. In
Cereals Policies in Sahel Countries: Acts of the Mindelo Conference, Eds CILSS/Club
du Sahel/OECD, Paris.

OANDA’s exchange rates. Retrieved November 21, 2003, from http://www.oanda.com/

Newbery, David M, 1989. The Theory of Food Price Stabilization. Economic Journal,
Vol. 99 (398) pp. 1065-82.

Paxon, Christina H., 1993. Consumption and Income Seasonality in Thailand. Journal of
Political Economy, 1 01 ( 1 ): 39-72.

Pitt, M.M. 1983. Food preferences and nutrition in rural Bangladesh.The Review of
Economics and Statistics 65: 105-114.

Raunikar, Robert and Huan, Chung-Liang. Demand Analysis: Problems, Issues, and
Empirical Evidence. Ames: Iowa State University Press, 1987.

Reardon et a1. Household Consumption Responses to the Franc CFA Devaluation:
Evidence from Urban Mali. Food Policy 24 (1999), 517-534.

Rogers, BL. and Lowdermilk, M. (1991). Price Policy and Food Consumption in Urban
Mali. Food Policy 16, 461-473.

Sadoulet, Elisabeth, and de Janvry, Alain. Quantitative Development Policy Analysis.
Chapter 6. Baltimore, MD: The Johns Hopkins University Press, 1995.

Sahn, E. 1989. Seasonal Variability in Third World Agriculture: The Consequences for
Food Security. Baltimore and London: John Hopkins University Press.

81

Sen, Amartya. Poverty and F amines: An Essay on Entitlement and Deprivation. Oxford:
Oxford U. P., 1981.

Singare, K., T. Reardon, Y. Camara, M. Wanzala, B. Teme, O. Sanogo. 1999. Household
consumption responses to the franc CFA devaluation: evidence from urban Mali. Food
Policy, 24(5), October: 517-534.

Sundberg, Shelly and Alayne Adams. Food Composition Table for Mali. September
1998. Michigan State University/U SAID.

Teklu, Tesfaye, 1996. Food Demand Studies in Sub-Saharan Africa: A Survey of
Empirical Evidence. Food Policy, Vol. 21, No. 6, pp. 479-496, 1996.

Timmer, C. P. and H. Alderman. 1979. Estimating consumption parameters for food
policy analysis. American Journal of Agricultural Economics 61: 982-987.

Timmer et a1. 1983. Food Policy Analysis. Published for the World Bank. John Hopkins
University Press.

United States Department of Agriculture Agricultural Research Service (2003). Nutritive
Value of Foods. Home and Garden Bulletin Number 72. Susan E. Gebhardt and Robin G.

Thomas.

USDA. (2003). Composition of Foods: Raw, Processed, Prepared. US. Department of
Agriculture, Agricultural Research Service, Nutrient Data Laboratory, Beltsville,
Maryland.

US. Department of Agriculture. Nutritive Value of Foods. Home and Garden Bulletin No.
72. Washington: US. Department of Agriculture, 1971.

Wu Lueng, Woot-Tsuen, Felix Busson, and Claude Jardin. Food Composition Table for

Use in Aﬁ'ica. Bethesda, Maryland: US. Department of Health, Education, and Welfare
and FAQ, 1968.

82

APPENDIX 2

83

Appendix A2~l Summary of Deaton and Zaidi (1999) Methods for Constructing
Expenditure Aggregates

Net Food Expenditure Aggregate

The households’ total food expenditure aggregates were computed by adding
expenditure on food at-home to expenditures incurred on meals away from home. The
net food expenditure was calculated as the total value of food purchases minus the
value of foods that the household donated (foods that were not for the household’s
own consumption). Giﬁs and remittances to other households are excluded, as their
inclusion would involve double counting if the transfers show up in the consumption
of other households.

Net Non-Food Expenditure Aggegate

The non-food expenditure aggregate was constructed by excluding the following
items: work related expenses, taxes paid, purchase of assets (i.e. car, motorcycle), and
lumpy expenditures such as marriages and births (Deaton, 2000). Taxes are excluded
because they are not part of consumption but a deduction from income. Gifts and
remittances to other households are excluded as their inclusion in the consumption
aggregate would involve double counting if the transfers show up in the consumption
of other households. Lumpy expenditures were also excluded because while almost
all households incur these types of expenditures at some stage, only a few of them are
likely to make such expenditures during the week of the survey. Unlike food items, for
which consumption was recorded daily, data on purchases of non-food items are often

collected from different recall periods (i.e., past 15 days or past month).

84

Figure A2~l: Agricultural Calendar in Mali

M3 M4

Plusez

Su'veyRomd Phase]

E

 

[Ibrvest

IHWSJ

 

 

Mllet
Nhize
808mm
Rice rainfed

Rice iniguatcd

Jul

AIgSepQINovDecJanFeblvhrAp'h/lame

85

Table A2-1: Mean Weekly At-Home Food Consumption (kg/AE) by Season and by
Income Group

Items

   
 
  
  
 
 
 
 
  
  
  
 
    
 
    
 
 
 
  
 

 

 

 

 

Rice 1.819
Millet-Sorghum 1.057

aize 0.216
Wheat 0.152
Fonio 0.025
Atieke 0.015
Cassava 0.008
Potato .
Sweet Potato ,‘ “ . - . . . .
""" Méii'ih'd'FEh""" """ - ' ' ' """ ""' ------ """'
Beef ." f ' - 0.238
Mutton 0.004
Poultry 0.006
Dry Fish .
Fresh Fish

Okra

Onion 0.260
Tomato 0.261
Beans 0.067
.chetksetablos ........................ -- .
Peanut Oil

Palm Oil

  
    
 
 
 
  
  
 
 
 
 
 
 
 
     
   

Sheanut Oil

Fresh Milk
Condensed Sweet Milk
Powdered Milk

Eggs

Peanuts

Quinqueliba
Other Beverage
Banana
Citronella
Dates

Lemon

Raisins
Tamarind
Orange
Seansonrn - s and S ices

Note: L — August= lean season, H= November— — harvest, PH= February= post-harvest and P= May— =

planting.

86

Table A2~2: Weekly Mean At-Home Food Items Consumption (MAEI by Phase

 

 

 

 

 

 

 

 

 

Phase
Items L I H 1 PH 1 P J Avg
Staples
Rice 1.703 1.769 1.812 1.681 1.741
Millet-Sorghum 1.104 1.142 1.081 1.162 1.122
Maize 0.145 0.104 0.107 0.109 0.116
Wheat 0.103 0. 065 0.080 0.068 0.079
Other Cereal 0.007 0.000 0.003 0.008 0.005
Atieke 0. 004 0. 001 0.004 0.005 0.003
Cassava 0. 008 0.035 0.010 0.002 0.014
Potato 0. 019 0. 004 0.100 0.024 0.037
Sweet Potato 0.014 0.091 0.003 0.005 0.028
Meat and Fish
Beef 0.224 0.246 0.222 0.190 0.221
Mutton 0.010 0.000 0.032 0.003 0.01 1
Poultry 0. 013 0.002 0.008 0.003 0.006
Dry Fish 0. 075 0.059 0.065 0.049 0.062
Fresh Fish 0. 068 0.052 0.017 0.029 0.042
Vegetables
Leave 0.159 0. 123 0.088 0.085 0.114
Okra 0.248 0. 054 0.029 0.087 0.105
Onion 0. 114 0.109 0.227 0.195 0.161
Tomato 0.100 0.193 0.213 0. 219 0.181
Other Vegetable: Fresh 0.145 0.214 0.310 0.110 0.195
All Other Vet able 0. 090 0.051 0.054 0.044 0.060
01
Peanut Oil 0.113 0.074 0.097 0.065 0.087
Palm Oil 0.019 0. 009 0.006 0. 011 0.011
Sheanut Oil 0.021 0. 030 0.029 0. 017 0.024
..-..--...-..SC r 0.315 0. 304 0.317 0.306 0.310
0t ers
Butter 0. 001 0.001 0.000 0.001 0.001
Buttermilk 0. 028 0.014 0.022 0.018 0.020
Fresh Milk 0. 002 0.002 0.001 0.000 0.001
Condensed Sweetened Milk 0. 006 0. 000 0.000 0.000 0.002
Powdered Milk 0. 019 0. 015 0.033 0.014 0.020
Eggs 0. 019 0. 002 0.002 0.000 0.006
Peanuts 0.145 0.188 0.159 0.144 0.159
Seeds 0.017 0.018 0.018 0.016 0.017
Other Nut&Seed 0. 000 0.000 0.000 0.000 0.000
Coffee 0. 002 0.001 0.001 0.001 0.001
Tea Lipton 0. 002 0.000 0.006 0.002 0. 003
Green Tea 0.003 0.002 0.001 0.001 0. 002
Quinqueliba 0.002 0.002 0.004 0.001 0.002
Other Beverage 0.001 0.000 0.002 0.000 0.001
Banana 0.006 0.013 0.013 0.001 0.008
Citronella 0.000 0.000 0.000 0.000 0.000
Dates 0.000 0.001 0.000 0.000 0.000
Lemon 0.016 0.063 0.009 0.002 0.023
Raisins 0.000 0.000 0.000 0.000 0.000
Tamarind 0.023 0.001 0.028 0.030 0. 021
Orange 0.009 0.000 0.000 0.000 0. 002
Seansonings and Spices 0.119 0.117 0.125 0.151 0.128

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting.

87

Table A2~3: Mean Budget Shares (%) Allocated to Individual At-Home Food Items

by Season and by Income Group

 

 

 

 

 

 

 

 

 

 

 

 

 

Share(%)

Lean Harvest Post-Harv Planting Avera e

Food LJMIH LIMIH QMIH LLMIH LIM H
Staples 29 29 26 40 34 29 40 30 30 37 31 30 37 31 29
Rice 68 61 59 65 68 61 65 67 61 60 64 59 65 65 60
Millet-Sorghum 22 22 22 23 23 25 28 18 20 34 29 28 27 23 24
Maize 4 3 5 5 3 3 1 5 3 1 3 3 3 4 3
Wheat 4 11 12 3 5 8 5 4 10 4 3 6 4 6 9
Fonio 0 l 0 0 0 0 0 0 1 0 0 1 0 0 0
Atieke 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0
Cassava 0 1 0 2 0 3 0 0 1 0 0 0 0 0 1
Potato 1 l 0 1 0 0 1 5 4 1 1 l 1 2 1
SweetPotato 1 0 0 1 1 2 0 0 0 0 0 0 0 0 1
Tota1100100100100100100100100100100100100100100100
Meatand Fish l6 16 21 17 13 23 16 21 22 15 13 16 16 16 20
Beef 66 55 61 69 63 66 48 36 81 57 72 66 60 57 68
Mutton 0 5 2 0 0 0 5 48 1 0 2 1 1 14 1
Poultry 0 1 7 0 5 0 7 1 0 3 0 2 2 2 2
DryFish 22 19 15 20 20 12 33 12 14 24 19 20 25 17 15
FreshFish 12 19 15 11 13 22 7 3 4 16 8 11 ll 11 13
Tota1100100100100100100100100100100100100100100100

Vegetables 12 12 14 ll 12 12 13 ll 12 ll 11 12 12 11 12
Leave 13 11 14 13 13 7 10 7 8 12 9 9 12 10 9
Okra 21 26 19 13 ll 10 11 8 7 26 22 13 18 17 12
Onion 18 21 23 31 25 27 32 24 25 23 24 29 26 24 26
Tomato 19 18 18 22 24 23 20 23 26 21 21 23 21 21 22
Beans 16 15 l3 13 19 21 17 26 25 13 19 16 15 20 19
Others 12 9 l3 8 8 12 10 13 10 5 5 10 9 9 11
Total100100100100100100100100100100100100100100100

Oil 5 3 5 3 3 3 4 4 3 3 3 3 4 3 4
Peanut Oil 86 77 76 73 74 70 66 86 86 77 74 70 76 78 75
PalmOil 10 13 13 16 10 10 19 3 3 14 19 18 15 11 11
SheanutOil 3 10 11 12 15 20 15 11 11 8 8 13 10 11 14
TotallOO100100100100100100100100100100100100100100

_S_ugar 6 6 6 6 6 6 5 7 6 5 7 6 6 6 6

 

Note: Non-bolded ﬁgures refer to budget shares within each commodity group while bolded ﬁgures
represent budget shares across commodity groups.

88

Table A2~3: Mean Budget Shares (%) Allocated to Individual At-Home Food Items
by Season and by Income Group (continued)

 

 

 

 

 

 

 

 

 

 

Share (%)
Lean Harvest Post-Harv Planting Avera e
Food LIMIH LIMIH LIMIH LIMIH LIM H
Others 13 13 10 10 11 10 11 9 12 10 9 ll 11 10 ll
Butter 3 1 0 0 2 0 0 0 0 0 0 0 1 1 0
Buttermilk 3 3 9 1 1 8 6 4 5 4 6 5 3 3 7
Fresh Milk 0 1 0 0 0 1 0 1 0 0 0 0 0 0 0
Cond.Sweet Milk 0 1 2 0 0 2 0 0 0 0 0 0 0 0 l
Powdered Milk 13 15 7 5 6 8 7 2 22 ll 5 21 9 7 14
Eggs 2 3 l 0 0 0 0 0 2 0 0 0 0 1 1
Peanuts 25 27 27 37 31 30 30 33 25 33 37 30 31 32 28
Seeds 6 4 3 6 7 3 9 7 3 7 6 3 7 6 3
Other Nut&Seed 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Coffee 3 4 2 4 2 0 1 1 1 1 2 1 2 2 1
Tea Lipton 2 2 4 0 0 0 1 0 1 0 0 l 1 1 1
Green Tea 0 0 5 0 3 5 0 3 0 0 1 1 0 2 3
Quinqueliba 0 0 1 0 0 2 0 2 1 0 0 1 0 1 1
Other Beverage 0 0 0 0 0 0 0 O l 0 0 0 0 0 0
Banana 0 1 l 1 2 0 0 0 3 0 0 0 0 1 l
Citronella 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Dates 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
Lemon 1 l 2 4 3 3 0 2 0 0 0 0 1 2 1
Raisins 3 0 0 0 0 0 0 0 0 0 0 0 1 0 0
Tamarind 2 3 3 0 0 0 5 3 3 3 5 3 3 3 2
Orange 2 0 1 0 0 0 0 0 0 0 0 0 1 0 0
Seas. and Spices 33 33 32 40 40 36 41 41 35 38 38 35 38 38 35
Tota1100100100100100100100100100 100100 100 100100100
Total FAH 81 79 83 87 79 83 89 80 84 81 74 77 85 78 82
FAFH 19 21 17 13 21 17 ll 20 l6 19 26 23 15 22 18
Total Food 100100100 100100100 100100 100 100 100 100100100100

 

Note: Non-bolded ﬁgures refer to budget shares within each commodity group while bolded ﬁgures
represent budget shares across commodity groups.

89

. .mﬁﬁwocog u w: 98 .mEszzﬁ n we .2:me .1. m deco—«ooze.
u 18.x ”2m 3:5 one .330— .O<mv so: no» mEszzﬁ em use .8323 do.“ magma—ma cos .28 6:8 no.2 2:sz=8 we
.EZNGOQC no.“ magma? m; .EEmEH 8m 2:89:82 N._ .0 £883 8m mammmzze 34 .< 5883 8m mESwoBmE coo .Eouoa
no.“ 3:ma mo .mBEquoﬁmo Sm 252m com .385 com 322823 comm 2m “co—mince 23%.. am now <Q¢ m.0<m 2C. 832

 

 

 

 

 

 

 

 

 

 

mm 22. mm 2 _ _ 2 SM 3. a wow 83 so:

mm 2m om om _ _ 8 mm x S N24 22cm :2:

om 4mm ow 2 _ _ mm 2 m S _m gm :3 223.2

mm com 2 2 _ 2 R as 3 mm am $8. 33
owaho><

mm mm om 2 _ _ em gm mm mm mom 5% :82

mm on. 8 2 _ _ om $4 2. am on. 38 emf

2 8m 2 2 _ _ 2. 9cm 2 2. Sm :2 2232

_m em 2 2 _ _ mm 22 mm _m am 2% 33
955:.—

Nm .3 cm 2 _ _ em mom 2. mm 4:. :8 :82

em Sm om ON _ _ 2. Se mm 3 e2. can :3:

2 2m 2 2 _ _ em mam 2. _m com 3.; 22::

mm eom 2 2 _ _ em EN 2 em 8... zoom 33

«noisy—tumou—

2 22 em 2 _ _ mm mm 2. am 24 emmm :82

em 82. a... _N _ _ :. 84 em 2. 32. £2 :3:

_N 5 om : _ _ mm 2 m :c x 8m $0.8. 2232

8 SM 2 2 _ _ em _mm 3 8 an EN 33
“meta:

mm C24 em 2 _ _ mm was 2. _e 2:. 3mm :82

R a? _m om _ _ 2. 2cm 8 K ”2. 22mm emf

_m as. E 2 _ _ am 24 3 2c. 0% 5a 22:2

mm m2. 8 2 _ _ mm mom 2. em 02 8. m 33
:31—

wE wE _ wE wE mE wE mE w: _ w w _ w Bov—

co: EEG—mu _Eo< ozom £222 E>mco£d £825. Ux> <:> _ 8m 589m 88?? Quocm cemmom

282:2 _ 2:885 _ .0950 coon

 

 

 

 

 

 

9.9.0 2:3:— z: can nemaom .3 mueem «Heatwave 59: me. hem b:3a=a>< 2.3.5: Z .225 :32 34.2 «Ben.

90

.NEEwEBE u w: 2: 2:82:52 n wE 2.3% u w deco—30:2 n .82 .08 2:... of. 8an

 

 

 

 

 

 

 

 

 

 

 

 

 

 

O 2 . .N N. O. 2. N ON 4. N N :82

O 2 . .N 2 .. 2. N NN 2 N N 22:

O. N. N NN .. . . e N ON 2 N N 2222

O N. O ON . . O N e ON 2.. N N 33

23... =<

O O O O O O O O O O .. N :82

O O O O O O O O O O N. N 22:

O O O O O O O O O O . . N 22...).

O O O O O O O O O O N O 33

uawsm

O O O O O O O ON NN O O N :82

O O O O O O O .N ON O O N :2:

O O O O O O O ON NN O O N 2223.

O O O O O O O ON ON O O O 33

=0

2 NN NN N 2 N NN .O . N .. 2. 5.23.

ON NN OO O 2 O ON Ne . N N 2. 2m...

2 ON NN N 2 N NN NN N e 2. N 2225. .

N. NN ON N 2 O NN ON . N .. .. 33

33503.»

N NN N . . N. N O N 2 2 O 2. =82

N NN O N. 2.. O O N 2 .N O N 22..

e .N N O. N. N O N 2 N. O a 2222

N NN N O O. O O 2. N. 2 O .. 33

222 2:. .32

2. N. 2 ..N 2. N.. N. N N. ON NN ON :82

2. N. ON ..N 2. 2. 2 O ... ON ..N ON 22:

2. N. 2.. NN 2. 2. . . .. ... NN ON ON 2225.

2. 2.. N. NN ON ON N. e 2 NN NN .N 33

.293. 350

ON 2 O ON . . NN O O N ON ON .2. :82

2 2 O NN O. «N O O N ON 2. NN 22:

NN 2 O 2. N. ON O O N NN NN 2. 2223.

2 2 O ON . . ON O O N ON ON 2. 33

I 33
wE wE wE wE wE wE wE w: w w w :82
co: EEG—m0 Eo< £62 £022 522—328 EEmE... U :> < :5 an 580.:— BEPE hwuocm 260“— 0801.3».
28052 22:82? .330 coon

 

 

 

 

 

 

 

9.9.0 2:8:— 3 2x; 2:529 wee..— oEe=¢< .3 18:22:80 25232 :32 "mtg. 03:.

91

.NESwoBE. n m: was £82.35... n w... 0NENNw n w auto—woo...— ” _Nov. .2.. 2.5. RC. 8ch

 

 

 

 

 

 

 

 

 

 

 

O N. N O. N. O. N O NN N. N N :82

N O. N N. O. O N N .N N. N N ONE

O N. O .N O. .. O O. ON N. N N 2222

O. N. . 2 N. O N N ON O. N N 33

22.8 ._<

O O O O O O O O O O . . N :82

O O O O O O O O O O N. N ON...

O O O O O O O O O O O. N 2.....2

O O O O O O O O O O O. N 33

uuusm

O O O O O O O ON NO O O N O.32

O O O O O O O ON OO O O O. ON...

O O O O O O O 2. ON O O O 2.222

O O O O O O O ON NO O O O 33

..o

.N .O ON O NN O. ON .N N O N O O.32

.N NO NO N NN .. ON 3 N O. O N ON...

O. OO NN N N. O ON .O N O O N 2.22

NN OO NN N ON N. NN .N N O. O N 33

83303.,

N NN N N. N. N O N N. ON O N :82

N NN O N. N. O. O N O. ON O O ONE

O ON N O. . . N O N N. N. O O 2.222

N ON O . . . . N O N N. N. O O 33

Sara ﬁca “so:

NO . . O. NN OO NO .. O O. ON NN ON :82

OO . . NN ON OO OO O. N N. .N ON ON ON...

2. N. N. NN 2. OO N. N O. . N NN ON 2222

NO O. .. ON .O NO .. N N. ON .N NN 33

835m .550

O. N. O NN O. NN O O N NN OO ON :82

O. . . O NN O .N O O N ON OO ON ON...

.N N. O OO . . NN O O N ON .N NO 22.2

ON N. O OO O. ON O O N ON .N .O 33

95—
wE wE wE mg. mp. wE mg. m: w w w 8 ,
O2. 5328 .29.. 2O... O.222 23.22.... 252.: o ..> < .5 a. 522. NERO $25 2.8. 022.-....
222.2 3.8 O8?

 

 

N:_E2.m

9.ch 2:3:— 3 ON: a

 

 

 

 

 

 

< 5 2.29 3520 EEO— oE¢=¢< .3 toga—Tancu 35.534 5.? «Sun.

92

.252?me u m: was .NENNwEE. u w... 2.8% u m £020.32... u Rev. 6.... 2...: 2; 6ch

 

 

 

 

 

 

 

 

 

 

 

 

O. N. O ON N. .. N N NN 2 N N :82

O N. O NN N. .. N N NN O. N N ON...

.. N. . NN .. N. N N NN O. N N 22.2

.. O. O NN N. N. N O. NN O. N O 33

22.8 __<

O O O O O O O O O O O. N :82

O O O O O O O O O O N. N ON...

O O O O O O O O O O .. N 2.5.2

O O O O O O O O O O N O 33

Luw=m

O O O O O O O .N ON O O O :82

O O O O O O O O. ON O O N ON...

O O O O O O O .N NN O O N 2.222

O O O O O O O NN NN O O O 33

.5

N. NN NO N N. N NN NN . N O N :82

N. .N NO N O. O OO ON . N N O ONE

O. ON NN O .N N ON NO N O O N 22.2

O. ON ON O O. O NN NO . N N N 33

83803.»

N .N N ._ N. N O O N. N. O N :82

O NN O O. N. O. O O N. NN O O ONE

O NN O O .. N O N N. O. O O 2222

O N. N O. .. N O O N. N. O O 33

:m_.& $35 «so:

OO O. ON ON NO NO ON N. N. ON NN ON :82

NO O. NO NN NO NO ON N. N. ON ON ON ON...

NO . . N. NN OO OO N. N N. NN ON ON 2.....2

NO N. .N NN ON OO N. O. N. NN ON .N 33

2.35m .050

ON N. O NN .. ON O O N ON ON .O :82

N. N. O ON O. NN O O N NN OO NN ON...

NN N. O .O N. ON O O N ON ON NO 22.2

ON N. O NN N. ON O O N ON .N NO 33

00m”—
wE wE wE wE wE wE wE w: _ m m m Bow.
8: .5228 .22.. 2.0.. 22.. 232.3... O.O.O.ON u ..> < .5 _ a. 222.. 22.2 2.2.... 2.8. 20..-...
222:: 3.835 _ .330 noon?

 

 

 

 

 

 

9.9.0 2:8:— .3 22.8957. a. 2x; 2.59.0 62...— 2.3.23. .3. cot—£5.80 25.53,. "Nag. «Bah

93

.NEN.wo.o_E H w: E... $5.835... n w... “gnaw H w auto—30:2 n 33. .9... 3...: 2:. 332

 

 

 

 

 

 

 

 

 

 

O N. O .N N. : N N ON N. N N :82
O. NN . NN NN N. O N. ON N. N O ON...
O. . . O NN O. N. N . ON N. N N 2.22:
N N. O O. O O N . ON N. N N 33
23.32
O O O O O O O O O O O. N :82
O O O O O O O O O O N. N ON...
O O O O O O O O O O N. N 2.2....
O O O O O O O O O O N N 33
.awzm
O O O O O O O N. ON O O N :82
O O O O O O O O NN O O N ON...
O O O O O O O .N NO O O N 2%.:
O O O O O O O NN NN O O O 33
..o
N. NN NN N N. O ON NN . N N O 52>.
N. NN OO O N. N .N ON . O O O ON...
N. NN ON N N. N NN NN . N N O 2.6.2
N. NN ON N N. O NN NO . N N O 33
835%...»
N NN O .. N. N O N O. N. O N :82
O .N O. N. N. O. O N O. .N O O ON...
N NN O O. N. N O N N. O. O N 22...).
N NN N N O N O N .. O. O N 33
ON... E... .8...
OO N. N. NN OO NO N. N O. ON . N NN 52).
NO . . ON NN NN OO N. . N. NN NN NN ON...
NN O. N. N. OO NN N. N O. NN NN .N 2.22.).
.N N. N ON ON NN .. N ON NN NN ON 33
8.9.5 .350
NN N. O OO N. NN O O N .N .N NO :82
ON N. O NN O. NN O O N NN NO ON ON...
ON NN O NO N. ON O O N NN NN NO 2.222
O. O. O NN .. ON O O N ON OO .O 33
I on.“
wE wE w... wE w... w... w... w: w w w .81
O8. 5228 O3. 2...... O.222 532.3... 5.52.. 0O? .25 .2. £22. OOOOONO NNOOON. 2.8. Naomi
£2055. N588; A530 woo...

 

 

 

 

 

 

 

 

9.9.0 2:3..— .3 Cash—oh 5 ANY... 359.0 6.5,.— oEcﬂ..< .3 68.52280 3.3232 "mama. 2......

94

.NEN.wo.o.E u w: E... .N80.w....E H w... N88... u w aN0...o.Noo..O. u .82 .0... 3.55 0:... .032

 

 

 

 

 

 

 

 

 

 

N O. O ON O. O. N N ON O. N N 522
O N. O NN N. .. N N NN N. N N ON...
O N. O .N O. .. N . ON N. N N 2.02.).
N N. O O. O N N N NN N. . O 33
£3.32
O O O O O O O O O O .. N :82
O O O O O O O O O O N. O ON...
O O O O O O O O O O N. O 2.22.).
O O O O O O O O O O N O 33
.aw5m
O O O O O O O NN NN O O N ON2).
O O O O O O O NN NN O O O ON...
O O O O O O O NN NN O O O 2.22.).
O O O O O O O ON ON O O N 33
..o
N. NN ON N N. N ON NO . O O N 5.22
NN ON ON N O. N NN NO N O N O ON...
O. ON NN N N. N .O NN N O O N 2.22:
O. ON NN O N. N OO NO . O N N 33
830.03.,
O ON N O. .. N O N O. O. O O 52).
O O. N .. N. N O N N. N. O O ON...
O N. N O. .. N O N O. O. O O 2.0.22
N ON N N O N O O N. N. O N 33
ON... OON .8...
NO O. N. ON ON NO O N N. NN ON .N ON2).
OO N. NN ON NO NO . . N O. . N ON ON ON...
NO N. N ON OO OO N N O. . N NN ON 2.22.).
ON O. N ON NN NN N N .N NN NN ON 33
8.9.5 .050
ON N. O OO N. NN O O N .N OO NO :82
N. O. O NN . . ON O O N ON NO OO ON...
.N O. O OO N. ON O O N NN ON NO 2.0.2.).
ON N. O .O N. NN O O O NN ON NO 33
003
wE wE wE w... wE m... w... w: w w w .03.
8.. EEG .20.. OOO. O.32 25%.... 2.52. 0.5 < ..> a. 222. 22.3 NNOON. NOS. Naomi
£3052 NEESS 3.00 .00on

 

 

 

 

 

 

 

@580 0.53..— 3 N02 513. 2.59.0 too..— 0Eol-.< .3 10.5.2230 3502.57. no-3. 03.;

95

Table A2-10: Contribution of At-Home Food Commodities to Protein Availability

 

 

 

 

 

 

 

 

 

 

 

At-Home Food Mean Daily Protein Availability NutrienTSEurce
Coomodity Groups (Grams/AB) (%)
L H PH P Av 1: H PH P Av
Rice 17 17 18 16 17 47 50 52 49 50
Millet-Sorghum 15 15 14 14 15 43 43 41 43 42
Maize 2 1 l 2 2 6 4 4 5 5
Wheat 1 1 l 1 1 3 2 3 2 3
Other Cereal 0 O 0 0 0 0 0 O O 0
Atieke 0 0 0 0 0 0 0 0 0 0
Cassava O 0 0 0 0 0 0 0 O 0
Potato 0 O 0 0 0 0 0 l 0 0
Sweet Potato 0 0 0 0 0 0 1 0 0 0
Meat andfisb 12 II II 39 II 230 f9 f8 16 138'
Beef 5 6 5 4 5 43 52 49 52 49
Mutton 0 0 1 0 0 l 0 5 1 2
Poultry 0 O 0 0 O 2 O 1 l 1
Dry Fish 5 4 4 3 4 41 36 41 38 39
Fresh Fish 2 1 0 1 1 12 11 4 9 9
Vegetabfec 5 4 4 3 4 39 37 6 7
Leave 1 1 1 1 1 19 18 12 16 17
Okra 1 0 0 1 1 16 12 9 16 13
Onion 1 1 1 l 1 9 14 17 18 14
Tomato 0 0 0 0 0 5 9 9 11 8
Beans 0 0 l 0 0 6 11 15 8 10
Others 2 l 2 1 2 44 37 37 31 38
611 36 30 40 36 30 36 36 30 36 36'
Peanut Oil 0 O 0 0 0 O 0 0 0 0
Palm Oil 0 0 0 0 0 0 0 O 0 0
SheanutOil OJ 04 03 03 OJ 03 03 0 0 0
..-...- §ugar go o a p o a a ‘0
ml others 8 39 49 7 38 13 T5 T5 T4 f4
Butter 0 0 0 0 0 0 0 0 0 0
Buttermilk 0 O 0 0 0 2 1 1 1 1
Fresh Milk 0 0 0 0 0 0 0 0 O 0
Condensed Sweetened 1V 0 O O 0 O l 0 0 0 O
Powdered Milk 1 1 1 l l 8 6 l4 7 9
Eggs 0 0 0 0 0 4 0 0 0 1
Peanuts 5 7 6 5 6 64 74 64 70 68
Seeds 1 l l l 1 9 9 9 9 9
Other Nuts and Seeds 0 0 O 0 O 0 0 0 0 0
Coffee 0 0 0 0 0 0 0 O 0 0
Tea Lipton 0 O 0 0 0 0 0 0 0 0
Green Tea 0 0 0 0 0 0 0 0 0 O
Quinqueliba 0 0 0 0 0 O 0 0 0 0
Other Beverage 0 0 0 0 0 0 0 0 0 0
Bananas 0 0 0 0 0 O 0 0 0 0
Citronella 0 0 0 0 0 0 0 0 0 0
Dates 0 0 0 0 0 0 O O 0 0
Lemon 0 0 O 0 0 0 0 0 0 0
Raisins 0 0 0 0 0 0 0 0 0 0
Tamarind 0 0 0 0 0 l 0 1 1 0
Orange 0 0 0 0 O 0 0 0 0 0
Seas. and Spices 1 1 1 1 1 ll 9 10 11 11
Total 6W7 W

 

Note: Non-bolded ﬁgures refer to budget shares within each commodity group while bolded ﬁgures
represent budget shares across commodity groups. L = August = lean season, H = November = harvest, PH
= February = post-harvest and P = May = planting.

96

Table A2-11: Mean Budget Share Allocated to Fruits, Nuts, and Dairy Products At
and Away From Home

 

 

Commodities L H PH P Avg.
Nuts
At-Home 3.9 3.9 3.5 4.4 3.9
Away-From-Home 0.8 0.5 1 .7 0.4 0.9
Fruits
At-Home 1.0 0.6 0.7 0.5 0.7
Away-From-Home 3 .8 1 0. 1 3 .8 3 .8 5 .4
Dairy
At-Home 2.8 1.5 2.0 2.4 2.2
Away-From-Home 6.9 6. 1 6.2 4.4 5 .9
Others
At-Home 92.3 94.0 93.8 92.7 93.2
Away-From-Home 88.5 83.2 88.3 91.4 87.8

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting.

97

CHAPTER 3

ESTIMATING THE IMPACT OF SEASONAL CHANGES IN REAL INCOMES
AND RELATIVE PRICES ON HOUSEHOLDS’ CONSUMPTION PATTERNS IN
BAMAKO, MALI, USING THE ALMOST IDEAL DEMAND SYSTEM MODEL

3.1. Introduction

In 2000, more than 58 percent of the Malian urban population lived below the poverty
line, and urban unemployment approached 70 percent (U SAID, 2000). Urban households
have faced higher and more variable food prices, lower real wages, growing
unemployment, and reductions in social services (e.g., health and education) since the
inception of the early 19803 structural adjustment reforms (Tefﬁ et al., 1997). Such
economic environments have resulted in urban households having real incomes that can
vary signiﬁcantly across seasons. As shown in the previous essay, in 2000-2001, Bamako
households’ mean real expenditures decreased by 38 percent between the lean and post-
harvest season, increased by 4 percent between the harvest and post-harvest season, and
dropped by 18 percent between the post-harvest and planting season.

Concerns about real income stability, which is an important determinant of
household food security, have drawn the attention of policy makers towards the design of
safety net programs to protect at-risk households’ food entitlements (Sahn, 1989).
However, the formulation of such programs requires substantial knowledge about urban
households’ seasonal food consumption patterns and the forces causing changes in those
patterns. To date, two empirical consumption studies (Rogers and Lowdermilk, 1981 and
Reardon and a1, 1999) have been conducted in urban areas of Mali. These consumption

studies have focused mainly on estimating the Engel relationship between food

98

expenditure and income using cross-sectional data. This study uses the complete systems
approach to estimate demand parameters using household-level panel data in order to
provide a clear understanding of households’ consumption patterns in Bamako. Panel
data, unlike cross-sectional data, which tends to reﬂect long-run adjustment processes,
allows estimation of short-run income and price elasticities (Timmer, 1983).

This study hypothesizes that Bamako households’ consumption patterns are
responsive to changes in their real incomes and that the relationship between household
income and food and non-food consumption patterns will change from one season to
another. This implies that the effectiveness of speciﬁc programs or policies will depend
on the economic conditions prevailing at their time of implementation (Skouﬁas, 2002).24
The impact of seasonal changes in real income and relative prices on households’
consumption patterns has not been investigated in Bamako, Mali, prior to this study.

The general objective of this study is to examine the impact of seasonal changes
in real incomes and relative prices on households’ consumption patterns in Bamako.

The speciﬁc objectives of this essay are as follows. First, the study seeks to estimate

demand parameters using the almost ideal demand system model for each season.25
Second, the analysis aims at computing income elasticities for different seasons in order
to determine if households’ consumption patterns are responsive to changes in their real

incomes. Third, the study derives own and cross-price elasticities for different seasons in

order to identify households’ seasonal substitutions among and between broad

 

2‘ Temporal targeting mechanisms, such as seasonal income transfers to low-income households and
seasonal imports of rice, are examples of programs or policies that are season-speciﬁc.

2’ The seasons are defined as follows: August = lean, November = harvest, February = post-harvest, and
May = planting.

99

commodity groups. Finally, the study performs sensitivity analyses on households’
consumption by varying estimated income and price elasticities.

3.2. Methods

3.2.1. Commodity Aggregates and Weak Separability

This study will assume that consumers’ preferences are weakly separable in order to
simplify the modeling of the consumers’ consumption decisions. The reasoning behind
the concept of weak separability is that the optimization problem is intractable for the
consumer if the demand for every commodity is a function of the prices of all other
commodities (Deaton and Muellbauer, 1980a). To simplify this problem, we may assume
that the consumer partitions total consumption into groups of goods, so that preferences
within groups can be described independently of the other groups (Pollak and Wales,
1992). Price changes in one good will then affect only other goods in the same group
directly. Commodities in any other group will only be affected through the change in
total expenditure as the price change makes the consumer richer or poorer.

Under the assumption of weak separability, the consumers’ simultaneous
decision-making process is broken into three steps by adopting a three-stage budgeting
process, as depicted, below, in Figure 3-1. In the ﬁrst stage, households allocate their
total expenditures among seven broad groups of commodities: (1) Food, (2) Durable
Goods, (3) Semi-Durable Goods, (4) Health, (5) Energy and Utilities, (6) Other Non-
Durables (Hygiene and Tobacco), and (7) Services. 26 In the second stage, households
allocate their food expenditure on seven food groups: (1) Staples, (2) Vegetables, (3)
Meat and Fish, (4) Oil, (5) Sugar, (6) Other Foods, and (7) Food Away From Home. In

the third and ﬁnal stage, households allocate their staple group expenditure to (1) Rice,

100

(2) Millet-Sorghum, (3) Maize, (4) Wheat, and (5) Roots and Tubers. Hence, it is thus

assumed that preferences are weakly inter-temporally separable, that food is weakly

separable from non-food commodities and that staples are weakly separable from the

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

other food groups.27
Total
Stage I l
l _.I_ __L;__ _.L_. l l
Foods Durable Semi-durable Health Energy and Other Non- Services
, Goods Goods Utilities Durable Goods
Stage II I
| I l | I , l l -
Staples Meat and Vegetables Oil Sugar I Other Food Away
Fish Foods From Home
Stage III 7
l - 1 _, 1
Rice Millet- Maize Wheat Roots&
I Sorghuml ‘ Tubers ‘

 

 

 

 

 

 

 

 

 

 

Figure 3-1: Three-Stage Budgeting Process for Urban Households in Mali

3.2.1. The Almost Ideal Demand System (AIDS)

The study will model, as shown in Figure 3-1, the allocation of (i) total expenditure

(Stage 1), (ii) food expenditure (Stage II) and (iii) staples expenditure (Stage III) for each

season separately and for the entire data pooled (yielding 160 observations). The Almost

Ideal Demand System (AIDS), developed by Deaton and Muellbauer (1980a, 1980b), is

used to estimate demand equation parameters. The AIDS is a demand system that is

superior to its predecessors and is recommended as a vehicle for testing, extending, and

improving conventional demand analysis because it is linear in the parameters and hence

 

2‘ Table A3-1 of Appendix 3 presents the deﬁnition of the various commodities and commodity groups.

27 The assumption of weak intertemporal separability allows each demand equation in each season to be
expressed as a function of prices and income in that season alone, so that goods in each season form a
closely related group with only general relations between seasons (Deaton, 1990). Thus, once households
make the consumption-saving decision, the problem left is for the household to allocate total income
among goods at given prices.

101

simple to estimate, and most satisfactory in terms of being able to estimate and test the
predictions of consumer demand theory (Green & Alston, 1990; Alston et al., 1994). The
AIDS model is derived from a consumer cost/expenditure minimization problem as
deﬁned by a cost/expenditure function that expresses the minimum expenditure necessary
to reach a speciﬁc utility level at a given set of prices. The AIDS, formulated in terms of

the budget shares, is speciﬁed in this study as follows:
Wm: 00+ 25' 'Yij In stt + Bi ln (Xst/Pst) + 9i In AEst + u sit (1)

The dependent variable W,,-, is budget share of good i of the stage 8 model (s=1, 2,
3) at season I (t = 1, 2, 3, 4). The independent variables of the equation include P,,-, as the
price of each good i of the stage s model at season t; X“ as the household real
expenditures per adult equivalent (AE), AB is household size in adult equivalents and PM
is an overall price index.28 Following Moschini (1995), the price index is approximated
by a log-linear analog of the Laspeyres price index in order to maintain the linear
speciﬁcation. 29 In equation (1), the 7,,- parameters measure the change in the ith
commodity’s budget share in response to a 1 percent proportional change in the jth
commodity price with real income held constant. The [3, parameters, or marginal budget
shares, represent the change in the ith commodity’s budget share with respect to a change
in real income, holding prices constant. The 0, parameters represent the change in the ith

commodity’s budget share due to a 1 percent change in household size, with incomes and

 

2’ Other independent variables could in theory be added, but in this study, degrees of freedom
considerations restricted their use.

I:
29 The log-linear analog of the Laspeyres price index is deﬁned as: In P. = Z W,0 In P,
i

l02

prices held constant. The restrictions, imposed on equation (1), to ensure theoretical

consistency for the almost ideal demand system are:

Adding-up: 2i (Xi = 1, 2i 'ij = 0 (1a)
Symmetry: Yij = 1m (1b)
Homogeneity: Eijj = 0 (lo)

Adding-up requires that the demand functions must satisfy the linear budget
constraint (marginal budget shares derived from the system must add up to one).
Homogeneity of degree zero in all prices and income means that the scaling of all prices
and income has no effect on the quantity demanded of each good. Symmetry entails that

the cross-price derivatives of the Hicksian demands are symmetric.

The Chow Test, which is simply an F test, will be performed to test the hypothesis
of the constancy of the parameters of the demand system across seasons. The study will
test for the stability of the coefﬁcients under the null hypothesis that the estimated
income and price elasticities do not vary across seasons. Once the F-statistic is
computed, we will compare that value against the critical F-values at a chosen
signiﬁcance level (e. g., 1%, 5%, and 10%). If the F-value is less than the critical F-value,
then we do not reject the null hypothesis, meaning that the impact of changes in Bamako
households’ real incomes and relative prices on consumption patterns is constant across
seasons.

The income and price elasticities, evaluated at the mean budget shares, will be

derived from the parameter estimates equations using the following:

m = 1+ [[3, / w i]: Expenditure elasticity (2)

103

é, = -1-B,+ [7,, / w ,]: Marshallian (uncompensated) own-price elasticity (3)
5,, = [7,, * / w,] - [3, [w, /w,]: Marshallian (uncompensated) cross-price elasticity (4)

11,, = g, + 11,. w, : Hicksian (compensated) price elasticities (5)

Where, 11, is the expenditure elasticity, g, is the Marshallian (uncompensated) own-
price elasticity, Q, is the Marshallian (uncompensated) cross-price elasticity, and 1],, is the
Hicksian (compensated) price elasticities. Following Lazaridis (2003), a system of share
equations based on equation (1) and subject to the restrictions (adding-up, homogeneity,
and symmetry) is estimated using iterative Seemingly Unrelated Regression (ISUR)
method for constrained systems developed by Zellner (1962).30 The adding-up property
of demand causes the error covariance matrix of system to be singular, so one of the
expenditure share equations is dropped from the system to avoid singularity problems.
The estimates are invariant to which equation is deleted from the system, if no
heteroskedasticity is present, because the coefficients of the omitted equation are
recovered by using the adding-up restrictions.

3.2.2. The Data31

The panel data used in this study is from a 2000-2001 survey undertaken in Bamako by
the Direction Regionale du Plan et de la Statistique (DRPS) of the Direction Nationale de
la Statistique et de l’Informatique (DNSI) and the Proj et d'Appui au Systéme

d'Information Décentralisé du Marché Agricole (PASIDMA) of Michigan State

 

3° Demand equations are related because the error term across equations are correlated by the fact that the
dependent variables need to satisfy the budget constraint (the budget shares must sum up to one). Although
in this case the OLS estimates are consistent and unbiased, the SUR estimation method yields estimates that
are more efﬁcient.

3' The deﬁnitions and summary statistics of the variables are presented in Table A3-2 of Appendix 3.

104

University (MSU), the Assemblée Permanente des Chambres d’Agriculture du Mali
(APCAM), and the Centre d’Analyse et de Formulation de Politiques de Développement
(CAF PD). The four surveys covered 40 Food Consumption Units (FCU), the sample size
being the same in each round. The same 40 households were tracked over time and
interviewed in all four periods. There was no sample attrition.

The total expenditures in CF A Francs aggregates were computed following
Deaton and Zaidi (1999).32 The budget shares were calculated as the expenditure on a
good as a fraction of total expenditure. Unit values, used as proxy for prices, were
computed as a ratio of total household expenditure on a good divided by the total quantity
consumed of the good. Households’ real incomes are proxied by total real expenditures
and are calculated by deﬂating their nominal income (total expenditures) by the
Laspeyres price index. The data on household size was converted into adult equivalents
using the following scales: male > 14 years = 1.0; female > 14 years = 0.8; children = 0.5
(Duncan, 1994).

3.3. Empirical Results

This study seeks to examine the impact of seasonal changes in real incomes and relative
prices on households’ consumption patterns in Bamako, Mali. In this section, the
empirical results are presented in three parts. First, the analysis begins with an evaluation
of the estimated coefﬁcients. Second, the income and price elasticities are examined.
Third, sensitivity analyses are performed on households’ consumption by varying

estimated income and price elasticities.

 

32 Detailed information on the construction of the expenditure aggregates is provided in Appendix A2-l.

105

3.3.1. Coefﬁcients

The aim of this section is to examine the estimated demand parameters in order to
investigate the effects of changes in real incomes and relative prices on the expenditure
share of each commodity aggregate by season, holding ﬁxed all other factors. Results of
the Chow test are also presented in order to assess whether the estimated income and

V price parameters are stable across seasons.

3.3.1.1. Stage I Coefﬁcients

Table 3-1, below, displays the estimates of the parameters, their associated t-values, and
the Chow test results for the Stage I model. The parameters of the dropped equations,
Services, were recovered using the adding up restrictions. The dependent variables are
the expenditure share of each commodity aggregate. In addition to prices and real
income, a household size (in adult equivalents) variable was included in the regression
models. A total of 7 equations were estimated for each season.

First, the results indicate that price, income, and household size factors account
for a substantial part of the observed variation in the budget share devoted to the
commodities considered. For instance, the goodness-of—ﬁt measure, R2, ranges between
0.65 and 0.82 for food, suggesting that, as a group, the price, income, and household size
variables explain about 65 to 82 percent of the observed variation in the food budget
share. The R2 for non-food expenditure categories is lower than that for food. For
instance, the R2 ranges between 0.16 and 0.48 for health and 0.13 and 0.21 for semi-
durable goods.

Second, the estimated results show that coefficients of a great number of the

explanatory variables in all seasons are statistically signiﬁcant at the 1%, 5%, or 10%

106

level, indicating a large degree of price and income responsiveness of budget shares. For
instance, the results reveal that the estimated income parameters for food are statistically
signiﬁcant at the 1 percent level in all seasons and for the pooled data.

Third, the signs of the price and income coefficients are consistent with the
theory. The marginal budget share estimates, [3, for food, are all negative, are ~0.130,
-0.224, -0.186, and —0.207 for the lean, harvest, post-harvest, and planting seasons and
-0.202 for the pooled data. These results suggest that food is a necessity, as food
expenditures take a smaller percentage of income as households get richer. In contrast,
the coefﬁcients for durable goods and health are all positive, suggesting that these are
luxuries, since at higher income levels, households increase the proportion of the total
budget allocated to these non-food items.

Finally, the Chow test results indicate a certain degree of non-constancy of price
and income parameters across seasons. For instance, the Chow tests results showed that
there was statistically signiﬁcant structural change, at 1 % signiﬁcance level, across
seasons in all the estimated income and price coefﬁcients for the food equation. This
means that the impact of real income and relative prices on the food budget share is not
constant across seasons. The null hypothesis of stability of the income parameters was
also rejected, at the 1 % signiﬁcance level, for durable goods and health. The Chow tests
results also showed that there was statistically signiﬁcant structural change, at least at the
10 % signiﬁcance level, in one or more of the estimated price parameters for all demand
equations in the Stage I model.

These ﬁndings conﬁrm the main hypothesis of this study, as they indicate that

Bamako households’ consumption patterns are responsive to changes in their real

107

incomes and relative prices and that the relationship between household income and
relative prices and food and non-food consumption patterns changes from one season to

another.

108

008038. as $.68 05 main 823002 083 £80883 «.8 00:7. 00:38 803:3 05 00.8 03238 8080 08053.0.

0: 03 22: 8%: E 08 88: 30:0 05 .08 828i m .08. E8808 2 05 8 800050»; b.8523.“ 0.8 386.8000 803588 05 85 88:05 828
-8 0.00 ”8850088 5 00:08. 08 828-8. 800.8318 :38 u m< .8238 .80 0088 u E 800 A0880... 8cm ocomwhzv 83838-00: 850 80 00E. n
on 88:8: 08 38:0 .80 0088 n 3 :28; 80 00.5 n X .800w 0.083360... .80 0080 n ma .800w 03830 .80 0088 n Na .8008 .80 00:0 u 3 800—8338
:35 .08 0036898 .02 .308 u Ex ”was—BE n .32 u n— 05 “88:80“. M 80508 n In .8822 .1. 820.302 n I .0083. :8_ n 88w=< u A ”302

 

 

 

 

  

 

 

   

 

 

 

 

 

 

 

8.8 8.8 one 8.8 38 who 83 88 who :8 88¢
5.8 E8 :88 828 :88 :88 888 888 888 828
88.8 88.8 888 no.8 888 _88 888 .88 888 888 m<
888 :_.8 888 :88 :88 88.8 888 888 888 88.8 :88 888
~28 Bod- _ 8.8 88.8 88.8 :88 83 88.8 88.8 88.8 2 88 £88 8
888 828 888 888 8.8 :88 888 A88 888 888 5.8 88.8
83 88.8 888 888 888 888 88 88.8 83 ~88- 88.8 :88- a
888 888 :88 :88 :88 888 888 888 888 5.8 888 E8
82.” 888 888 888 88.8 888 88 88.8 88.8 88.8 588. 888 a
8: .8 88 8 5.. 8 :38 888 888 88 .8 888 :88 8 _ .8 8 _ . 8 888
o _ f _ 8.8 38.8 88.8 :88- ”8.8 an}. :88 8.8 88.8 88.8 888 a.
88 .8 :88 888 8 _ . 8 :88 :88 88 .8 :88 8 _ . 8 : _8 5 .8 888
3.2 88.8 888 :88 :88 88.8 and 888 :88 888 888 28.8 8
E .8 :88 :88 5.8 888 888 88 .8 888 888 08.8 888 A8,. 8
82 88.8 888 88.8 8 8.8 $88 8%.... :88 .888 o _ 88 888 a _ 8.8 8.
888 888 888 88.8 888 :8. 8 888 8.8 888 888 :88 888
.93. :88 888 o _ 8.8 88.8 a _ 88 8.5. E88 E88 2: .o 88.8 88.8 8
888 5.8 :88 :88 888 888 888 :8 8 888 888 888 5.8
83 8: .8 288 N88 3.2 E; :8 ~88 888 3 _ .8 8.8.8 3 _ .8 8x
84.8 :88 :88 888 888 888 888 :88 888 888
. - . - . - . - 88 .. o ._ m: 88.. SM: 82 .28
32:0 38.: a E n 320 828; .8 _ E _ m _ 0
880 2825 88.: 833:3
80 0.3.5 9.3.88an 833...; 2.0—002.2—

 

380: — owaum .3.. 3.38% 88h. 30:0 88 83:58— 8888...:— "Tm 038,—.

109

8080.88. a: 3.28 05 8.8 8020800.. 0.03 8802.88. 8. 002.... 2088.00 80.20m 05 20.. 03.235 80.80

888288 02 8a 0.0....- 828. 2. 0.8 80. 3020 05 20.8 829.8 ._0>0_ “2080.. o. 05 8 888.288 b.8888... 0... 820.0508 0082.80 05 85 38.02.
8287. 20.. 80852028. 2. 00:08. 03 82.3-..- ..20_0>_=c0 :38 u m< 803.8 80 00.3 u 2. 22a A0830... 828 020.82.: 83828-202 .050 80 00....
u on 88:8: 82.. 3.020 .80 00.... n mm 8.8.. .«0 00.... u E. .muoow 03823.28 .80 00.5 n ma .mcoow 03822 .20 00.... u an .8008 .80 00.5 n 2. 42033200
.30.. 20.. 83.82098 .8. .80. u Ex 82822.. n .3.). u n. .028 888F808 H 88.50“. n F. .8028. u .0nE0>0Z u I .2088 28. u 8.83.. n -_ ”0.02

 

 

 

 

 

 

 

N3 83 mm... 88¢ 8.8 m3 m..¢ .N.¢ 8.¢ 88¢ Essa-m
.28 888 8...: 5.8 .88 5.8 8.88 5.8 .88 888
88¢- .3? 88¢- 88¢ 88¢- ¢¢¢.¢ 88¢- 88¢ 28¢- 38¢ 0<
5.8 5 .8 .88 888 888 :88 8.. .8 :88 .88 888 .88 .88
8: .¢¢.¢ .¢¢.¢- .¢¢.¢- 88¢- ¢¢¢.¢ 83 28¢- 88¢ 88¢- .888 .888 8.
.88 .88 888 .28 .88 888 888 888 .88 888 5.8 88.8
.88. m¢¢.¢ ¢¢¢.¢ ...¢¢.¢ 88¢ 88¢ 88.. .¢¢.¢ 88¢ 88¢ m¢¢.¢ 888 8..
888 888 :88 88 8 .88 .5. 8 8m .8 a. .8 .88 .88 :88 :88
8......- .¢¢.¢- N¢¢.¢ ¢~¢.¢ ¢¢¢.¢ 38¢- 3m. .88. ¢¢¢.¢ ¢.¢.¢ 88¢- m¢¢.¢- 8..
8.8 .888 5.8 :88 5.8 5.8 888 5.8 8.8 888 .888 888
88¢ 88¢ 88¢ N¢¢.¢ mm¢¢ 88¢ 88.. . .¢.¢- .¢¢.¢ 28¢- .888 888 8..
:88 .88 8. .8 88 8 .888 888 «8 .8 a. .8 8. .8 :88 5w 8 .88
:¢.. .88 .¢¢.¢ 28¢- ¢¢¢.¢- ¢.¢.¢- ¢8..~ 88¢ 83. 88¢ ¢m¢.¢ 8 ..¢ 8.
.88 888 58 88.8 888 .88 888 .38 888 m... :88 :88
8.. .. :8? 38¢- 88¢- .¢¢.¢- 88¢- ¢.~.. 88¢- 88¢- 38¢ . .¢.¢- 88¢- 8.
8. .8 888 :88 8 . .8 5 .8 .38 «8 .8 88¢ 8. .8 : .8 8 .¢ 58
¢.¢... 3 ¢.¢- 88¢- m¢¢.¢- ¢ . ¢.¢- m¢¢.¢ m8.” ¢¢¢.¢ m . ¢.¢- N¢¢.¢- 88¢ m . ¢.¢- ..
8.. .8 .888 888 888 5.8 .88 88 .8 5.8 .R. 8 :88 888 5.8
8.8 38¢ E . .¢ E¢¢ 88¢ ¢¢¢.¢ 8.. 88¢ ¢m¢.¢- 8... .¢ m . ¢.¢ 88¢- 8x
888 .88 5.8 5.8 888 888 83 .88 8.88 .28
88.8 ¢ . ¢. 8 88¢- N83- ..33- 88¢- . :8 8.5.88 .¢¢.¢ .8 .¢ .28
32.0 _ 8.8.. _ .. _ E _ 2 F A >6.8 8.8.. _ .. _ 2.. I _ A 8.2.2;
530: $0000 o_pw..:Q.._Eom
.0 0.32m 0.2882093— "038...; ~20—020809

 

90552008 _0802 _ $85 20.“ 8.58m 80,—- 3020 82.. 8:22.88”— ..808888 "TM 038%

110

donuts“: a: 9:23 05 ME? 82382 395 280828 3 8:7. 3:930 3228 2: .08 032E?“ 9.9:»

235% o: 2m 20;... 6.2.3 E 2m $8 325 05 .8 mos—~38 ._o>o_ :8an S of 8 38.2ch 3.85:“: 2m $56508 voumﬁumu 05 65 2865
3283 Son "3355th E 3:32 2a mos—$9 853230 :22“ n m< Hoqum mo 83 u E was A0838. vcm 259$: 3383.3: 550 .3 nova
n on .325: uca 355.3 corn u mg £30; no 83 u 3 "$50» 235358? 3.5 n mm 53% 0383 .«o 83 n mg 608 he Bra n E 303223
:33 .3 836833 in: .88 u &x Eczema n >22 n L van 76>;me M 80350“— n In. awning u .onEoSZ u I .2033 32 u ~m=m=< n A “302

 

 

 

 

 

 

 

 

 

oo..c.--.mc..o.--.mm.o 8.3 8.3 _3 «3 m3 ”3 33 8229.-“
Sq: 8m: $28 838 3.: 3.: 93.8 3.8 838 5.8

:3- $3 83 53. 38.8 83 83 23 83- 83 :3 83 83- :3 $3 m<
3.8 3.8 5.8 Go: 3.: 5.: 5 .8 5.8 3.8 5.: 38.: 30.8

83 ”N3 83 $3 83 NS.” 83 83 23 R3- :3 82 3.8 3.8 3.8 3.8 83- E
888 5.8 $8 2.8 3.: 8m: 3.8 388 :8: 838 E: 5.8

83 S3 23 R3- :3 £3 33. 23- .808 23 $3. $3 83 :3- 83. £3 83- a
$8 838 3.: 838 E: 858 858 $8 $88 3.8 2.8 888

3.8 R3- 38 E3- 83- 3.. 83 :3- 83- 883 83.8 $2 $3 $3 $3 8.3 33 3
$8 an: 3.8 $8 5.8 5.8 858 638 838 8w: 3.8 5.:

S3 s3- 53. B3- 83 B3 83 83 83 83 33 $2 53. S3 83 83 :3- E
3.8 838 5.8 688 9:8 608 5.8 8.8 38 £58 3.8 :38

3.8 83 3.8 83. 3.8 38% S3 S3 83 83 83- wow. 53. 83 83 B3- 83- a
3.8 8&8 63.8 888 8.8 :8: 888 858 3.: 38 E8 388

S3- __3 83- 83.8 :3- SE 53 83 $3- 53 $3 3.. 83- S3- 33. 83.8 83 mg
. .. 88.8 3.: 3.8 8.88 2.88 388 3.8 888 83.: 5.8 888 E8

38 3.8 83- £3. 38 SE 83. 83 3.8 3.8 53- 33.. 83- 3.8 3.8 3.8 83 a
3.8 3.8 30.8 388 :8: 338 3.8 an: 3.8 688 30.: E:

53 83 $3 :3 3.8 M32 38 :3- R3- 38 3.8 $2 83- 83 :3- 88.8 3.0 .8
888 5.8 3.: 8.8 308 f8 38 $.88 $38 3.:

8:3 8:8 53. 83.3 23 $3 $3 :3 83 E3 83 8.3 83 $3 83- .28

38; n. E I A 396 _ESL E _ E _ z _ 4 >66 .38; E _ E _ z _ 4 3.22.9

moomtom €000 2350.52 550 82:5 nan Awbcm
8o 2.3m uszczoaxm 335...; Eon—Saun—

 

 

Auuznzacov .352 _ «am .88 3:53— amuh EEO “Ea 8355mm— houoﬁuhum "Tm Saab

111

3.3.1.2. Stage II Coefﬁcients

Results of the Stage 11 model, presented below in Table 3-2, indicate that the explanatory
variables account for 46 to 64 percent of the observed variation in the budget share
devoted to staples and for 14 to 61 percent in that of meat and ﬁsh. Furthermore, the
estimated results show that with the exception of vegetables and oil, the estimated income
coefﬁcients are statistically signiﬁcant in at least one season for each of the commodity
aggregates. This means that Bamako households’ consumption of staples, meat and ﬁsh,
oil, and other foods is responsive to changes in their real incomes. Also, a great number
of the price variables in all seasons are statistically signiﬁcant at the 10% level. For
instance, 4 of the 7 price variables were statistically signiﬁcant in the staples’ equation
for the pooled data.

The marginal budget shares, B, for staples are —O.172, -O.212, -O.185, and —O.252
for the lean, harvest, post-harvest, and planting seasons and —O. 195 for the pooled data
and are all statistically signiﬁcant at the 1 percent level. These results suggest that staples
are necessities; thus, Bamako households’ expenditures on staples take a smaller
percentage of income as households get richer. The income coefficients for meat and ﬁsh
range between 0.048 during the planting season and 0.169 during the harvest season. The
estimates of B for meat and ﬁsh are all positive and statistically signiﬁcant at the 1
percent level in all seasons except May and for the pooled data. This means that meat and
ﬁsh commodities are luxuries for Bamako households, since at higher income levels,
households increase the proportion of the food budget allocated to these goods.

With the exception of the price of other foods, the null hypothesis of stability of

the income and price parameters for the staple equation could not be rejected at the 10 %

112

signiﬁcance level. These results suggest that the impact of real income and relative
prices, except for the price of other foods, on the budget share for staples is stable across
seasons. In contrast, the Chow tests results show that there was statistically signiﬁcant
structural change, at least at the 10 % signiﬁcance level, in all the income and price
coefﬁcients, except for the price of meat and ﬁsh, in the meat and ﬁsh equation. The null
hypothesis of stability of the price parameters was also rejected, at least at the 10 %
signiﬁcance level, for the following demand equations: vegetables, sugar, and other

foods.

ll3

decorum“: a: 9:25 2: was: 3.0302 803 €22:an m: 85.0. conga»

moqum 2: .8 032:2“ echo Banana 0: Pa 22: 3%: E 0.8 «m8 326 2.: .58 323?; :26. 28th o. 05 8 232.:ch b.8238.“ 2a
3:23:08 anEzmo 2t :2: 38%.: 83a?“ Eon ”885228 E 80:82 2a mo:_a>-._. 358223 2.6m u m< .oEo: Eot 33m 88.8 n E can £50..
850 .3 8:8 u on .Ewam 80 3:8 n 3 £0.80 3:; n E .mo_aﬁowo> 80 02.5 n ma .5: can 32: 80 8E n ma £293 80 SE n _a ace—353 :36
ha 2:88:35 woe .8. _89 u Ex ”wanna—q u .32 u m can «332.368 u Sago... u E .3032. N 838032 N I .538 :8. n ~3w=< n A ”202

 

 

 

 

 

 

 

.3 9:25 9.3%..»an 333...; “56:08:—
Ecez Emmwum .88 £3QO «8.! 326 can 83535 $88.25— "NA 935,—.

 

B. E... 5.: 8% 8% 8B 83 mm... 5: Rs 83-8
888 888 3.8 5.8 «2.8 $8 388 85 98 9.8
$2. nod 83 $3. an; 83.. Rod- 83 238 ”8.8 m<
38 .8 A88 $88 Go. 8 888 §8 88. .8 $88 :88 an. 8 888 E . 8
SN.” 83 83. 83 89¢ 83. S: «8.8 88.8 08.? ~88 mad- 8
S88 3.8 $98 888 A88 388 $8 5.8 5.8 38: A38 :9:
8:2 89c. 83 ”8.8 :58. 88.8 83 Sad. 83 mod 83. :3- 8
:88 pm: 888 5.: $8 2.8 $8 5.8 and 38.8 A38 :38
23 8a.? 83. _86 29¢. .86 8“; ~88 on; :3 ”:3. 89¢. a
«88 608 8&8 E .8 398 3.8 888 5.8 888 E8 5.8 888
£2 .85. ~88 53. 89¢ 83. $3 89o- 89c 83 08.8 88.0 E
«S 8 8m. 8 :88 888 58 5:8 8% .8 5.8 $8 888 2 08 8a.:
82 88c 33 25.8 23. ”Nod 82 N88 83- Rod- _ :3 08.8 2
2 a 8 3.8 5.: :88 3.. : 3N. : «8 8 8w _8 388 so. _8 23.8 Gm. 8
w _ E .88 8.3. 888 N88 83 82 a _ ¢.¢- m _ 9o- 08.? wood- Rod- 8
$8 838 888 3a.: 3.08 an: $88 $8.8 :98 am: 888 288
2a.. :3- 298 $38- wood- 83. 83.. 83 838 E; 838 83 a
888 5.8 $28 5.8 5.8 5.8 38 82:8 538 888 A38 $.88
82 2 3 89¢ .88 .o $3 .8 .o E: 33. N28 £3- 82. 8 _ .8 8x
5.8 $88 83.8 888 A88 88m: :58 38.8 5.8 $88
82. wood. 3.2- 22. Ed- :8. a: 8:8 888 F3 .28
326 :28: a 8 E _ m _ ._ >66 38.1 a _ E 8 z _ ._ _
ﬁr. 28 “~22 838m 932...;

114

62838. a: 363 2: maﬁa 3:382 803 3208qu m: 35m cones—do

moatom 05 no.“ 032.86 Echo 235% o: 0.8 82:. 3%: 5 2a 3.8 30:0 05 .58 83.3-8 :05. «522. 2 05 “a 585ch 3.33:8,“ 98
356508 388:8 2t 35 3885 8383 Ron ”385588 E corona. 03 8:33.. ”228,33 :35 u m< .oEo; 80¢ 33m woo.“ u E 98 £58
350 80 out: n on .396 80 out: n ma .=o go out: n v: .mozﬁowg mo 028 n ma .5: can «8:. 80 out: u N: £293 80 8F. u E ace—8,33 :23
Ba 236530 woo.“ _mo. .82 n EX award—«E u 882 n m can 5052-32. n .0350"— u mm 4329. u 803E302 n I £088 :8. u Hm=w3< n A ”302

 

. ebb mHD 3.: cm: B5 mbb MEV bio M— .o a: .o 83.1
$8 $8 $8 988 58 $3 :88 $8 :88 :38
add mddd dddd mddd add. :3. god. add. :3. mad- m<
5.8 : _8 828 3.8 :88 3.: d8 .8 G08 88 :88 :98 $8
Sod dddd Sod- dddd- Nddd- dddd add wood 53 dddd _ddd mddd- a
888 §8 :_.8 odd 808 8:8 888 :08 808 A88 :88 :38
add wood .88 dddd mood- Edd SN: Sod- mddd- wood. wood. _ _dd- 8
ad .8 :N. 8 : _ .8 A38 :8. 8 :8. 8 as .8 :N. 8 am. 8 $88 5.8 :38
Sod mddd _ddd- dad- dddd Edd- «dd: dad add dddd- Edd mddd a
S 8 388 608 a. . 8 5.8 888 :88 an 8 A _ d8 :88 A38 8. .8
$3 Edd Vddd- dad dsd Edd- dd: dddd Edd mddd dddd .88 dd
38 Add: :88 :88 888 8:8 888 :88 an: :88 828 :88
dd _ ._ dddd Edd mddd dddd add. 2 5. god dmdd- dmdd dmdd- d :3 3
am 8 388 308 :2 .8 388 3.8 an .8 SN. 8 :88 :88 :88 :28
83 add. god. 398 dddd mddd dad dddd wood wood. Edd- dmdd 8.
an .8 £88 888 E .8 E8 888 8d .8 E8 :28 :88 : 08 :88
Ed dmdd- dddd mddd dmdd- Edd 83 ded- dddd- mmdd- : :3 dad- a
as 8 :88 $8 :38 A28 308 :88 £88 :58 :38 38 § .8
mddd Edd dad- dad «mod dddd 32 add add Edd- Edd Edd- 8x
:68 Add. 8 $88 : m. 8 ad. 8 £38 688 8.18 :88 £88
Jumbo? dmdd Sod dadd- add §d dmdd add :3 dad use
326 38: F r E _ E _ d 326 .38: 8 8E _ m _ 4
=0 333 o> 83229
be 9.25 9.33:0me 333...; Eon—.095

 

 

 

 

 

 

 

 

€253.83 .232 : owSm be 3:53— «moh .325 can moan—Edam ~80—52am "WM 935.

115

:08850. a: wEE... 0:: w:_:: 80.0300. 0.03 82052:: :8 00...: 5888.0 :00_>.0m 05 .8 03282: 88.0

8.88:8: o: 0.: 0.0:... 8::: E 0:: ::0: 30:0 05 .8 329-:— m_0>0_ 800.0: 2 0:: 8 ::moc_:w_: b.8885: 0.: 35.0.8808 80.2.88 0:. 85 0888:.
82.8.-. 2:: 80:05:08: :_ 80:80. 0.: :0=_:>-._. 3:03:30 ::.8: u m< .0Eo... E28 :85: 808 n n: 8:: .258 .050 8: 00:: n 8: 5&3: 8o 00:: n m:
.2: .80 00:: u v: .m0_:80w0> 80 00:: n m: .52.: 8:: 80:: 8o 00:: u N: £038: 8o 08.... u E 80321.0 2::: .0: 0.3.8598 800.: .80. .82 u :5: ”082

 

 

 

 

 

 

 

::.: :0: ::.: ::: 2.: 8.: 8.: 8.: :2 2.: 2.3.
.88 3:8 3:8 3:8 ::.8 3:8 8:8 : .8 ::.8 3:8

8:: :8: 8:8 :8: 8:: :8: .::.8 8:: 8::. .::.8 :::.8 :8:. .8: :.:.8 :::.8 8:
::.8 8:8 ::.8 ::.8 3:.8 3:8 ::.8 .::.8 3:8 ::.8 3:8 ::.8

8:: :8: :8: 8:: .8: :8. 2::. ::.8 :.:.8 :::.8 2:8 8:. :::.8 ::.8 8:8 :::.8 8:: ::
3:8 3:8 ::.8 ::.8 3:8 3:8 :88 ::.8 3:8 ::.8 ::.8 3:8

:88 :8:. :88 .88 :8:. :8: :8: :8: 3:: :8: :8: 8:: 8:8 .::.8 2:: 2::- 2:: ::
::.8 ::.8 3:8 ::.8 3:8 3:8 :88 3:8 3:8 .88 ::.8 ::.8

8:: ::c: :8:. 8:: :::.8 ::..: 8:8 .::.8 2:: 2:8 2:: 8:: 8:: :::.8 8::. 8:: :8: 8
::.8 ::.8 :..8 8.: ::.8 3:8 «::.8 ::.8 :..8 3:8 ::.8 ::.8

8::- 8:: 8:: :::.8 :8: 2:. 8:: 8::. :8: 8::. :8: 8:: 8:: .::.8 8::- 8:: ::.8 ::
::.8 ::.8 3:8 ::.8 3:8 3:8 ::.8 ::.8 ::.8 3:... .88 ::.8

8::- 8::- :8: :8:- 8:: ::2 :88 :::.8 8::. 8::. ::.8 8:. 8:: :8: :8:. :8: 8:: m:
::.8 3.8 3:8 3:8 3:8 3:8 ::.8 .::_8 3:... ::.8 .88 :..8

.8: :8: 8:: 3:: 8:: ::... :.:.8 8:: :8:- .::.8 :8:. :8... :8:. 8::. .8: 2::- 5:: ::
::.8 .::,8 ::.8 ::.8 2:8 ::.8 ::.8 .28 .88 3:8 3:8 ::.8

8:: :8:- :8:- :8: :8: :.... 8::- 8:: :8: 8::. ::.8 :2... :::.8 :8: :8: 8::- 8::- ...
::.8 3:8 .::8 3.8 ::.8 3:8 8:8 .88 ::.8 ::.8 ::.8 ::.8

:8: :8: 8:: 8:: :8: :8. :88 :::_8 8::. .::_8 8:8 :.:.: 8::. :8:. :::.8 :8:. 2::. :8
3:8 3:8 3:8 ::.8 3:... ::.8 3:8 .::.8 3:8 3:8

8:: 8:: :8: 8:: 8:: ::..: :2: ::..: :::.: ::..: ::..: 8:: :8: ::..: :8: .28
:28..— : _ z: _ I _ 4 39.0 ::.8:_ : — E F m _ .. 32.: 8.8,. .. 8 z: _ .. _ :

0E2. ES... :83: too... :88“. .050 28.6 333:...»

a: 0..qu 8::—2.03m “03......5 3085139

 

 

30:32.88 .083). = 0wmum :8 ::..:3— ::0H 32.0 8:: 33:58— ..0:0E:..:m “NJ. 03:8.

116

3.3.1.3. Stage III Coefﬁcients
Table 3-3, below, presents the estimated parameters for the Stage 111 model. Price,
income, and household size variables can explain about 10 to 20 percent of the observed
variation in the share of rice, millet-sorghum, maize, and wheat. The relatively low R2
may be due to the fact that households’ demand for speciﬁc staples, such as rice, is a
function of their tastes and preferences for these goods, which are usually controlled for
in the estimation process through the inclusion of many socio-demographic variables.

The estimates of B for rice are —0.085, -0.028, -0.076, and —0.009 for the lean,
harvest, post-harvest, and planting seasons and —O.109 for the pooled data and are
statistically signiﬁcant at the 10 percent level only for the pooled data. The income
coefficients for millet-sorghum range between —0.009 during the harvest season and —
0.055 during the planting season, and none of them are statistically signiﬁcant. The
marginal budget shares for maize range between —0.027 for the pooled data and —0.185
for the post-harvest season and are statistically signiﬁcant, at the 10 percent signiﬁcance
level, only for the pooled data. The results indicate that rice, millet-sorghum, and maize
are necessities for Bamako households since at higher income levels, households will
reduce the proportion of the food budget allocated to these goods. The estimated income
coefﬁcients for wheat are all positive, ranging between 0.002 during the planting season
and 0.110 during the lean season, and are statistically signiﬁcant during the lean, post-
harvest, and for the pooled data. This means that wheat is a luxury for Bamako
households.

The Chow test results indicate that there was statistically signiﬁcant structural

change, at l % signiﬁcance level, across seasons in all the estimated income and price

117

coefﬁcients for the maize equation. The Chow tests results also showed that there was
statistically signiﬁcant structural change, at least at the 10 % signiﬁcance level, in one or

more of the estimated price parameters for all demand equations in the Stage 111 model.

118

£26.50. u: wanna 05

9.7.: 3.0280. 0.03 0.0.0883 m: 00....»- mcoumaco ﬂops-H0383. 0.: 8.. 03.2.02. 0.2.0 20.05:... 2. 0.8 0.0.:- .8=S_ E 0... .00. 32.0 05 5.. 0020?:
“_0>0_ 2.09.0: o. 05 ... .53....nwﬁ 2.0030380 03 0.2.0.0500“. 38830-0 05 .05 0.8%... 0020?. Eon $000558: E v2.2.0. 0... 0020?... 3:032:00
:3... u m< ...... .303... :5. 30... ..o out: n w: 400:3 .... out: n v: .0808 ..o 00.... u m: €398.05... mo 00:: u N: .00.. ..o 00.... u 3 4:29:60 :3...
.0: 0.3.2.098 038.... B0. .88 u EX 85.53 u .82 u m :5 $022.32. n .0350... n I: 40022 n 898032 u I .5800 :02 n .maw:< n 1. .082

 

 

 

 

 

 

\

...... 2... x... .N... 2... cm... mm... mm... R... .N... 2......

.88 3.8 9.88 5.8 3.8 8...: .3: ..w... 3.8 80.8

.8... 8.... ......o 3.....- Noo... 2...... 8.... 2...... . ...... . ...... m...
5 8 .38 .38 an... .38 3.8 «88 5.... .88 .38 8.8 .....8
$0.. 8.....- 0.......- 2.... 2.....- 8.... man. 8.....- VS..- 3.... 8.....- 8.... m...
888 .3.: .mm... .88 .38 .8... «88 .88 .88 .8... 5.: .28
880 0...... 08... .8... .8... .2....- E..~ 8.....- 8.....- 8.... 3.....- 2...... ....
.88 .ma... .....8 .38 .28 .38 $8 $8 an... .88 as... .....8
«...... .2.... 2.....- N2... 2...... 8.... So. So... 8:... 3...... 9...... ...... ....
:88 :2. 3... .2.: 8...: 2...... 3.8 .88 .3... 6.8 .3.: 6.8
2....” VS... a _ ... 8.... 8.... 8......- 22 mm. ...- m... ...- . .2.- . . ...- 2......- N...
.3 .8 .88 .3.: .38 .3... G. .8 3 .8 .3. .. .98 ......N. .... : ... . .8
E.” R. ...- 3.....- . .0..- .....- 8.....- ...... 8...... E... .2... 8. ... .- ......- ...
a. 8 .3.... ......8 E8 .N . .8 .38 3 .8 .38 .. .8 .8. .. .28 .... . _.
$0.. 2.....- 3.....- 3.....- 2......- 2.....- ...... 2.....- 2......- S.....- .8..- mwo..- ..x

...m... ...... .88 .28 ...... .38 am... .28 .3: .28

8.... 2m... .3...- .8... E... ....m... .8... ....m... 3... .2.... .28
32.0 :28. _ n. _ .... _ m _ -. 32.0 30.8.: n. H .... _ z _ -.

EEwSm-BEE 002 0.033-:3
.2. 0.3.5 0.3%..0num ”033...; 2055:3—

 

 

3—52 :— 0»:.@ .8.— mzsmom «mo-_- EEU :5. 0.03830”.— ..0uoﬁauam .m-m 03:.

119

-:0..0...m0. a: ”3.0.00 0...
9.8: 00.0500. 0.0.5 0.0.0E0.0.. m... 00:.m 0:0..0000 Eons-.....ﬂoom 0... .0. 0...0..0>0 0.0..0 0.0.2.0.. 0: 0.0 0.0....- .m0._0.. :. 0.0 .00. 30:0 0... .0. 0020?:
:05. .:00.0: o. 0... .0 .:00....:w.m 3.00.8..80 0.0 35.05000 00.00....00 0... .0... 0.00.0:. 0020?. 0.0.. 80355.0: :. 08.000. 0.0 m0:.0>-.- ..:0.0>.=..0
.300 n m< ::0 .80.... 0:0 0.00. ..0 00.... M Q. ..00...» ..o 00.... u .... 60.0.: ..o 00.... u m... .E....w.00-.0:.:. ..o 00.... u m: .00.. .0 00.... n ... ..:0_0>.=..0 ....00
.0: 0.3.0588 0.00.0 .00. .80. n ASA .m:..:0... n >02 u n. 0:0 $030.38.. n E0230". u m... ..m0>.0_.. n .0..E0>oz u .. £0.08 :00_ u .m=w=< u -. ”0.02

 

 

 

 

 

...... 8... 8... N... 2.. 0m... 8... ...... ...... 8... .83.
.88 .88 .088 .88 ...N... .88 .3.-8 .2.: .8... .8...

8......- 8....- 8.....- 8.... 8.....- ..8.. 2.....- .Vo....- 8.....- 8.....- .8..- 8.....- 8.....- 8.....- 8....- 8..
. 5.8 ....8 .8... .88 .88 .8... .88 .88 .88 .88 .8... .8...

2.... ...... 8....- ..8... 8.... .8... 8...... .8... 8....- 8.... ...... m8... 8.....- 808- 3.....- .8... m8..- 8
- .88 .88 .88 .88 .88 .8... .88 ...... .8... .8... .8... .88

8.... .8... 8.....- 8.... ...... 8.... v8...- .8..- 8.....- .8..- 8....- 80.0 8.... .8..- 2....- ..8... 8.... ....
.88 .88 .8... .8... .8... .88 888 ...... .88 .28 .8... ..08

8.....- 8....-:....- 08... 8....- 8... 8.... .8..- 2.....- 88... 8.... N8... 8...- 808- 8...- 8....- 8...- 8
.88 .1... .8... .88 .88 .8... 888 .8... .88 .88 .28 .88

...........8.... 2.... m8..- 8.... 08... 0...... 08... .8... .8... .8..- 0...-m .8... 8.....- 8... 8.... .8... 8
...-.8 .88 ...m... ...... .8... .88 .88 .88 .808 .88 .8... .8...

..8..- 88 3.... 80..- 8.... 80.. 8.....- N.....- 8...... 8....- 8.... 8.... 8.... 8.... 8....- 800 0...... ...
.88 .28 .88 .88 .8... ....8 .88 ....8 .88 .8... .88 .3....

8... ...... 2m... .8... 8;... ...... .8... 8.... 8.... 8.... ...... 88m .8..- .8..- 8....- 8.....- ......- .3.
.88 ....8 .88 .28 .8... .8... .8... .88 .88 .88

8.... .8... 8.... 8m... 8.... 8.....- 8... 8.....- ..8...- CIR...- 8... 8.... .8... .8..- 08... .88
8.8... .. .... . I . 4 32.0828..— . .:._ : . . 30.0.8.8... .. _ 2.. _ z . .

.00; 0N5: mo_nu_.a>

0.003-80.03.

 

 

 

.0 0.0.5 0.3.059... "030...; .:0..:0 0n.

2.0.3.250. .255. :— 0w0.w .0. 3.500% .00-F 30:0 ::0 00.05....m 3350.08 "m4“ 030..-

 

120

3.3.2. Income Elasticities

Table 34, below, reports the estimated income elasticities for the Stage I, II, and 111
models for each season and the pooled data and the Chow test results. The income
elasticity estimates, in absolute terms, are used to classify commodities into one of three
categories: inferior goods (11(i) <0), normal goods (0< n(i) <1), and luxury goods (11(i)
>1). In this section, the estimated income elasticities are examined in order to determine
if in any given season (1) households’ consumption patterns are responsive to changes in
their real incomes and (2) there are evidence of seasonal changes in income
responsiveness.

Table 34: Estimated Income Elasticities for Stage; II, and III Models
Commodities I L I H l PH l P I Pooled I Chow

 

 

 

Stage 1
Food 0626* 0.463" 0.577“ 0.574“ 0.516“ 6.576“
Durable Goods 1.496 2.277“ 1.361 1.468 1.912“ 5.202“
Semi-Durable Goods 0.865 1.125 1980* 0.563 1.600“ 1.568
Health 2.104“ 1.368 1.847" 2.364" 1.721 ‘ 3.33"
Energy and Utilities 1.386 0.634 0.343" 1.094 0.829 1.587
Other Non-Durables 0.502 0116*" 0.267" 0.646 0.492" 1.598
Services 0.816 1.164 1.437 1.301 1.012
Stage II
Staples 0439" 0.386“ 0.467“ 0.274“ 0.418‘I 1.027
Meat and Fish 1810* 2.064“ 2.087" 1.359 1.775" 2.202"
Vegetables 0.968 1.150 0.916 1.208 1.006 1.275
Oil 1.208 1.819" 1.270 0.639 1.364" 0.902
Sugar 0.751 0.631 0.966 0400*" 0.874 0.712
Other Foods 0.863 0.329“ 0.850 0.560" 0.687"' 1.182
Food Away From Home 1.422 1.550 1.232 2.349 1.606
Stage III

Rice 0.837 0.945 0.862 0.983 0.796“ 1.618
Millet-Sorghum 0.772 0.957 0.742 0.817 0.841 1.682
Maize 0.286 0.617 -0.521 0.199 0.793” 5223‘
Wheat 2.626" 2.028 3.068“ 1.039 1.944" 0.618
Roots&'1‘ubers 4.547 1.868 3.920 4.123 3 .292

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest and P = May =
planting; ‘, ", and "“" denotes estimated income parameters are signiﬁcant at the 1%, 5% and 10% level,
respectively. Tests of statistical signiﬁcance are not available for the dropped equations because their
parameters were recovered using the adding-up restrictions.

121

3.3.2.1. Stage I Income Elasticities: Food vs. Non-Food Commodities

For most commodities, the estimated income elasticity of demand is statistically
signiﬁcant at the 1, 5, or 10 percent level, suggesting that Bamako households’
consumption patterns in any given season are responsive to changes in their real incomes.
For instance, the pooled data estimated income elasticities are statistically signiﬁcant, at
the 1 percent level, for all commodities, except energy and utilities. The estimated
income elasticities are also statistically signiﬁcant, at least at the 10 percent level, for
food and health during the lean season, for food, health, and other non-durable goods
during the harvest season, for all commodities except durables goods during the post-
harvest seasons, and for food and health during the planting season.

All the estimated income elasticities are positive, suggesting that the food and
non-food commodities are normal goods, as their consumption will increase with income.
The results indicate that food and other non-durable commodity groups are clearly
necessities across all seasons (0< n(i) <1), indicating that Bamako households will tend
to spend proportionally less on these commodities at higher income levels. The income
elasticities of demand for durable goods (housewares and education) and health were
found to be greater than 1 in all seasons, suggesting that these commodities are luxury
products for Bamako households. These ﬁnding suggest that as households’ income
increase, they tend to spend proportionally more on these non-food commodities.
Moreover, the results show that semi-durable goods (clothing and footwear) are
necessities in August and May and luxuries during the other seasons.

The results also indicate that there is considerable variation in the estimated income

elasticities across seasons and across commodity groups in any given season. Between

122

the lean and harvest season, which corresponds to the period when households’ real
expenditures decreased by 36 percent, the income elasticity for food decreases from
0.626 to 0.463 while those of most non-food commodities increase, especially durable
(from 1.496 to 2.277) and semi-durable goods (from 0.865 to 1.125). In contrast,
between the harvest and post harvest seasons, which corresponds to the period when
households’ real expenditures increased by 4 percent, the income elasticity of demand for
food increases from 0.463 to 0.577, while that for durable goods, energy and utilities, and
services decreased. Between the post-harvest and planting seasons, which corresponds to
the period when Bamako households’ real expenditures fell by 18 percent, the income
elasticity of food decreases from 0.577 to 0.574, while that for durables goods (1.361 to
1.468), health (1.847 to 2.364), energy and utilities (0.343 to 1.094), and services (0.267
to 0.646) increases.

Two main conclusions can be drawn from these ﬁndings. First, the high absolute
level of these income elasticities, even for food, underscores the extreme level of poverty
and unmet “basic needs” that prevail in Bamako. In consequence, the results suggest that
policies that aim at increasing households’ real income will cause substantial
improvements not only in the quantity of food available in urban households but also in
the demand for non-food commodities. Rapid growth in the demand for non-food
commodities could translate into sizable rise in employment, to the extent that these
commodities can be produced domestically.

Second, the results indicate that the responsiveness of food consumption to changes
in real income remains fairly stable across seasons compared to that of non-food

commodities. For instance the food income elasticity ranges between 0.463 and 0.626

123

while the elasticity for durable goods ranges between 1.361 and 2.277. This is evidence
that households engage in food consumption smoothing. Speciﬁcally, the results indicate
that Bamako households engage in food consumption smoothing from seasonal shocks in
real incomes at the expense of non-food commodities such as health and durable goods
(e. g. housewares and education).
3.3.2.2. Stage 11 Income Elasticities: Staples vs. Non-Staples Commodities
The estimated income elasticties for the Stage II commodity groups are statistically
signiﬁcant at least at the 10 percent level for the following commodities: staples, meat
and ﬁsh, oil, and other foods for the pooled data and the harvest season; staples and meat
and ﬁsh for the lean and post-harvest seasons; and staples, sugar, and other foods during
the planting season. All the estimated income elasticties of demand in the Stage II
commodity groups are positive, suggesting that these commodities are normal or luxury
goods. The results show that staples, sugar, and other foods have income elasticities that
are less than 1 in all seasons, indicating that a 1 percent change in income results in a less
than proportionate increase in expenditures on these goods. In contrast, the income
elasticity for meat and ﬁsh and food away from home exceeds 1 in all seasons, suggesting
that the amounts households spend on these commodities increase more than
proportionally with income. The income elasticity of demand for vegetables is close to
unity in all seasons. The fact that there are no inferior goods in the commodity aggregates
considered is consistent with the ﬁndings of previous consumption studies (Rogers and
Lowdermilk, 1991 and Reardon et al., 1999).

The estimated income elasticity for staples derived from the pooled data

estimation is slightly lower than that found by Reardon et a1 (1999), which was 0.47 aﬁer

124

the CFA Franc devaluation. These ﬁndings suggest that staples have been a necessity
over time and that Bamako households’ consumption of staples is becoming slightly less
responsive to changes in their real incomes. The income elasticities derived from the
pooled data estimation also indicate that staples have the smallest income elasticity of
demand (0.418) while that of vegetables (1.006), food away from home (1.606), and meat
and ﬁsh (1.775) are considerably larger. Reardon et al (1999) found that the income
elasticity for staples (0.47) was greater than that for vegetables (0.180) and meat and ﬁsh
(0.160) after the CF A Franc devaluation. These ﬁndings suggest that Bamako
households’ consumption of vegetables and meat and ﬁsh is becoming increasingly
responsive to changes in their real incomes.

Furthermore, the null hypothesis of stability in the income parameters could only
be rejected for meat and ﬁsh, suggesting that the impact of real income on the demand for
meat and ﬁsh is not constant across seasons. In contrast, the results indicate that the effect
of changes in real income on the demand for staples, vegetables, oil, sugar, and other
foods is not statistically signiﬁcantly different across seasons.

Two main conclusions can be drawn from these ﬁndings. First, the high absolute
level of the income elasticities of demand for vegetables and meat and ﬁsh, which
contain essential micronutrients, suggests that improvements in households’ real incomes
will have a substantial impact on the quality of their diets. Hence, many nutritional
deﬁciencies, such as vitamin A deﬁciency, could be addressed by policies that focus on
stimulating income growth.

Second, the fact that the income elasticity of staples varies less across seasons

than that of non-staple commodities suggests that households engage in food

125

consumption smoothing by protecting their consumption of staples at the expense of non-
staple foods, which contain essential micro-nutrients. The results of the previous essay
indicated that Bamako households will tend to diversify their diets during periods of low
grain prices (harvest and post-harvest seasons) and revert to necessities as food prices
begin to increase and real income levels decline.

3.3.2.3. Stage III Income Elasticities: Rice vs. Other Staples Commodities

Only a few of the estimated income elasticties of the Stage 111 model are statistically
signiﬁcant at least at the 10 percent level: rice, maize, and wheat for the pooled data, and
wheat during the lean and harvest seasons. These results indicate that, with the exception
of wheat, Bamako households’ consumption of rice, millet-sorghum, and maize is not
responsive to changes in their real incomes in any given season. However, the pooled
data results are statistically signiﬁcant at the 10 percent level for all staples, except
millet-sorghum, suggesting that Bamako households’ consumption of staples is
responsive to changes in their real incomes in the long-run. The pooled data results also
show that the estimated income elasticity of rice (0.796) is greater than that found by
Rogers and Lowdermilk (1991) of 0.562 and Reardon et al (1999) of 0.23. These results
suggest that rice is becoming less of a necessity for urban households over time.

3.3.3. Own and Cross Price Elasticities

The own-price elasticities of demand are used to test the law of demand hypothesis,
which says that normal goods must have downward sloping demand curves. The cross-
price elasticities are used to classify commodities into one of three categories: substitutes
(5i; > 0), complements (£9- < 0), and independent (£3 = 0). In this section, the estimated

own-price and cross price elasticities are examined in order to determine if in any given

126

season (1) Bamako households’ consumption patterns are responsive to changes in
relative prices; (2) there is evidence of seasonal change in price responsiveness; and (3)
households substitute among and between broad commodity groups.

3.3.3.1. Stage I Price Elasticities

3.3.3.1.]. Uncompensated and Compensated Own-Price Elasticities

Table 3-5, below, shows the compensated and uncompensated own-price elasticity of
demand for commodities in the Stage 1 model by season and for the pooled data. The
own-price elasticities, both compensated and uncompensated, are all negative, suggesting
that there is an inverse relationship between price and quantity demanded for each of the
commodity groups of the Stage I model.

Table 3-5: Compensated and Uncompensated Own-Price Elasticities for Stage I
Commodities by Season

 

 

 

Commo- Marshallian (Uncompensated) Hicksian (Compensated) I Chow
dities L I H I PH I P I Pooled L I H I PH I P I Pooled I Test

Food -0.698 -0.545 -0.574 -0.643 -0.612 -0.480 -0.353 -0.320 -0.364 -0.392 7 3.742

DG -0.684 4.132 -0.768 -0.958 -0.890 -0.465 -0.7 l 4 -0.647 -0.8 16 -0.654 1.720

SDG -0.332 -0.7 l 7 -0.945 - l .003 -0.676 -0.180 -0.604 41.683 -0.964 —0.493 2.430

Health -0.638 -0.722 - l .057 -0.895 -0.692 -0.466 -0.61 8 -0. 889 -0.588 -0.532 0.9 l 3

EU -0.578 -0.380 -0.383 -0.31 I -0.384 -0.461 -0.328 -0.346 -0.207 -0.307 I .552

ONE -l.643 -0.l49 -l.018 -1.305 -l.093 4.624 -0.l44 -l.008 -l.274 -l.073 2.028

Services -0.211 41.134 -0.049 41.650 -0. 162 -0.107 ~0.017 -0. 100 -0.553 ~0.060

 

 

 

Note: DG = Durable Goods; SDG = Semi-Durable Goods; EU = Energy and Utilities;
and OND = Other Non-Durable Goods. Bold values denote that the estimated price
elasticities are statistically signiﬁcant at the 10% level. Tests of statistical signiﬁcance
can’t be performed for Services because their parameters were recovered using the
adding-up restrictions.

For most commodities, the estimated own-price elasticity of demand is

statistically signiﬁcant at the 10 percent level, indicating that households’ demand for

these commodities is responsive to own-price changes in all seasons and for the pooled

data. For instance, the pooled data own-price elasticities are statistically signiﬁcant for

all commodities, except other non-durable goods. All the estimated own-price elasticities

127

are statistically signiﬁcant during the lean season. The own-price elasticities are also
statistically signiﬁcant for food, health, energy and utilities, and other non-durable goods
during the harvest season, food, durable goods, and energy and utilities during the post-
harvest season, and food, health, and energy and utilities during the planting season.

All the statistically signiﬁcant compensated own-price elasticities are less than 1
(in absolute value) in all seasons, suggesting that the demand for these commodities is
inelastic. The statistically signiﬁcant uncompensated own-price elasticities derived from
the pooled data range between —0.384 for energy and utilities and -0.890 for durable
goods, suggesting that the demand for energy and utilities is the least responsive to own-
price changes.

The compensated own-price elasticities are much smaller in magnitude than the
uncompensated ones, suggesting that prices occupy a smaller role when the income
effects are removed. The results show that income effects from changes in the price of
food are very strong, implying that the price of food has a substantial impact on the real
income of Bamako households. For instance, the pooled data results show that
households would react to a 1 percent increase in the price of food by reducing food
consumption by 0.397 percent, when their purchasing power is held constant. However,
with no compensation for the price increase, the price change decreases the quantity
demanded of food by 0.615 percent. These results show that changes in food prices
substantially affect households’ purchasing power because of the large proportion of
income that is devoted to food. Furthermore, a comparison of the uncompensated and

compensated own-price elasticities across seasons reveals that changes in the price of

128

food have the greatest impact on Bamako households’ real incomes during the planting
season and the smallest effect during the harvest season.

3.3.3.1.2. Uncompensated and Compensated Cross-Price Elasticities

Tables 3-6, below, presents the compensated and uncompensated cross-price elasticities
for non-food commodities by season and for the pooled data. The results indicate that the
demand for food is not very responsive to changes in the price of non-food commodities
in any given season and for the pooled data. For instance, the results show that the price
of health during the lean season and for the pooled data, energy and utilities in all periods
except the lean season, other non-durable goods during the harvest and post-harvest
seasons, and services for the pooled data have a statistically signiﬁcant, but very small,
impact on the demand for food.

In contrast, the results indicate that the price of food has strong and statistically
signiﬁcant uncompensated effects on the demand for non-food commodities. For
instance, the uncompensated results show that a 1 percentage increase in the price of food
reduces urban households’ expenditures on traditional and formal health services by
0.451 percent for the pooled data and 0.848 percent during the planting season. A
comparison of the uncompensated against the compensated cross-price elasticity of
demand for health with respect to the price of food indicates that the income effects from
changes in the price of food are stronger than the pure substitution effect, in that changes
in the price of food will substantially increase or decrease households’ real incomes.

The results, in Table 3-6, also show that the price of food has a negative
statistically signiﬁcant uncompensated effect on the demand for energy and utilities

during the harvest (-0.299) and post-harvest (-0.099) seasons and for the pooled data (-

129

0.212). However, once households are compensated for changes in the price of food, the
results show that the compensated cross-price elasticity of demand for energy and utilities
with respect to the price of food is very small (-0.035 during the harvest and 0.051 during
the post-harvest seasons). Bamako households’ consumption of energy and utilities tends
to increase during the harvest season, which also corresponds to the winter months, and

post-harvest season, which coincides with hot dry-season winds (the Harmattan).

130

Table 3-6: Stage I Compensated and Uncompensated Price Elasticities

 

 

 

 

 

 

Commo- L Marshallian (Uncompensated) I Hicksian (Compensated) I Chow
dities 1.. H PH P Pooled 1.. H PH P Pooled Test
Food
Price of food -0.698 -0.545 -0.574 -0.643 -0.612 -0.480 -0.353 -0.320 -0.364 -0.392 3.742
Price ofDG 0.006 0.106 0.015 0.045 0.067 0.098 0.191 0.066 0.101 0.132 3.590
Price ofSDG 0.028 0.061 0.052 0.002 0.061 0.138 0.107 0.128 0.042 0.120 3.387
Price ofhealth 0.037 0.016 0.032 0.006 0.012 0.088 0.051 0.084 0.080 0.061 4.458
Price ofEU 0.039 -0.045 -0.050 -0.007 -0.018 0.092 -0.007 0.013 0.048 0.030 5.503
Price ofOND 0.010 -0.072 -0.034 0.045 0.002 0.033 -0.053 -0.013 0.072 0.024 3.582
Price ofS -0.049 0.016 -0.017 -0.022 -0.027 0.031 0.063 0.043 0.021 0.024 5.268
DG
Price offood -0.289 -0.514 -0.273 -0.208 -0.383 0.232 0.434 0.325 0.506 0.433 4.572
Price ofDG -0.684 -1.132 -0.768 -0.958 -0.890 -0.465 -0.714 -0.647 -0.816 -0.654 1.720
Price ofSDG -0.136 ~0.186 0.106 -0.084 -0.165 0.127 0.042 0.287 0.017 0.061 2.062
Price ofhealth -0.235 -0.100 -0.085 -0.208 -0.185 -0.113 0.072 0.039 -0.017 0.000 1.610
Price ofEU -0.011 -0.136 -0.115 -0.121 -0.110 0.116 0.050 0.032 0.019 0.069 2.378
Price ofOND 0.040 -0.049 -0.117 0.035 -0.029 0.095 0.048 -0.067 0.105 0.049 2.630
Price ofS -0.182 -0.l60 -0.109 0.076 -0.149 0.008 0.068 0.031 0.186 0.043 3.132
SDG
Price offood -0.027 -0.024 -0.446 0.022 -0.258 0.274 0.444 0.424 0.296 0.426 2.365
Price ofDG -0.021 -0.129 0.016 -0.030 -0.141 0.106 0.078 0.192 0.024 0.063 1.210
Price ofSDG -0.332 -0.717 -0.945 -1.003 -0.676 -0.180 -0.604 -0.683 -0.964 -0.493 2.430
Price ofhealth -0.048 -0.097 -0.211 0.067 -0.168 0.023 -0.011 -0.031 0.141 -0.014 1.832
Price ofEU -0.014 -0.079 -0.033 0.038 -0.066 0.059 0.013 0.181 0.092 0.081 1.205
Price ofOND -0.054 0.011 0.096 0.129 -0.013 -0.022 0.060 0.169 0.156 0.054 1.388
Price ofS -0.370 -0.091 -0.457 0.213 -0.276 -0.260 0.022 -0.253 0.256 -0.118 2.323
Health
Price offood -0.357 -0.287 -0.406 -0.848 -0.451 0.376 0.282 0.405 0.300 0.270 1.490
Price ofDG -0.509 -0.077 -0.126 -0.242 -0.221 -0.201 0.174 0.038 -0.013 0.005 1.143
Price ofSDG -0.320 -0.152 -0.290 -0.088 -0.216 0.049 -0.015 -0.045 0.075 -0.007 1.017
Price ofhealth -0.638 -0.722 -1.057 -0.895 -0.692 -0.466 -0.618 -0.889 -0.588 -0.532 0.913
Price ofEU -0.259 -0.032 0.132 -0.114 -0.081 -0.081 0.080 0.331 0.111 0.078 3.315
Price ofOND 0.011 0.035 -0.001 -0.068 0.008 0.088 0.093 0.068 0.045 0.078 1.937
Price ofS -0.032 -0.134 -0.099 -0.107 -0.067 0.236 0.004 0.091 0.070 0.109 1.703
EU
Price offood -0.105 -0.299 -0.099 -0.290 -0.212 0.378 -0.035 0.051 0.242 0.138 1.978
Price ofDG -0.003 -0.004 -0.004 -0.087 -0.008 0.200 0.112 0.026 0.019 0.098 1.278
Price ofSDG -0.121 -0.047 0.177 -0.009 0.008 0.122 0.016 0.222 0.066 0.107 1.298
Price ofhealth -0.192 0.026 0.248 0.009 0.001 -0.079 0.074 0.279 0.151 0.080 1.767
Price of EU -0.578 -0.380 -0.383 -0.311 -0.384 -0.461 -0.328 -0.346 -0.207 -0.307 1.552
Price ofOND -0.115 0.333 -0.003 -0.121 0.024 -0.065 0.360 0.009 -0.068 0.058 2.087
Price ofS -0.272 -0.263 -0.277 -0.285 -0.258 -0.096 -0.199 -0.242 -0.203 -0.174 1.703

 

 

 

Note: DG = Durable Goods; SDG = Semi-Durable Goods; EU = Energy and Utilities; and OND = Other
Non-Durable Goods; S = Services. L = August = lean season, H = November = harvest, PH = February =
post-harvest and P = May = planting; Bold values denote that the estimated price elasticities are statistically
signiﬁcant at the 10% level. Tests of statistical signiﬁcance cannot be performed for Services because their
parameters were recovered using the adding-up restrictions.

131

 

 

 

 

Table 3-6: Stage I Compensated and Uncom ensated Price Elasticities (continqu

Commo- Marshallian (Uncompensated) Hicksian (Compensated) I Chow

dities L I H I PH I P TPooled L I H I PH I P IPooledI Test
OND

Price of food 0.141 -0.559 -0.270 0.417 0.036 0.316 -0.511 -0.153 0.731 0.246 1.780

Price ofDG 0.307 0.186 -0.183 0.151 0.089 0.381 0.207 -0.159 0.213 0.153 1.607

Price of SDG -0.l95 0.128 0.568 0.180 0.096 -0.106 0.140 0.604 0.225 0.152 2.408

Priceofhealth 0.156 0.156 0.142 0.038 0.130 0.197 0.165 0.167 0.121 0.177 0.907

Price ofEU -0.193 0.677 -0.001 -0.198 0.085 -0.150 0.686 0.027 -0.137 0.130 1.082

PriceofOND -1.643 -0.149 -1.018 -l.305 -1.093 -1.624 -0.144 -1.008 -1.274 -1.073 2.028

Price of 8 0.923 -0.554 0.496 0.072 0.165 0.987 -0.542 0.523 0.120 0.214 2.292
Services

Price of food -0.l99 -0.224 -0.448 -0.493 -0.334 0.085 0.260 0.183 0.138 0.094

Price ofDG -0.110 -0.089 -0.101 0.114 -0.072 0.009 0.125 0.027 0.241 0.059

Price ofSDG -0.501 -0.095 -0.516 0.145 -0.239 -0.358 0.022 -0.325 0.235 -0.118

Price ofhealth 0.085 -0.085 -0.050 -0.048 0.006 0.152 0.003 0.081 0.121 0.102

Price of EU -0.133 ~0.256 -0.408 -0.383 -0.258 -0.064 -0.l62 -0.253 -0.259 -0.164

Price ofOND 0.253 -0.281 0.135 0.014 0.046 0.283 -0.231 0.189 0.077 0.088

Price ofs -0.211 -0.134 -0.049 -0.650 -0.l62 -0.107 -0.017 -0.100 -0.553 -0.060

 

 

 

Note: DG = Durable Goods; SDG = Semi-Durable Goods; EU = Energy and Utilities; and OND = Other
Non-Durable Goods; S = Services. L = August = lean season, H = November = harvest, PH = February =
post-harvest and P = May = planting; Bold values denote that the estimated price elasticities are statistically
signiﬁcant at the 10% level. Tests of statistical signiﬁcance cannot be performed for Services because their
parameters were recovered using the adding-up restrictions.

3.3.3.2. Stage 11 Price Elasticities

3.3.3.2.]. Uncompensated and Compensated Own-Price Elasticities

Table 3-7, below, presents the compensated and uncompensated own-price elasticity of

demand for the Stage 11 model commodities by season and for the pooled data. First, the

results indicate that the estimated own-price elasticities, both compensated and

uncompensated, are negative and most of them are statistically signiﬁcant at the 10

percent level. For instance, the pooled data own-price elasticities are statistically

signiﬁcant for all staples, meat and fish, oil, and other foods. The estimated own-price

elasticities of demand for staples and other foods are statistically signiﬁcant during the

lean season; for staples, oil and other foods during the harvest season; for vegetables

during the post-harvest season; and for meat and ﬁsh and other foods during the planting

season.

132

Table 3-7: Compensated and Uncompensated Own-Price Elasticities for Stage II
Commodities by Season

 

 

 

 

 

Commo- Marshallian (Uncompensated) Hicksian (Compensated) I Chow
dities L H I PH I P I Pooled L I H I PH I P I Pooled I Test
Staples -0.500 -0.513 ~0.61 1 -0.749 -0.506 -0.366 -0.380 -0.449 -0.654 -0.364 1.080
MP -0.968 -0.967 -0.799 -0.754 -0.905 -0.697 -0.640 -0.479 -0.571 -0.641 1.718
Veg -0.870 -1.321 -0.591 -1.254 -0.958 -0.739 -1. 186 -0.477 -1.119 -0.835 3.113
011 -1.273 -0.695 -0.743 -1.128 -0.770 -1.220 -0.641 -0.698 -1.1 1 1 -0.724 1.665
Sugar -0.868 -0.898 -1.017 -1.267 -0.955 -0.822 -0.856 -0.950 -1.239 -0.897 0.723
OF -0.687 -0.585 -0. 85 1 -0.619 -0.632 -0. 584 -0.550 -0.761 -0.561 -0.558 2.032
FAFH -0.919 -0.756 -0.813 -0.895 -0.825 -0.658 -0.484 ~0.612 -0.411 -0.533

 

Note: MF = Meat and Fish; Veg = Vegetables; OF = Other Foods; and FAF H = Food
Away From Home. L = August = lean season, H = November = harvest, PH = February =
post-harvest and P = May = planting; Bold values denote that the estimated price
elasticities are statistically signiﬁcant at the 10% level. Tests of statistical signiﬁcance
cannot be performed for FAF H because their parameters were recovered using the
adding-up restrictions.

Second, the results show that all the estimated uncompensated and compensated
own-price elasticities are less than 1 (in absolute value), suggesting that the demand for
these commodities is inelastic. For each of these food commodities, a 1 percent change
in the commodity’s own price has a less than proportionate effect on the quantity
demanded of that commodity. The statistically signiﬁcant uncompensated own-price
elasticities derived from the pooled data indicate that staples have the smallest own-price
elasticity (-0.506) while meat and ﬁsh have the largest (-0.905).
3.3.3.2.2. Uncompensated and Compensated Cross-Price Elasticities
Table 3-8, below, presents the compensated and uncompensated cross-price elasticities
for the Stage 11 model by season and for the pooled data. First, the results show that
changes in the price of staples have strong and statistically signiﬁcant uncompensated
effects on the demand for meat and ﬁsh, vegetables, oil, and sugar. For instance, the
uncompensated results show that a 1 percentage increase in the price of staples reduces

urban households’ expenditures on meat and ﬁsh by 0.760 during the post-harvest season

and on vegetables by 0.254 percent during the lean season.

133

Moreover, a comparison of the uncompensated against the compensated cross-
price elasticity of demand for non-staple commodities with respect to the price of staples
indicates that changes in the price of staples will substantially increase or decrease
Bamako households’ real incomes. For instance, the results derived from the pooled
estimation indicate that households would react to an increase in the price of staples by
reducing the consumption of oil (-0.237) and increasing that of vegetables (0.085), when
their purchasing power is held constant (i.e. no income effects). However, when
households are not compensated for the price increase, an increase in the price of staples
substantially reduces households’ purchasing power because staples occupy a sizeable
proportion of households’ food budget. Hence, the total effect of a one percent increase
in the price of staples is to reduce the consumption of vegetables (-0.254 percent) and oil

(-0.946 percent).

134

Table 3-8: Uncompensated and Compensated Price Elasticities for Stage 11 Model

 

 

 

 

 

 

 

 

 

 

Marshallian (Uncompensated) Hicksian (Compensated) Chow

Commodities L I H I PH I P IPooled L H I PH I P IPooled Test
Staples
Price of St -0.500 -0.513 -0.611 -0.749 -0.506 -0.366 -0.380 -0.449 -0.654 -0.364 1.080
Price of MP 0.009 0.074 -0.087 0.059 0.035 0.075 0.135 -0.015 0.096 0.097 1.705
Price of Veg -0.041 0.103 0.000 0.054 -0.017 0.019 0.148 0.059 0.085 0.034 0.808
Price of oil 0.036 -0.039 0.027 0.028 -0.037 0.056 -0.027 0.043 0.036 -0.023 0.767
Price of S -0.026 -0.011 0.076 0.213 0.044 0.001 0.014 0.109 0.232 0.072 0.730
PriceofOF 0.021 0.044 0.124 0.093 0.056 0.074 0.084 0.173 0.121 0.101 2.383
Price of FAFH 0.061 -0.043 0.004 0.028 0.001 0.142 0.025 0.080 0.084 0.078 1.032
Meat and Fis
Price of St -0.402 -0.419 -0.760 -0.224 -0.384 0.153 0.294 -0.034 0.248 0.215 1.913
Price of MP -0.968 -0.967 -0.799 -0.754 -0.905 -0.697 -0.640 -0.479 -0.571 -0.641 1.718
Price of Veg 0.079 -0.204 -0.162 0.023 -0.039 0.324 0.039 0.097 0.175 0.178 2.708
Price of oil -0.014 -0.034 -0.049 -0.026 -0.033 0.065 0.027 0.026 0.010 0.027 2.285
Price of 8 -0.042 -0.153 0.058 -0.198 -0.115 0.069 -0.016 0.204 41.103 0.003 2.922
Price of OF -0.258 ~0.1l7 o0.300 -0.040 ~0.173 -0.042 0.101 -0.079 0.100 0.020 3.080
Price of FAFH -0.205 -0.l69 -0.075 -0.139 -0.146 0.128 0.195 0.266 0.141 0.178 2.202
Vegetables
Price of St -0.254 0.037 -0.155 -0.156 -0.253 0.043 0.435 0.164 0.264 0.085 1.090
Price of MP 0.213 -0.130 -0.021 0.048 0.066 0.358 0.052 0.120 0.211 0.216 0.820
Price ofVeg -0.870 -l.321 -0.591 -1.254 -0.958 -0.739 -1.186 -0.477 -l.119 -0.835 3.113
Price of oil -0.006 0.043 0.039 0.056 0.044 0.036 0.077 0.072 0.088 0.078 1.408
Price of S 0.040 0.294 -0.222 0.195 0.076 0.099 0.371 -0.158 0.279 0.143 1.005
Price ofOF -0.074 -0.052 -0.054 -0.065 -0.018 0.041 0.069 0.042 0.060 0.091 1.287
PriceofFAFH -0.017 -0.021 0.087 -0.032 0.008 0.161 0.181 0.237 0.217 0.192 0.687
Oil
Price of St 0.017 -0.946 -0.020 0.239 -0.698 0.387 -0.317 0.421 0.461 -0.237 0.778
Price ofMF 0.042 -0.143 -0.083 -0.034 -0.079 0.223 0.146 0.111 0.052 0.124 1.052
Price of Veg -0.052 0.091 0.093 0.296 0.125 0.112 0.306 0.251 0.367 0.292 1.192
Price ofoil -1.273 -0.695 -0.743 -1.128 -0.770 -1.220 -0.641 -0.698 -1.111 -0.724 1.665
Price of S -0.333 0.250 -0.286 -0.012 0.142 -0.259 0.371 -0.197 0.032 0.233 0.663
Price of OF 0.295 -0.l72 -0.029 0.014 -0.014 0.439 0.020 0.106 0.080 0.134 0.783
Price ofFAFH 0.097 -0.206 -0.202 -0.012 -0.054 0.319 0.115 0.006 0.119 0.196 0.963
Sugar

Price of St -0.226 -0.l44 0.206 1.013 0.063 0.004 0.075 0.543 1.152 0.358 1.452
Price ofMF 0.056 -0.139 0.300 -0.253 -0.123 0.168 -0.039 0.448 -0.199 0.007 1.600
Price of Veg 0.117 0.585 -0.402 0.402 0.155 0.218 0.660 -0.282 0.446 0.262 1.772
Price ofoil -0.218 0.148 -0.l36 0.002 0.085 -0.185 0.167 -0.101 0.012 0.115 1.803
Price of S -0.868 -0.898 -l.017 -1.267 -0.955 -0.822 -0.856 -0.950 -l.239 -0.897 0.723
Price of OF 0.267 —0.195 0.225 -0.238 -0.081 0.357 -0.129 0.327 -0.l97 0.014 2.005
PriceofFAFH 0.122 0.012 -0.143 -0.058 -0.040 0.260 0.123 0.015 0.025 0.119 1.685

 

Note: St = Staples; MF = Meat and Fish; Veg = Vegetables; S = Sugar; OF = Other

Foods; FAFH = Food Away From Home. L = August = lean season, H = November =
harvest, PH = February = post-harvest and P = May = planting; Bold values denote that

the estimated price elasticities are statistically signiﬁcant at the 10% level. Tests of

statistical signiﬁcance cannot be performed for F AF H because their parameters were
recovered using the adding-up restrictions.

135

Table 3-8: Uncompensated and Compensated Price Elasticities for Stage II Model

 

 

 

 

(continued)

Marshallian (Uncompensated) Hicksian (Compensated) I Chow
Commodities L H I PH I P IPooled L I H I PH I P IPooled] Test
Other Foods 7
Price ofSt -0.075 0.162 0.274 0.213 0.062 0.190 0.276 0.569 0.408 0.294 1.112
Price ofMF -0.182 0.099 -0.245 0.055 -0.082 -0.053 0.152 -0.115 0.131 0.020 1.120
Price ofVeg -0.070 0.039 -0.056 0.002 0.014 0.047 0.078 0.050 0.065 0.098 1.392
Price ofoil 0.124 -0.004 0.005 0.006 0.015 0.162 0.006 0.036 0.021 0.039 1.733
Price ofS 0.130 -0.102 0.157 -0.l73 -0.040 0.183 -0.081 0.216 -0.133 0.005 2.857
Price ofOF -0.687 -0.585 -0.851 -0.619 -0.632 -0.584 -0.550 -0.761 -0.561 -0.558 2.032
Price ofFAFH -0.103 0.061 -0.134 -0.045 -0.092 0.055 0.119 0.005 0.070 0.033 1.090

FAFH

Price ofSt -0.200 -0.486 -0.258 -0.674 -0.397 0.236 0.050 0.171 0.142 0.144
Price ofMF -0.109 -0.071 0.061 -0.224 -0.091 0.104 0.176 0.250 0.092 0.149
Price ofVeg -0.074 -0.061 0.027 -0.145 -0.064 0.119 0.121 0.180 0.117 0.132
Price ofoil 0.014 -0.027 -0.043 -0.048 -0.019 0.076 0.019 0.001 0.016 0.036
Price ofS 0.000 -0.056 -0.079 -0.156 -0.062 0.087 0.046 0.006 0.008 0.045
Price ofOF -0.134 -0.092 -0.127 -0.207 -0.147 0.036 0.071 0.003 0.035 0.027
Price ofFAFH -0.919 -0.756 -0.813 -0.895 -0.825 -0.658 -0.484 -0.612 -0.411 -0.533

 

 

 

Note: St = Staples; MF = Meat and Fish; Veg = Vegetables; S = Sugar; OF = Other

Foods; FAF H = Food Away From Home. L = August = lean season, H = November =
harvest, PH = February = post-harvest and P = May = planting; Bold values denote that

the estimated price elasticities are statistically signiﬁcant at the 10% level. Tests of

statistical signiﬁcance cannot be performed for FAF H because their parameters were
recovered using the adding-up restrictions.

Second, the results show that, with the exception of sugar, changes in the price of

non-staples commodities have very small uncompensated and compensated effects on

Bamako households’ consumption of staples in any given season and for the pooled data.

For instance, the results show that a 1 percent increase in the price of vegetables during

the lean season will reduce the quantity demanded of staples by 0.041 percent. A

comparison of the uncompensated and compensated cross-price elasticities shows that

changes in the price of these non—staples commodities have very small income effects

because they occupy a relatively small proportion of households’ budget. On the other

hand, the results indicate that changes in the price of sugar have a positive and

statistically signiﬁcant impact (0.213) on staples’ consumption during the planting

season. A possible explanation for this ﬁnding is that households react to the high grain

136

prices that prevail during the planting season due to low food availability by increasing
their consumption of millet-sorghum‘as an attempt to maintain their calorie levels by
preparing meals such as porridge, usually made with millet and sorghum ﬂour and sugar,
that are consumed in the morning and evening. Moreover, households may also increase
their consumption of tea as a substitute for eating.

Third, the results show that most of the uncompensated cross-price elasticities
indicate net complementarity between staples and food away from home, meat and ﬁsh
and other foods, vegetables and staples, oil and staples, and other foods and food away
from home during the lean season and for the pooled data. However, if households were
compensated for the price changes, the compensated cross-price elasticities indicate that
they would tend to substitute between these commodities.
3.3.3.3. Stage III Price Elasticities
3.3.3.3.1. Uncompensated and Compensated Own-Price Elasticities
Table 3-9, below, presents the compensated and uncompensated own-price elasticities for
the Stage 111 model by season and for the pooled data. The sign of the estimates of own
price elasticities of the Stage 111 model are all negative, indicating that there is an inverse
relationship between price and quantity demanded. All the statistically signiﬁcant
compensated own-price elasticities were smaller than the uncompensated own-price

elasticties, as expected for normal goods.

137

Table 3-9: Compensated and Uncompensated Own-Price Elasticities for Stage III
Commodities by Season

 

 

 

 

Commo- Marshallian (Uncompensated) Hicksian (Compensated) Chow
dities L H I PH I P IPooled L I H I PH I P IPooled Test-
Rice -l.027 -0.607 -0.644 -0.821 -0.767 —0.593 -0.217 -0.193 -0.340 -0.338 1.618
MS -1.380 -0.588 -0.659 -0.598 -0.691 -1.211 -0.389 -0.368 -0.307 -0.487 2.807
Maize -l.903 -1.840 -l.788 -l.977 -1.968 -l.694 -1.492 -1.691 -l.861 -l.759 4.332
Wheat -l.605 -1.453 -2.786 -1.490 -1.759 -l.449 -1.390 -2.667 -1.445 -1.660 1.093
RT -0.678 -0.154 -1.405 -0.777 -0.651 -0.648 -0.155 -1.364 -0.711 -0.591

 

 

 

Note: MS = Millet-Sorghum and RT = Roots and Tubers. L = August = lean season, H =
November = harvest, PH = February = post-harvest and P = May = planting; Bold values
denote that the estimated price elasticities are statistically signiﬁcant at the 10% level.
Tests of statistical signiﬁcance cannot be performed for Roots&Tubers because their
parameters were recovered using the adding-up restrictions.

The own-price elasticities, both compensated and uncompensated, are negative
and for the most part, statistically signiﬁcant. The compensated own-price elasticities
derived from pooled data are smallest for rice (-0.338) and largest for maize (-1.759),
implying that the quantity demanded of rice is far less responsive to own-price changes
than that of maize. The estimated compensated own-price elasticities of maize in all
seasons and for the pooled data and wheat during the lean and post-harvest seasons and
for the pooled data are greater than 1 (in absolute terms), suggesting that the demand for
these commodities is elastic. This means that for each of these commodities, a 1 percent
change in the commodity’s own price has a more than proportionate effect on the
quantity demanded of that commodity. However, one should note that the estimated high
price elasticities of demand for staples, especially maize, are only valid within the price
range observed during the survey year. Without this caveat, it is hard to reconcile the

very high price estimated elasticity of demand for maize with the high year-to-year price

volatility of maize, which implies an inelastic demand.

138

3.3.3.3.2. Uncompensated and Compensated Cross-Price Elasticities

Tables 3-10, below, presents the uncompensated and compensated cross-price elasticity
of demand for staples by season and for the pooled data. With the exception of the lean
season, the results indicate that the price of rice has a positive and statistically signiﬁcant
effect on the consumption of millet-sorghum once the income effects are removed in all
seasons. These results suggest that rice and millet-sorghum are net substitutes, in that
households would turn towards purchasing more millet-sorghum in the face of higher rice
prices. However, once the income effects are accounted for, the uncompensated cross-
price elasticity of rice with respect to millet-sorghum is statistically signiﬁcantly
negative, meaning that rice and millet-sorghum tended become complements. Thus, an
increase in the price of rice would result in reduced consumption of millet-sorghum as the
income effects from rice price changes are stronger than the pure substitution effect.
Rogers and Lowdermilk (1999) found that the effects of changing rice prices did not have
a statistically signiﬁcant impact on millet-sorghum purchases. They attributed this result
to the fact that rice and millet-sorghum occupied different functions in urban households’
diets. Household tended to consume rice at mid-day while millet-sorghum were

consumed in the morning and evening.

139

Table 3-10: Uncompensated and Compensated Price Elasticities for Stagelll Model

 

 

 

 

 

 

 

 

 

 

Marshallian (Uncompensated) Hicksian (Compensated) Chow
Commodities L H I PH I P IPooled L H I PH I P IPooled Test
Rice
PriceofRice -l.027 -0.607 -0.644 -0.821 -0.767 -0.593 -0.217 -0.193 -0.340 -0.338 1.618
Price of MS 0.118 -0.103 -0.456 -0.238 -0.168 0.288 0.054 -0.284 0.031 0.018 2.723
Price of M 0.069 0.155 0.043 0.260 0.158 0.205 0.278 0.143 0.323 0.263 1.357
Price of W 0.028 -0.111 0.099 -0.063 -0.007 0.084 -0.077 0.138 -0.015 0.037 2.077
Price of RT -0.017 -0.083 0.134 -0.033 -0.022 0.015 -0.038 0.195 0.001 0.021 1.292
MS
Price of Rice 0.304 -0.363 -1.504 -0.467 -0.440 0.737 0.133 -0.743 0.055 0.029 2.977
Price of MS -1.380 -0.588 -0.659 ~0.598 -0.691 -1.211 -0.389 -0.368 -0.307 -0.487 2.807
Price of M 0.404 -0.031 0.589 -0.122 0.186 0.539 0.125 0.757 -0.053 0.300 1.962
Price of W -0.207 0.135 0.265 0.236 0.108 -0.152 0.179 0.331 0.288 0.155 3.097
Price of RT 0.053 -0.107 -0.080 -0.020 -0.044 0.086 -0.049 0.022 0.017 0.003 1.667
Maize
Price of Rice -0.013 -0.222 0.208 1.569 0.244 0.654 0.883 0.645 2.451 1.141 4.087
Price of MS 0.411 -0.284 1.139 -0.716 0.162 0.672 0.160 1.306 -0.223 0.561 5.418
Price ofM -1.903 -1.840 -1.788 -1.977 -1.968 -l.694 -1.492 -1.691 -1.861 -1.759 4.332
Price of W 0.355 0.210 -0.196 -0.426 -0.005 0.441 0.308 -0.158 -0.337 0.085 5.345
Price ofRT -0.124 0.012 -0.l62 -0.092 -0.116 -0.074 0.141 -0.103 -0.030 -0.028 4.023
Wheat
Price ofRice -0.563 -1.580 0.222 -0.601 -0.623 0.656 -0.867 1.586 -0.152 0.359 1.267
Price ofMS -0.938 0.525 0.928 1.354 0.239 -0.461 0.812 1.449 1.605 0.665 0.850
Price of M 0.693 0.870 -0.703 -0.504 -0.040 1.075 1.095 -0.400 -0.445 0.200 2.123
Price of W -1.605 -1.453 -2.786 -1.490 -1.759 -1.449 -1.390 -2.667 -1.445 -1.660 1.093
Price ofRT 0.087 0.266 -0.151 0.405 0.338 0.179 0.350 0.032 0.437 0.437 0.570
RT
Price ofRice -0.202 -0.316 1.141 -0.922 -0.422 0.204 -0.323 1.449 0.013 0.186
Price of MS 0.287 -0.l64 -0.054 -0.384 -0.263 0.446 -0.167 0.063 0.138 0.001
Price ofM -0.435 0.383 -0.238 -0.178 -0.210 -0.308 0.381 -0.170 -0.055 0062
Price of W 0.254 0.266 -0.006 0.522 0.404 0.306 0.265 0.021 0.616 0.466
Price ofRT -0.678 -0.154 -1.405 -0.777 -0.651 -0.648 -0.155 -1.364 -0.711 -0.591

 

Note: MS = Millet-Sorghum; M = Maize; W = Wheat; and RT = Roots and Tubers. L =
August = lean season, H = November = harvest, PH = February = post-harvest and P =
May = planting; Bold values denote that the estimated price elasticities are statistically
signiﬁcant at the 10% level. Tests of statistical signiﬁcance cannot be performed for
Roots&Tubers because their parameters were recovered using the adding-up restrictions.

Furthermore, both the pooled data and seasonal results indicate that the price of rice

has a positive effect, both compensated and uncompensated (except during the lean

season), on the consumption of maize, meaning that rice and maize are net substitutes.

Hence, households are more likely to move towards maize during the planting season

when the price of rice tends to be very high. Concerning wheat, the results indicate that

140

the price of rice has a statistically signiﬁcant large uncompensated effect on the
consumption of this commodity only during the harvest season. During that period, a 1
percentage increase in the price of rice results in a 1.580 percent decrease in the
consumption of wheat.

Moreover, the results provide practically no evidence to support the hypothesis that
cereals and roots and tubers are substitutes (e. g., Timmer and Alderman, 1979; Pakpahan,
1988), implying that Bamako households, in the face of higher prices of cereals, would
not consume more roots and tubers. The results show that the price of roots and tubers
has a negative, both compensated and uncompensated, statistically signiﬁcant, but very
small, impact on the demand for rice during the lean season.

3.3.4. Sensitivity Analysis

Much research in food policy has focused on the effects of changes in households’
income levels on the food income and price elasticity at a given point in time (Alderman
(1990), Rogers and Lowdermilk (1991), and Dorosh et a1. (1994)). These previous
studies have shown that low-income households are much more sensitive to changes in
incomes and prices than high-income households (Timmer, 1983). The robustness of the
estimated price and income parameters are tested in this section using sensitivity analyses
in which different scenarios are simulated by manipulating real income levels and tracing
their effects.

More speciﬁcally, in this section, sensitivity analyses are performed in order to
determine the effect of changes in Bamako households’ real incomes on (1) the income

elasticity of food and (2) the own-price elasticity of food in any given season.

141

3.3.4.1. Effects of Changes in Households’ Real Incomes on the Income Elasticity of
Food

This ﬁrst scenario examines the impact of changes in households’ real incomes on the
income elasticity for food (as an aggregate commodity).33 The expected change in the
food income elasticity is derived using the following equation:

Anfood = 1+ [Brood / Aw food]. where Ame = FE/ ATE (3)

In this equation, FE represents food expenditure per adult equivalent and TB is

total real expenditure per adult equivalent. Food expenditures are held constant while
several simulations are performed on households’ average weekly real total expenditures
per adult equivalent, used as a proxy for real income. The effects of changes in
households’ total expenditures on the income elasticity for the food are traced, holding

everything else constant. The baseline parameters and variables are presented below in

 

 

 

 

 

 

 

Table 3-11.
Table 3-11: First Scenario Base Parameters and Variables

Variables Baseline

Lean Harvest Post-Harvest Planting

Share of food 0.348 0.416 0.439 0.486
Food expenditures/AB 2368 2223 2131 2149
Total real expenditure/AB 8123 5242 5458 4461
Food income elasticity 0.626 0.463 0.577 0.574

 

The impacts of changes in households’ total expenditures on the income elasticity
for the food are presented, below, in Figure 3-2. One should note that the model, by
construction (see equation (3)), will show a uniform decline in the income elasticity of
demand for food as households reach higher income levels. Thus, all the sensitivity

analysis is testing for is the rate of change of the estimated income elasticity of demand

 

33 The income elasticity was derived using the following formula: m = 1+ [[5, / w ,]

142

for food as households’ expenditures increase. First, the results indicate that households’
demand for food becomes highly inelastic as their real income increases. For instance, the
results indicate that as households’ real incomes increase from 2000 to 10,000
FCFA/AE/week (or by 400%), the income responsiveness of their demand for food
decreases from 3.5 to 0.4 (or by 775 %) during the lean season. Second, the results show
that there is a uniform shift in the entire income-food consumption relationship across
seasons, suggesting that the impact of real income on the demand for food is not constant
across seasons. The results reveal that there is a substantial difference in the income
responsiveness of demand for food between the lean and the other seasons. Bamako
households’ demand for food is most responsive to changes in real income in August,
during the lean season, when their real incomes are high and the demand for non-food
commodities is high. There is no sizable difference in the impact of real income on

households’ demand for food between the harvest, post-harvest, and planting seasons.

Elasticity

 

 

 

 

 

 

 

 

 

 

 

 

o I I I I T I I I I I I T I I I I

2000 3000 4000 5000 6000 7000 8000 9000 10000
Income (FCFA/AE/WEEK)

r—O—Aug *Nov ”in-Feb _"" 'May I

 

 

Figure 3-2: Effect of Changes in Real Incomes on
the Income Elasticity of Food by Season

143

3.3.4.2. Effects of Changes in Households’ Real Incomes on the Food Price Elasticity
The second scenario investigates the effects of changes in households’ real expenditures
on the compensated own-price elasticity of food (i.e., no income effects). 34 The expected
change in the compensated own-price income elasticity of food is calculated using the
following equation:

Ami = g“ + m . Aw i, i = Food and AWfood = FE/ ATE (4)

In this equation, FE represents food expenditure per adult equivalent and TE is
total real expenditure per adult equivalent. Food expenditures are held constant while
several simulations are performed on households’ average weekly real total expenditures
per adult equivalent. The effects of changes in households’ total expenditure levels on the
own-price elasticity of demand for food are traced. The baseline parameters and variables
are presented, below, in Table 3-12.

Table 3-12: Second Scenario Base Parameters and Variables

 

 

 

 

 

 

 

Variables Baseline
Lean Harvest Post-Harvest Planting
Share of food 0.348 0.416 0.439 0.486
Food expenditures/AB 2368 2223 2131 2149
Total real expenditure/AB 8123 5242 5458 4461
Food own-price elasticity -0.480 -0.353 -0.320 -0.364

 

Figure 3-3, below, shows the impacts of changes in households’ real incomes on
the compensated own-price elasticity of demand for food. One should note that the
model, by construction (see equation (4)), will show a uniform increase in the own-price
elasticity of demand for food as households expenditure levels increase. However, the

model allows a discussion on the magnitude of the decline as real income increases. The

 

3’ The compensated price elasticity was derived using the following formula: m = £0- + n, . w,-

144

results indicate that the responsiveness of households’ food consumption to changes in
the price of food decreases from well above 2 percent to near zero as their real income
increases from 1000 to 10000 FCFA/AE/week. Moreover, there is a uniform shiﬁ in the
entire compensated own-price elasticity-income relationship across seasons, indicating
that the effect of own-price changes on the demand for food is not constant across
seasons. Figure 3-3 clearly indicates that households’ demand for food is far more
responsive to own-price changes during the lean season than during the harvest, post-

harvest, and planting seasons.

 

 

 

 

 

 

 

 

Elasticity
0
-0.5 a
-1 .
-1.5 -
a
-2 -
-2.5 ~
-3 4
-3.5 -
-4 -
-4.5 y
1000 Income 10000
I——Aug —t—-Nov— ' 'Feb ' 'K- MayI

 

 

Figure 3-3: Effect of Changes in Income Levels on the
Own-Price Elasticity of Food by Season

3.4. Conclusions

In this essay, the Almost Ideal Demand System was applied to a three-stage demand
model for different seasons in order to estimate the impact of seasonal changes in
Bamako households’ real incomes and relative prices on their consumption patterns.
First, the results indicate that price, income, and household size factors account for a
substantial part of the observed variation in the budget share devoted to the commodities

considered in the Stage I, II, and 111 models.

145

Second, the study ﬁnds that Bamako households’ consumption is responsive to
changes in real incomes and relative prices in any given season and that that there are
seasonal changes in income and price responsiveness for all the commodities in the three
demand models. This implies that the impact of a uniform food policy on the quantity and
quality of food available in Bamako households will vary by season.

Third, the results indicate that Bamako households engage in food consumption
smoothing from seasonal shocks in real incomes. Food consumption smoothing was
achieved at the expense of non-food commodities such as health and durable goods
(housewares and education), of non-staple foods, and through signiﬁcant substitutions
among and between broad commodity groups.

Fourth, the estimated price elasticities indicate that (1) the price of food has strong
and statistically signiﬁcant uncompensated effects on the demand for non-food
commodities, such as health and education; (2) the price of staples has striking impacts
on the demand for non-staple foods, which are sources of high-quality protein and
micronutrients and; (3) the price of rice has a positive effect on the consumption of
maize, meaning that rice and maize are net substitutes.

The ﬁndings of this essay have several implications for development planning in
Mali. First, the high absolute level of the income elasticities, even for food, underscores
the extreme level of poverty and unmet “basic needs” that prevail in Bamako. As a
consequence, the results suggest that policies that aim at increasing households’ real
income will cause substantial improvements not only in the quantity of food available in

urban households but also in the demand for non-food commodities. Rapid grth in the

146

demand for non-food commodities could translate into sizable rise in employment, to the
extent that these commodities can be produced domestically.

Second, the empirical results for food commodity groups showed that as Bamako
households’ real income increases, they will increase their expenditure on non-staple
commodities (e. g., meat and ﬁsh and vegetables) more rapidly than on staple foods. As a
consequence, households will diversify their diets, through greater consumption of non-
staple commodities, as their income grows. This ﬁnding suggests that the pattern of
production within the agricultural sector in Mali will have to change with economic
growth, as increased specialization in livestock and horticultural production will be
required. Hence, greater allocation of resources and investment in the production and
marketing of horticultural commodities offer the potential to substantially reduce
malnutrition, especially vitamin A deﬁciency, increase employment, and reduce poverty

in urban areas.

147

REFERENCES

Alston, J. M., K.A. Foster, and RD. Green, (1994). Estimating Elasticities with the
Linear Approximate Almost Ideal Demand System: Some Monte Carlo Results. Review
Economics And Statistics, Vol.76, 351-356.

Alston J M ,& Chalfant J. 1993. The silence of Lambdas: A test of the almost ideal and
Rotterdam models. American Journal of Agricultural Economics 75 300—3 13.

Deaton, A and Muellbauer J. Economics and consumer behavior. Cambridge,
Cambridge University Press: 1980a.

Deaton, A. The Analysis of Household Surveys: A Microeconomic Approach to
Development Policy. World Bank. The John Hopkins University Press. Baltimore,
London: 1997.

Deaton, A. & Zaidi, S., 1999._Guidelines for Constructing Consumption Aggregates for
Welfare Analysis. Papers 192, Princeton, Woodrow Wilson School - Development
Studies.

Deaton, A and J. Muellbauer. 1980. An almost ideal demand system. The American
Economic Review 70: 312-326.

Dembelé, N. N. and J .M. Staatz, June 1999. The Impact of Market Reform on
Agricultural Transformation in Mali. MSU Agricultural Economics Staﬁr Paper No. 99-
29.

Duncan, H. Boughton. 1994. A Commodity Subsector Approach to the Design of
Agricultural Research: the Case of Maize in Mali. Ph.D. dissertation, Michigan State
University.

Green R, and Alston J M. 1990. Elasticities in AIDS models. American Journal of
Agricultural Economics May 442—445.

Moschini, G. 1995. Units of measurement and the stone index in demand system
estimation. American Journal of Agricultural Economics 77: 63-68.

Pakpahan, Agus 1988. Food Demand Analysis in Urban West Java, Indonesia. Ph.D.
dissertation, Michigan State University.

Paxon, Christina H., 1993. Consumption and Income Seasonality in Thailand. Journal of
Political Economy, 101(1): 3 9-72.

Pollak, Robert A. and T.J. Wales. Demand System Speciﬁcation and Estimation. New
York: Oxford University Press, 1992.

148

Raunikar, Robert and Huan, Chung-Liang. Demand Analysis: Problems, Issues, and
Empirical Evidence. Ames: Iowa State University Press, 1987.

Reardon et al., 1999. Household Consumption Responses to the Franc CFA Devaluation:
Evidence from Urban Mali. Food Policy 24 (1999), 517-534.

Rogers, BL. and Lowdermilk, M. (1991). Price Policy and Food Consumption in Urban
Mali. Food Policy 16, 461-473.

Sadoulet, Elisabeth, and de J anvry, Alain. Quantitative Development Policy Analysis.
Chapter 6. Baltimore, MD: The Johns Hopkins University Press, 1995.

Sahn, E. 1989. Seasonal Variability in Third World Agriculture: The Consequences for
Food Security. Baltimore and London: John Hopkins University Press.

Teklu, Tesfaye, 1996. Food Demand Studies in Sub-Saharan Aﬁ'ica: A Survey of
Empirical Evidence. Food Policy, Vol. 21, No. 6, pp. 479-496, 1996.

Timmer et a1. 1983. Food Policy Analysis. Published for the World Bank. John Hopkins
University Press.

USAID. 2000. http://www.usaid.org/Mali urban proﬁle

149

APPENDIX 3

150

   

and Poultry
Fish

    
   
   
   
 
 
    
 
  
 
 
   

omatoes (fresh and concentrate)
Vegetables

Oil

 
 

Milk
Sweetned Milk
milk

ea Lipton

and Other Beverages

amarinds
Fruits (Dates, Orange, Raisin)

15]

Table A3-2: Deﬁnitions of Variables and Summary Statistics

 

 

 

 

 

 

 

 

Vari:blel Description Au ust November February May Pooled
lw Hem SD eanISD MeanISDIMeanLSD MeanISDJ
Share of food 0.348 0.127 0.416 0.149 0.439 0.170 0.486 0.152 0.422 0.157
Share ofDG 0.146 0.156 0.183 0.195 0.089 0.103 0.097 0.099 0.129 0.148
W3 ShareofSDG 0.175 0.161 0.100 0.111 0.133 0.142 0.069 0.065 0.119 0.130
w4 Share ofhealth 0.082 0.118 0.076 0.104 0.091 0.106 0.130 0.181 0.095 0.132
wS Share ofEU 0.085 0.069 0.081 0.062 0.108 0.086 0.096 0.075 0.092 0.074
w6 Share ofOND 0.037 0.046 0.043 0.043 0.037 0.032 0.048 0.046 0.041 0.042
w7 Share ofServices 0.127 0.124 0.100 0.090 0.103 0.118 0.075 0.072 0.101 0.104
pl Price of food 204 100 234 110 218 94 264 101 230 103
p2 Price of DO 6242 5329 5292 3325 5269 5080 6421 5744 5806 4938
p3 Price of SDG 1678 991 1576 955 1922 1070 1365 724 1635 956
p4 Price of health 1854 1371 2159 1733 1871 868 2610 2094 2124 1598
p5 Price of EU 123 62 121 44 171 87 136 79 138 72
p6 Price ofOND 126 27 108 18 113 30 126 37 118 30
p7 Price of Services 225 61 214 87 170 54 158 40 192 68
AE Household size 13 7 13 7 l3 8 13 8 l3 8
X/P TEAE 9521 6086 7155 3172 6818 3898 6076 4287 7392 4633
StageII
wl Share of staples 0.307 0.084 0.346 0.094 0.348 0.088 0.348 0.098 0.337 0.092
w2 Share ofMF 0.150 0.075 0.159 0.087 0.153 0.079 0.135 0.060 0.149 0.076
w3 Share ofVeg 0.135 0.039 0.118 0.036 0.124 0.039 0.112 0.041 0.122 0.039
w4 Share ofoil 0.044 0.024 0.030 0.019 0.036 0.021 0.027 0.014 0.034 0.021
w5 Share of sugar 0.061 0.026 0.066 0.032 0.070 0.036 0.070 0.039 0.067 0.034
w6 ShareofOF 0.119 0.041 0.106 0.042 0.106 0.031 0.103 0.047 0.108 0.041
w7 Share ofFAFH 0.184 0.114 0.176 0.130 0.164 0.095 0.206 0.147 0.182 0.123
pl Price of staples 251 41 268 44 251 46 257 44 257 44
p2 Price of MP 960 304 1040 334 999 277 996 233 999 288
p3 Price of Veg 554 193 523 126 442 136 420 125 485 156
p4 Price of oil 678 229 614 294 671 336 577 198 635 270
p5 Price of sugar 460 96 439 74 464 98 438 42 450 81
p6 Price of OF 1037 648 734 381 622 267 680 377 769 465
p7 Price of FAFH 128 82 109 60 105 56 94 56 109 65
AE Household size 13 7 l3 7 l3 8 l3 8 13 8
X/P FEAE 2368 669 2223 676 2131 829 2149 5 10 2218 681
StageIlI
wl Share ofrice 0.524 0.138 0.520 0.142 0.548 0.137 0.537 0.155 0.532 0.142
w2 Share ofMS 0.205 0.120 0.209 0.117 0.209 0.117 0.300 0.148 0.231 0.131
w3 Shareofmaize 0.164 0.130 0.164 0.153 0.121 0.105 0.071 0.062 0.130 0.123
w4 Share of wheat 0.067 0.080 0.046 0.069 0.048 0.063 0.054 0.090 0.054 0.076
w5 Share ofRT 0.039 0.035 0.061 0.074 0.074 0.055 0.038 0.037 0.053 0.054
pl Price of rice 267 18 275 27 261 18 263 15 267 21
p2 Price ofMS 147 40 151 41 133 33 174 49 151 44
p3 Price of maize 187 76 182 51 194 95 229 97 198 83
p4 Price of wheat 734 395 954 615 880 310 831 253 850 421
p5 Price of RT 238 97 157 107 199 36 304 151 224 1 18
AE Household size 13 7 13 7 l3 8 l3 8 l3 8
X/l’ SEAE 913 437 981 346 889 303 851 340 909 359

 

Note: All prices and expenditures are in CF A Francs. D0 = Durable Goods; SDG = Semi-Durable Goods

EU = Energy and Utilities; and OND = Other Non-Durable Goods; TEAE = Total Expenditure per Adult
Equivalent; S = Staples; MF = Meat and Fish; Veg = Vegetables; S = Sugar; OF = Other Foods; F AF H =
Food Away From Home; F EAE = Food Expenditures per Adult Equivalent; MS = Millet-Sorghum; M =

Maize; W = Wheat; and RT = Roots and Tubers; and SEAE = Staples Expenditures per Adult Equivalent.

152

CHAPTER 4
ESTIMATING THE EFFECTS OF SEASONAL CHANGES IN REAL INCOMES
AND RELATIVE PRICES ON HOUSEHOLDS’ DEMAND FOR NUTRIENTS
IN BAMAKO, MALI
4.1. Introduction
The state of poor nutrition, caused by households’ inability to meet minimum energy,
protein, and other essential nutrients’ requirements, is particularly severe in Mali. In
2001, the Mali demographic and health survey (DHS) reported that 22 percent of women
between the ages of 15 and 19 years and 11 percent of those between 20 and 24 years
suffered from chronic energy deﬁciency. More than 10 percent of the population has
blinding disorders, such as trachoma, due to a vitamin A deﬁciency (DHS, 2001). An
estimated four out of ﬁve children (82 percent) under 5 years old have anemia, which is
caused by a deﬁciency in iron, and about 63 percent of women present a form of anemia
(DHS, 2001). Iron deﬁciency contributes signiﬁcantly to reduced resistance to infection,
impairment of some cognitive functions, and maternal deaths (FAO, 1997).

Furthermore, as shown in essay 1 (chapter 2), in 2000-2001, Bamako households’
mean real expenditures varied considerably across seasons. Households’ real
expenditures decreased by 36 percent between the lean and post-harvest season, increased
by 4 percent between the harvest and post-harvest season, and dropped by 18 percent
between the post-harvest and planting season. The seasonal variation in households’ real
expenditures could be partly explained by seasonal changes in the relative prices of goods
and services, the size and timing of remittances, and the fact that households’

expenditures on many non-food commodities tend to be highly seasonal.

153

Empirical evidence (Sahn, 1989 and Dostie, 2000) suggests that seasonal changes
in real income affect the quantity and quality of foods available in households and,
thereby constitute an important determinant of household food security. Therefore,
understanding how the demand for nutrients responds to changes in real incomes and
relative prices is crucial for the formulation and implementation of policies to help
improve nutrition in developing countries such as Mali. In the last few years, the
empirical literature (e. g., World Bank (1981), Behrman and Deolalikar (1987), Strauss
and Duncan (1990), and Bouis and Haddad (1992)) has largely focused on the role of
income on nutrient consumption despite the widespread implementation of food subsidy
programs in developing countries. Moreover, the development literature has two
divergent views on the issue of whether nutrient intake responds to income. The
traditional view postulates that increases in income will lead to nutrient improvement in
households, hence that economic grth would eradicate hunger and malnutrition
(World Bank, 1981). In contrast, recent studies (e. g. Behrman and Deolalikar (1987)) and
Bouis (1994)) argue that income growth will not result in substantial improvements in
nutrient intake. The current literature claims that low-income households will increase
food expenditures with rising income, but that the marginal increase in income is spent
on food attributes other than nutrients.35 A potential explanation for the diverging views
is that the relationship between income and nutrient intake depends on the country, use of
cross-section, panel, or time series data, model speciﬁcation, and estimation technique

(Dawson and Tifﬁn, 1998).

 

3’ Examples of food attributes include degree of processing and taste (Behrman and Deolalikar (1987)). For
instance, Bamako households’ preference for rice has largely been attributed to taste factors and to the fact
that rice takes less time, fuel, and labor to prepare (Rogers and Lowdermilk, 1991).

154

This study tests the hypothesis that households’ demand for nutrients is
responsive to changes in their real incomes and relative prices and that the magnitude of
the nutrient income and price elasticities will change from one season to another. Most of
the empirical evidence on the determinants of nutrient demand (Behrman and Deolalikar
(1987), Bhargava (1991), and Subramanian and Deaton (1996)) has focused on the
effects of income on the demand for nutrients. However, the empirical evidence on the
effects of food price changes on the demand for nutrients is relatively scarce. Therefore,
this study, through the estimation of nutrient-price elasticities by season and for the entire
year, attempts to make a signiﬁcant contribution to food policy formulation in Mali. The
ﬁndings of this study would be important for policy design, as it would mean that (1) the
policies that aim at increasing households’ real incomes will also improve their nutrition,
(2) food prices can be used as instruments to reduce malnutrition in households, and (3)
the effectiveness of such policies will be contingent upon whether or not they are
systematically synchronized with the short-run response of households’ consumption
patterns to income and price changes.36 The effects of seasonal changes in real incomes
and relative prices on households’ demand for nutrients have not been assessed in
Bamako, Mali.

The general objective of this study is to estimate the impact of seasonal changes
in real incomes and relative prices on the households’ demand for nutrients in Bamako,
Mali. The speciﬁc objectives of this study are threefold. First, the study estimates
nutrient-income elasticities in order to determine if (1) household demand for nutrients is

responsive to changes in real incomes and (2) the nutrient-income elasticities are stable

 

3’ Temporal targeting mechanisms, such as seasonal income transfers to low-income households and
seasonal imports of rice, are examples of programs or policies that are season-speciﬁc.

155

across seasons. Second, the study seeks to compute nutrient price elasticities in order to
identify whether (1) the demand for nutrients is responsive to changes in relative food
prices and (2) the sign and magnitude of the nutrient price elasticities depend on the
season considered. The ﬁnal task of the study is to perform sensitivity analyses on the

estimated nutrient income and price elasticities using several simulation scenarios.
4.2. Methods

4.2.1. Nutrient Demand Model
This study is primarily concerned with the relationship between households’ real
incomes, relative prices, and nutrient availability, which is investigated for four seasons
(planting, lean (pre-harvest), harvest, and post-harvest) and for the entire year. The
demand for total calories, calories from staples, calories from other foods, protein,
calcium, iron, and vitamin A is estimated separately for each season and the entire year
using Engel functions.37 Following Skouﬁas (2002), the nutrient demand functions are
speciﬁed as a log-linear function of the form below:
lnNkht=0L+BlnYm+ylan+61nAEht+ukm (1)
where, k indexes a nutrient (calories, protein, calcium, iron, and vitamin A)

h indexes an household (h = 1,. . .,40)

t indexes seasons (t = 1, 2, 3, 4)
N is nutrient demand (i.e., amounts of nutrients available in household per
adult equivalent (AE)

Y is total real household expenditure per adult equivalent (AE)

 

’7 These particular nutrients were chosen because of the main types of nutrient deﬁciencies that persist in
Mali.

156

P is a vector of food prices (P1 = price of rice, P2 = price of millet-sorghum,
P3 = price of beef, P4 = price of dry ﬁsh, and P5 = price of green leaves)

AB is household size in adult equivalents”

u is an error term

Household demand for nutrients is expressed as a function of food prices, real
incomes, and household size. Real expenditures per adult equivalent (Y) are used as a
proxy for income and are calculated by deﬂating nominal expenditures by the Laspeyres
price index. Unit values, used as proxies for prices, were computed as the ratio of total
household expenditure on a good divided by the total quantity consumed of the good.
The prices of rice and millet—sorghum (PI and P2) are included in the analysis in order to
measure the effect of staple prices on nutrient demand. The prices of beef (P3) and dry
ﬁsh (P4) are chosen to assess the impact of meat and ﬁsh prices on the demand for
nutrients as these foods are important sources of protein. The price of green leaves (i.e.
potato leaves, spinach) (PS) is included in the analysis to account for the effect of
vegetable prices on nutrient demand estimates since green leaves are the main sources of
vitamin A and calcium in urban households’ diets.

This study assumes that all the explanatory variables (prices, income, and household
size) are exogenous (i.e., uncorrelated with the error term). The Ordinary Least Squares
(OLS) method is chosen to estimate the parameters of the nutrient demand functions
because it yields estimates that are unbiased and consistent under the exogeneity
assumption. OLS has been widely applied in many empirical studies to estimate nutrient

demand functions (Rogers and Lowdermilk (1991), Subramanian and Deaton (1996), and

 

38 The data on household size was converted into adult equivalents using the following scales: male > 14
years = 1.0; female > 14 years = 0.8; children = 0.5 (Duncan, 1994).

157

Skouﬁas (2002)). The demand for total calories, calories obtained from staples, calories
derived from other foods, and protein, calcium, iron, and vitamin A is estimated, as
speciﬁed in Equation (1), by Ordinary Least Squares (OLS) separately for each season
and for the pooled data. 39 The stability of the estimated nutrient income and price
elasticities across seasons is assessed using the Chow test.

4.2.2. Data

The panel data used in this study is from a 2000-2001 survey undertaken in Bamako by
the Direction Regionale du Plan et de la Statistique (DRPS) of the Direction Nationale de
la Statistique et de l’Informatique (DNSI) and the Projet d'Appui au Systéme
d'Information Décentralisé du Marché Agricole (PASIDMA) of Michigan State
University (MSU), the Assemblée Permanente des Chambres d’Agriculture du Mali
(APCAM), and the Centre d’Analyse et de Formulation de Politiques de Développement
(CAFPD).The survey was conducted in four rounds and covered the same 40 Food
Consumption Units (FCU) in each round.

The nutrient estimates were derived from the at-home food consumption data on
the quantities of food consumed and data on the nutrient composition of foods.40
Nutrient values exclude nutrients from the inedible or non-servable components of foods
(i.e., bones). Losses from trimming, cooking, plate wastage, and spoilage are not

accounted for in these values.41 The nutrient estimates computed this way represent

 

’9 It is legitimate to estimate each of these demand functions separately, using OLS, rather than as a system
of equations because all the independent variables are assumed to be exogenous (Deaton, 1997). In this
case, there is no simultaneity bias, which is the bias that results from using OLS to estimate an equation in
a simultaneous equation model (Wooldridge, 1999).

‘0 The food composition data come from the food composition table for Mali prepared by Sundberg and
Adams (1998) and from the USDA’s Nutrient Data Bank System (2003).

‘" The Food and Agricultural Organization (FAO) assumes that losses from trimming, cooking, plate
wastage, and spoilage represent about 10 percent.

158

nutrients in foods that are available for household consumption and not actual nutrient
intakes by individuals.

Summary statistics of the variables along with detailed information on the
nutrients contributed by major food groups and speciﬁc food items are presented in
Tables A4-l through A4-3 of the Appendix.

4.3. Empirical Results

The demand ﬁmctions, as speciﬁed in equation (1 ), were estimated by ordinary least
squares for calories, total calories, calories obtained from staples, calories derived from
other foods, protein, calcium, vitamin A and iron for the pooled data and for each season
separately. The estimated coefficients can be interpreted directly as price and income
elasticities since both the dependent and independent variables are expressed in
logarithms. Estimates of the nutrient demand functions, their associated t-values and F
statistics for each season and for the pooled data, and the Chow test results are presented
in Tables A4-4 of Appendix 4.

First, the results indicate that the prices of major food commodities, real income,
and household size factors account for part of the observed variation in the amounts of
nutrients available for household consumption at any given season. For instance, the
goodness-of-ﬁt measure, R2, for the calorie equation ranges from 0.126 during the
harvest season to 0.510 during the lean season, suggesting that, as a group, the price,
income, and household size variables explain about 12 to 51 percent of the observed
variation in calorie availability.

Second, the estimated results show that the demand for nutrients in Bamako

households is responsive to changes in real incomes and relative prices. Out of 35

159

estimated nutrient-income elasticities, 22 are statistically signiﬁcant at least at the 10
percent level. Out of 175 price parameters, 40 are statistically signiﬁcant at least at the 10
percent level.

Third, the null hypothesis of stability in the nutrient income parameters across
seasons was rejected at the 10 percent level for all the estimated coefﬁcients, except for
calcium, suggesting that there is a statistically signiﬁcant shift in the estimated nutrient-
income elasticities across seasons. Moreover, the Chow test results indicate a degree of
non-constancy of many price parameters across seasons, as the test of stability in the
price coefficients was rejected at the 10 percent level for 13 out of 35 estimated
coefficients.

4.3.1. Nutrient-Income Elasticities

The nutrient-income elasticities provide information on the response of nutrient demand
to a change in households’ real incomes, holding other factors ﬁxed. In this sub-section,
the effects of seasonal changes in Bamako households’ real incomes on the demand for
nutrients are examined in order to determine if the demand for nutrients is responsive to
changes in real incomes and if nutrient-income elasticities are stable across seasons.
4.3.1.1. Calories

Table 4-1, below, presents the calorie-income elasticity of demand by season and for the
pooled data and the results of the Chow test. First, the results indicate that real income
has a statistically signiﬁcant impact, at the 1 % level, on the demand for calories in all
periods except the harvest and planting seasons. Many previous studies, such as Strauss
and Thomas (1990), Rogers and Lowdermilk (1991), Bouis and Haddad (1992), and

Subramanian and Deaton (1996), have also found a statistically signiﬁcant relationship

160

between calories and income. The pooled data results, in Table 4-1, show that on average,
a 1 percent increase in households’ real annual incomes increases calorie availability by
0.162 percent. This estimate is almost ﬁve times smaller than that found by Rogers and
Lowdermilk (1991) of 0.760, suggesting that Bamako households’ demand for calories
becoming may be increasingly less responsive to changes in income.

Table 4-1: Calorie-Income Elasticities by Season and for the Pooled Data
Nutrient L H PH P Pooled Chow

TotalCalories 0.193* 0.102 0.171" 0.166 0.162" 4.390“

Note: *, **, and *** denote signiﬁcance at the 1 %, 5 %, or 10 % level, respectively.
L = August = lean season, H = November = harvest, PH = February = post-harvest and P
= May = planting.

 

Second, the results, in Table 4-1, also indicate that improvements in households’
real incomes will have a positive impact on calorie purchases during the lean and post-
harvest seasons. Moreover, the Chow test results indicate that the relationship between
households’ income and amounts of calories available for consmnption is not constant
across seasons. A 1 percent increase in urban households’ real incomes will increase
calorie availability by 0.193 % and 0.170 % during the lean and post-harvest seasons,
respectively.

Furthermore, the results, in Table 4-1, also allow a comparison of the estimated
calorie-income elasticities against the food income elasticities previously derived in the
second essay.42 Calorie-income elasticities are expected to be lower than food-income
elasticities because households will tend to substitute between and within commodity

groups to maintain constant calorie consumption (Subramanian and Deaton, 1996). The

 

’2 The food income elasticties estimated for the Stage I regression model in the second essay of this study
are 0.626 in August, 0.463 in November, 0.577 in February, 0.574 in May, and 0.516 for the pooled data.

161

results of this study are consistent with those ﬁndings, as the estimated food income
elasticties are substantially larger than the estimated the calorie-income elasticities in
each season and annually. This ﬁnding is evidence that households are upgrading the
quality of their diets, substituting more expensive sources of calories for cheaper sources,
as their income increases. These results are consistent with ﬁndings of the descriptive
analysis (Essay 1) that shows that diet diversiﬁcation occurs as households’ incomes rise.

Following Skouﬁas (2002), the effects of speciﬁc foods on the demand for
calories are examined by performing separate regressions for calories from staples and
calories from other foods. Table 4-2, below, presents the income elasticity of calories by
food source for each season and annually. The income elasticity of demand for calories
from staples is statistically signiﬁcant at the 1 percent level for the pooled data and at the
10 percent level for the lean season, whereas the income elasticity of demand for calories
from other foods is statistically signiﬁcant, at least at the 5 percent level, in all seasons
and for the pooled data.

Table 4-2: Calorie-Income Elasticities by Food Source

 

Nutrient L H PH P Pooled Chow
Calories ﬁom staples 0136*" -0.034 0.080 0.114 0.070* 1800*"

Calories ﬁ'om other foods 0.336* 0.401* 0.326" 0.310" 0.364“ 5.860*
Note: *, **, and *** denote signiﬁcance at the 1 %, 5 %, or 10 % level, respectively.

L = August = lean season, H = November = harvest, PH = February = post-harvest and P
= May = planting.

The pooled data results show that an increase in real incomes will have a positive
and less than proportionate impact on both the amounts of calories from staples (0.070)
and other foods (0.364). These results also indicate that on average, annually, calories

from staples are far less responsive to changes in real incomes than calories ﬁ'om other

foods. In addition, the income elasticity of demand for calories from other foods exceeds

162

that for calories for staples in all seasons considered. For instance, during the lean
season, a 1 percent increase in real incomes will improve calorie availability from staples
and other foods by 0.136 and 0.336 percent, respectively. This ﬁnding suggests that
households will increase their consumption of other foods more rapidly than that of
staples as their real incomes increases, indicating that households will tend to shift to
more expensive sources of calories as they get richer.

A comparison of the estimated coefﬁcients by season reveals that focusing only
on the income elasticity of demand for total calories could be misleading because it may
mask opposing changes in the income elasticity of calories for speciﬁc foods (Skouﬁas,
2002). For instance, during the harvest season, a 1 percent increase in real income will
result in a 0.102 percent increase in total calorie availability (Table 4-1). Once the
income effects are decomposed by food source, the results indicate that much of the
increase in calorie availability may be attributed to increases in the amounts of calories
obtained from other foods (0.401). The income elasticity for calories from staples, which
is negative (-0.034) and not statistically signiﬁcant, indicates that increases in real income
in the harvest season will have no effects of the amounts of calories derived from staples.
Furthermore, the Chow test reveals that there is evidence of instability in the estimated
income parameters across seasons for both calories obtained from staples, at the 10
percent signiﬁcance level, and calories derived from other foods, at the 1 percent
signiﬁcance level. Hence, Bamako households will respond to marginal increases in their

real incomes over that period by increasing their consumption of other foods only.

163

4.3.1.2. Protein, Calcium, Vitamin A, and Iron

Table 4-3, below, presents the estimated nutrient-income elasticities for protein, calcium,
vitamin A, and iron for each season and for the pooled data. First, the results indicate
that, on average annually, the demand for nutrients is responsive to changes in Bamako
households’ real incomes. The nutrient-income elasticities derived from the pooled data
are statistically signiﬁcant at the 1 percent level for protein, vitamin A, and calcium, and
at the 5 percent level for iron. However, when separate regressions are run by season, the
results show that during the lean season the demand for all nutrients, except calcium, is
responsive to incremental changes in real incomes while, during the harvest season, none
of the nutrient-income elasticities is statistically signiﬁcant. This means that once prices
and household size are controlled for, changes in households’ real incomes have no effect
on the demand for protein, vitamin A, calcium, and iron during the harvest season. This is
quite surprising because, as shown in Table 4-2, changes in real incomes have a
statistically signiﬁcant impact on calories obtained from other foods during the harvest
season.

A possible explanation for these seemingly contradictory ﬁndings is that the
sizeable decline in households’ total real expenditures (36 percent (Essay 1)) between the
lean and post-harvest season may push households towards subsistence levels of food
consumption. In this context, households will protect their food consumption levels
primarily through staples. Hence, although marginal increases in households’ real
incomes will be devoted to acquiring non-staple foods, the increase in non-staples
consumption may not be sufficiently large enough to have a substantial impact on the

availability of protein and micronutrients.

164

Table 4-3: Nutrient Income Elasticities by Season and for the Pooled Data

 

 

Nutrient L H PH P Pooled Chow
Protein 0.210" 0.090 0.192" 0.213M 0.191 * 4.150"
Vitamin A 0.725" 0.492 0.680" 0.597 0.721 "' 2.240"
Calcium 0.128 0.097 0.160 0.276* * 0.198* 1.630
Iron 0.210" 0.012 0.087 0.157 0.129" 2.680“

 

Note: *, **, and *** denote signiﬁcance at the l %, 5 %, or 10 % level, respectively.
L = August = lean season, H = November = harvest, PH = February = post-harvest and P
= May = planting.

Behrman and Deolalikar (1987) found that income does not have a statistically
signiﬁcant effect on the consumption of protein, calcium, and iron for households in rural
South India. Bouis and Novenario-Reese (1997) found that iron consumption, in rural
Bangladesh, is responsive to changes in income but that of vitamin A is not. They
attributed this ﬁnding to the fact that vitamin A is available in very speciﬁc foods such as
vegetables while, iron can be found in many staple grains.

Second, the results indicate that all the nutrient-income elasticities are positive
and less than 1, indicating that increments in Bamako households’ real incomes will have
a positive, but less than proportionate, impact on household demand for nutrients in any
given season and for the pooled data. The pooled data results show that a 10 percent
growth in real incomes will increase the demand for protein (+1.91%), calcium (+1.98%),
vitamin A (+7.21%) and iron (+1.29%). These results are consistent with those of Pitt
(1983), whose study in rural Bangladesh shows that increases in households’ incomes
result in less than proportionate increments in the consumption of all nutrients.
Furthermore, the pooled data results indicate that the income elasticities of protein and
micronutrients are substantially lower than those of calories from staples (in Table 4-2),
suggesting that households will increase their consumption of foods that contain essential

nutrients more rapidly than that of staple foods, as their real income increases. These

165

ﬁndings remain consistent even when the analysis is broken down by season. For
instance, during the lean season, a 1 percent increase in real income will improve the
demand for calories from staples by 0.210 percent while that for vitamin A increases by
0.725 percent. The results clearly indicate that households are upgrading the quality of
their diets, substituting less expensive sources of calories for cheaper sources, as their
income increases.

Third, the results indicate that the income elasticities vary noticeably across the
range of nutrients (e.g., from 0.129 for iron to 0.721 for vitamin A for the pooled data)
and across seasons, especially for micronutrients (e.g., from 0.492 during the harvest
season to 0.725 during the lean season for vitamin A). The higher income elasticity of
demand for vitamin A during the post-harvest season (0.680) can be partly explained by
the low availability (higher prices) of spinach and green leaves during the cool dry
season, which corresponds to the growing season for most horticultural crops. In
addition, the results reveal that the income elasticities for calories (from 0.102 to 0.193)
vary less across seasons than those for vitamin A (from 0.492 to 0.725). These results
suggest that the adj ustrnents Bamako households make to their food baskets to maintain
calorie consumption more or less constant across seasons will have a greater impact on
the consumption of foods that contribute essential vitamins and minerals, such as calcium
and vitamin A, to urban households’ diets. This ﬁnding is further substantiated by the
Chow test results, which shows that the null hypothesis of stability in the estimated
income parameters across was rejected at least at the 5 percent level for all nutrients,

except calcium.

166

4.3.2. Nutrient-Price Elasticities

The effects of seasonal changes in relative prices on the demand for nutrients are
examined in order to determine if (1) Bamako households’ demand for nutrients is
responsive to changes in relative food prices, and (2) there is evidence of seasonal
changes in nutrient price responsiveness. These ﬁndings can be useful in designing for
food policies that aim at improving nutrient availability in Malian households.

4.3.2.1. Rice Price Effects on the Demand for Nutrients

Table 4-4, below, presents the demand for various nutrients with respect to the price of
rice for each season and for the pooled data and the Chow test results. The pooled data
results indicate that, on average annually, the price of rice has no statistically signiﬁcant
effect on the demand for any of the nutrients considered. This is quite surprising because,
as shown in the ﬁrst essay of this study, rice contributed on average in 2000-2001, 39
percent of the total calories available for consumption at home, 28 percent of total protein
availability, 13 percent of calcium, and 19 percent of iron. The elasticity of total calories
with respect to the price of rice computed by Rogers and Lowdermilk (1991) was also not
statistically signiﬁcant. They attributed this ﬁnding to urban households being able to
ﬁnd ways to preserve their calorie consumption through substitutions between rice and
other foods. However, they indicated that further work was needed to identify which
foods substitute for rice in the face of higher rice prices. The results of the previous essay
of this study indicated that (1) rice and millet-sorghum are net substitutes, once the
income effects are removed in all seasons; (2) the price of rice has a positive effect, both

compensated and uncompensated (except during the lean season), on the consumption of

167

maize and; (3) the price of rice has a statistically signiﬁcant large uncompensated effect
on the consumption of wheat during the harvest season.

Table 4-4: Elasticity of Demand for Nutrients With Respect to the Price of Rice

 

 

Nutrient L H PH P Pooled Chow Test
Total Calories -l.260** 0.032 0.383 -0.396 -0.245 1.320
Calories from staples -0.965*** -0.157 0.359 -0.780 -0.316 0.840
Calories ﬁom other foods -1 .989* 0.507 0.303 1.027 -0.025 1.570
Protein -1.246"‘ * 0.483 0.266 -0.430 -0.083 1.180
Vitamin A -l .159 -0.l90 2.766 0.329 0.324 0.680
Calcium -0.927 1.269" 0.136 -0.121 0.292 1.090
Iron -1.830* 0.399 0.354 -1.023 -0.203 1990*"

 

Note: *, **, and **"' denote signiﬁcance at the 1 %, 5 %, or 10 % level, respectively.
L = August = lean season, H = November = harvest, PH = February = post-harvest and P
= May = planting.

When separate regressions are run by season, the results indicate that the price of
rice has a statistical signiﬁcant impact on the demand for all nutrients, except calcium
and vitamin A during the lean season and calcium during the harvest season. For
instance, during the lean season, a 1 percent increase in the price of rice reduces the daily
availability of all nutrients: total calorie availability by 1.260 percent, calories from
staples by 0.965 percent, calories from other foods by 1.989 percent, protein by 1.246
percent, and iron by 1.830 percent. The amounts of calories obtained from other foods are
nearly twice as responsive (-1.989) to changes in the price of rice than calories obtained
ﬁom staples (-0.965) during the lean season. These results suggest that increases in the
price of rice will substantially reduce the demand for nutrients during the lean season,
when grain availability is relatively low in urban markets.
4.3.2.2. Millet-Sorghum Price Effects on the Demand for Nutrients

Table 4-5, below, presents the elasticity of demand for nutrients with respect to the price

of millet-sorghum for each season and for the pooled data. First, the pooled data results

168

indicate that, on average annually, the price of millet-sorghum has a negative and
statistically signiﬁcant impact on total calorie availability (-0.l74), calories ﬁ'om staples
(-0.279), iron (-0.240), and calcium (-0.157). Rogers and Lowdermilk (1991) also found
that the demand for calories is responsive to changes in the price of millet-sorghum. The
calorie-price elasticity for millet-sorghum derived from the pooled data (-0. 174) is
slightly lower than that of Rogers and Lowdermilk (1991) of —0.236, suggesting that the
effect of changes in the price of millet-sorghum on the demand for calories has decreased
over-time.

Table 4-5: Elasticity of Demand for Nutrients With Respect to the Price of Millet-

 

 

Mhum
Nutrient I L H PH P Pooled Chow Test

Total Calories -0.359" -0.008 -0.402"“" -0.228 -0. 174‘ 3.360“
Calories ﬁ'orn staples -0.524"' -0. 14 -0.504"' -0.373 -0.279‘ 4.400‘
Calories from other foods -0.056 0314”" -0.112 0.147 0.083 1.69
Protein -0.378" 0.012 -0.3 17*" -0.242 -0. 181 2.430”
Vitamin A 0.369 0.11 -0.916 0.709 -0.081 0.67
Calcium -0.368 -0.062 -0.318 0.022 -0.157"' 1.22
Iron -0.730* -0.015 -0.621" 0.01 -0.240"'" 4.460“

 

Note: *, ", and *** denote signiﬁcance at the 1 %, 5 %, or 10 % level, respectively.
L = August = lean season, H = November = harvest, PH = February = post-harvest and P
= May = planting.

Second, the seasonal, statistically signiﬁcant, results indicate that the nutrient
elasticities with respect to the price of millet-sorghum are all negative, except for calories
obtained from other foods during the harvest season, and are all less than 1. These results
suggest that a 1 percent change in price of millet-sorghum has a negative and less than
proportionate effect on the effective demand for nutrients in any given season. Once the

price effect of millet-sorghum on total calorie availability is decomposed by speciﬁc

foods, the results reveal that the amounts of calories obtained from staples are far more

169

responsive to changes in the price of millet-sorghum than calories obtained from other
foods in all seasons.

Third, evidence of cross-price effects is present during the harvest season,
suggesting that changes in millet-sorghum prices will cause changes in the mix of foods
purchased. During the harvest season, when grain prices are generally low, a 1 percent
increase in the price of millet-sorghum will increase the availability of calories from
other foods by 0.314 percent. This “perverse” price effect can be explained by the fact
that changes in the price of millet-sorghmn have strong income effects, as expenditures
on millet-sorghum occupy on average 22 percent of households’ staple budget. The
positive elasticity of demand for calories from other foods with respect to the price of
millet-sorghum suggests that households substitute between commodities within this food
group in face of higher millet-sorghum prices by switching from high cost calorie sources
to low cost calorie sources. For instance, households substitute dry ﬁsh for beef during
the harvest season. Hence, the reallocation of households’ budget induced by higher
millet-sorghum prices result in increased calorie availability from other foods through an
increase in the consumption of foods that are cheap sources of calories.
4.3.2.3. Beef Price Effects on the Demand for Nutrients
Table 4-6, below, presents the nutrient price elasticity of demand for calories, protein,
calcium, vitamin A, and iron with respect to the price of beef for each season and for the
pooled data. The pooled data results show that the price of beef has a positive statistical
signiﬁcant effect on the amounts of calories obtained from other foods. A 1 percent
increase in the price of beef is expected to increase the availability of calories from other

foods by 0.181 percent, on average, annually. This is due to the fact that households’

170

allocate on average 61.7 percent of their meat and ﬁsh expenditures on beef and thus,
changes in beef prices have strong income effects. Households tend to substitute between
dry ﬁsh and beef. The null hypothesis of stability in the estimated parameters across
seasons was rejected, at the 10 percent level, for calories from other foods, suggesting
that there is a statistically signiﬁcant shift across seasons in the response of the amounts
of calories obtained from other foods to changes in the price of beef. The seasonal results
show that the price of beef has a positive statistically signiﬁcant effect, at the 5 percent
level, on the amounts of calories obtained from other foods during the lean season
(0.287). The price of beef has no statistically signiﬁcant effect on the availability of any
of the nutrients considered in all other seasons.

Table 4-6: Elasticity of Demand for Nutrients With Respect to the Price of Beef

 

 

Nutrient L H PH P Pooled Chow Test
Total Calories 0.039 0.073 0.120 0.031 0.045 1.300
Calories from staples -0.059 0.043 0.034 0.021 -0.015 0.550
Calories from other foods 0.287" 0.100 0.275 0.061 0.181" 2.210“
Protein -0.039 0.026 0.097 -0.037 -0.010 1.150
Vitamin A 0.414 0.162 -0.276 -0.203 0.026 0.330
Calcium 0.070 -0.058 0.059 -0.096 0.014 0.830
Iron -0.150 -0.052 -0.112 -0.116 -0.1 14 1.520

 

Note: *, **, and *** denote signiﬁcance at the 1 %, 5 %, or 10 % level, respectively.

L = August = lean season, H = November = harvest, PH = February = post-harvest and P
= May = planting.

4.3.2.4. Dry Fish Price Effects on Nutrient Availability

Table 4-8, below, presents the elasticities of demand for protein, calcium, vitamin A, and

iron with respect to the price of dried ﬁsh for each season and for the pooled data.43 First,

the pooled data results indicate that the price of dry ﬁsh has a negative and statistically

 

‘3 As mentioned in the ﬁrst essay of this study, in 2000-2001, dry ﬁsh contributed on average 1 percent of
the total calories available for consumption, 7 percent of total protein availability, 21 percent of calcium,
and 2 percent of iron.

l7l

signiﬁcant effect on the demand for calcium. A 1 percent increase in the price of dry ﬁsh
reduces the demand for calcium by 0.293 percent. Second, the results indicate that the
price of dry ﬁsh has a statistically signiﬁcant negative impact on the demand for calcium
during the harvest season and on that of calories, protein, and calcium during the post-
harvest season. A 1 percent increase in the price of dry ﬁsh will reduce household
demand for calcium by 0.236 percent during the harvest season. During the post-harvest
season, when households’ real incomes are high as the price of most staples decrease, a 1
percent increase in the price of dry ﬁsh will reduce the demand for calories, protein, and
calcium by 0.179 percent, 0.230 percent, and 0.434 percent, respectively.

Table 4-7: Elasticity of Demand for Nutrients With Respect to the Price of Dry Fish

 

 

Nutrient L H PH P Pooled Chow Test
Total Calories -0.057 -0.001 -0.179** 0.017 -0.058 1880*"
Calories from staples -0.069 -0.008 -0.l37 0.100 -0.043 1.010
Calories from other foods -0.043 -0.002 -0.274 -0.153 -0.104 1940*”
Protein —0.067 -0.052 -0.230** 0.025 -0.110 2.100"
Vitamin A 0.195 0.060 -0.163 -0.208 0.042 0.150
Calcium -0.170 -0.236** -0.434** -0.224 -0.293* 2820*
Iron 0.131 -0.021 -0.l76 0.038 -0.020 1810*"

 

Note: *, **, and *** denote signiﬁcance at the 1 %, 5 %, or 10 % level, respectively.

L = August = lean season, H = November = harvest, PH = February = post-harvest and P
= May = planting.

4.3.2.5. Effects of the Price of Green Leaves on the demand for nutrients

Table 4-8, below, presents the elasticities of demand for nutrients with respect to the
price of green leaves for each season and for the pooled data.44 First, the pooled data

results indicate that the price of green leaves has a negative statistically signiﬁcant impact

on calcium and vitamin A availability as, a 1 percent increase in the price of green leaves

 

“ As shown in the ﬁrst essay of this study, in 2000-2001, green leaves contributed on average annually
0.50 percent of the total calories available for consumption, 1.20 percent of total protein availability, 16
percent of calcium, 17 percent of vitamin A and 4 percent of iron.

172

will reduce calcium and vitamin A availability by 0.144 percent and 0.262 percent,
respectively.

Second, the seasonal estimates indicate that the price of green leaves has a
negative, less than proportionate, statistically signiﬁcant impact on the demand for at
least one nutrient in any given season, except the post-harvest season. During the lean
season, a 1 percent increase in the price of green leaves will reduce the demand for total
calories by 0.146 percent, calories from staples by 0.146 percent, protein by 0.123
percent, and calcium by 0.223. During the harvest season, a 1 percent increase in the
price of green leaves will reduce calcium availability by 0.223 percent. During the
planting season, a 1 percent increase in the price of green leaves is predicted to decrease
the availability of calories from other foods by 0.208 percent.

Table 4-8: Elasticity of Demand for Nutrients With Respect to the Price of Green
Leaves

 

 

Nutrient L H PH P Pooled Chow
Test

Total Calories -0.146** -0.093 0.028 -0.030 -0.027 1.650
Calories from staples -0.146** -0.112 0.008 -0.015 -0.039 1.080
Calories from other foods -0.164 -0.054 0.041 -O.208*** -0.047 1820*
Protein -0.123 * "' "' -0.122 -0.051 -0.006 -0.054 1.600
Vitamin A -0.191 -0.248 -0.183 -0.498 -0.262*"‘ 1.040
Calcium -0.223*** -0.267** -0.148 0.028 -0.144* 2.140"
Iron -0.129 -0.137 -0.023 -0.032 -0.047 1.450

 

Note: *, **, and *” denote signiﬁcance at the 1 %, 5 %, or 10 % level, respectively.

L = August = lean season, H = November = harvest, PH = February = post-harvest and P
= May = planting.

4.3.3. Sensitivity Analyses

In this section, the estimated nutrient-income parameters are used to simulate the impact

of changes in households’ real incomes on the demand for nutrients by season and for the

pooled data. The analysis is further disaggregated to take into account the effects of

173

including estimates of nutrient availability from away-from-home foods on average daily
nutrient availability per adult equivalent. Detailed information on how the estimates of
nutrient availability from away-from-home foods were computed is presented in Essay 1.
Table 4-9, below, presents the baseline results by season and for the pooled data and by
income group.

Table 4-9: Baseline Values

 

 

 

 

 

 

 

 

 

 

Nutrients from at-home foods I Total Amounts of Nutrients
Nutrient L I H I PH rP I PooledI L I H I PH I P I Pooled
Total Calories
Low 2101 2122 2092 2015 2083 2290 2313 2279 2196 2270
Middle 2101 2060 2134 1912 2052 2335 2289 2372 2125 2280
High 2588 2529 2530 2335 2496 2806 2742 2743 2532 2706
Mean 2263 2237 2252 2088 2210 2477 2448 2465 2285 2419
Calories From Staples
Low 1426 1564 1571 1550 1528 1555 1704 1712 1690 1665
Middle 1500 1444 1454 1360 1440 1667 1605 1616 1512 1600
High 1674 1689 1661 1626 1663 1815 1832 1801 1763 1803
Mean 1533 1566 1562 1512 1543 1679 1714 1710 1655 1689
Calories From Others
Low 675 558 520 465 555 735 609 567 507 604
Middle 602 616 680 552 612 669 684 756 613 680
High 914 840 869 709 833 991 91 1 942 769 903
Mean 730 67 1 690 575 667 798 734 755 630 729
Protein
Low 56 56 56 51 55 61 61 61 55 60
Middle 55 51 51 46 51 61 57 57 52 57
High 71 69 67 59 67 77 74 73 64 72
Mean 61 59 58 52 57 66 64 64 57 63
Calcium
Low 484 384 404 349 405 528 418 440 381 442
Middle 432 382 336 31 1 365 480 425 374 346 406
High 584 477 574 445 520 633 518 622 483 564
Mean 500 414 438 369 430 547 453 479 403 470
Vitamin A
Low 324 257 276 169 257 353 280 301 184 280
Middle 453 354 260 258 33 l 504 394 288 286 368
High 550 484 672 487 548 597 525 728 529 595
Mean 443 365 402 305 379 485 400 439 333 414
Iron
Low 22 23 23 21 22 24 25 25 23 24
Middle 21 21 19 19 20 24 23 21 21 22
High 27 26 24 25 25 29 28 26 27 28
Mean 23 23 22 22 23 26 25 24 24 25

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting. Total amounts of nutrients available = Nutrients from at-home

174

foods + Nutrients from away-from-home foods. Calories from away-from-home foods
can’t be disaggregated into calories from staples and other foods.

The impact of changes in real incomes is simulated using the following equation
to compute the expected change in the demand for nutrients:
A Na: AX: * ler (2)
where, A N.“ is the percentage change in the availability of Nutrient k at time t,
AX is the percentage change in real income at time t,
m, is the income elasticity of demand for nutrient k at time t; and
Table 4-10, below, presents the effects of a 20 percent increase in real incomes on
household demand for nutrients in Bamako by season and for the pooled data, holding all
other factors ﬁxed. The increased demand for food induced by the increments in income,
in face of constant food prices, would imply that food availability in urban households

would need to be increased (i.e., through greater production or imports, or reduced

 

 

exports).

Table 4-10: Effect of a 20 Percent Increase in Real Incomes in Percentage Changes
Nutrient L H PH P Pooled

Total Calories 3.9 2.0 3.4 3.3 3.2

Calories From Staples 2.7 -0.7 1.6 2.3 1.4

Calories From Other Foods 6.7 8.0 6.5 6.2 7.3

Protein 4.2 1.8 3.8 4.3 3.8

Calcium 2.6 1.9 3.2 5.5 4.0

Vitamin A 14.5 9.8 13.6 11.9 14.4

Iron 4.2 0.2 1.7 3.1 2.6

 

Note: L = August = lean season, H = November = harvest, PH = February = post-harvest
and P = May = planting.

The pooled data results show that the 20 percent increase in real incomes would
improve average calorie availability per adult equivalent by 3.2 percent, which is rather

small. However, once the effects are disaggregated by source of calories, the results

175

indicate that the impact of a 20 percent increase in real incomes would be larger on the
amounts of calories obtained from other foods (7.3 percent) than on those obtained from
staples (1.4 percent). Hence, the increase in real incomes would likely result in improved
diet diversity in Bamako households, as the amount of vitamin A available for
consumption, which is mainly supplied by vegetables, would increase by 14.4 percent.

Table 4-11, below, presents the effects of the 20 percent increase in real incomes
on the amounts of nutrients available by season and by income group. The pooled data
results show that the 20 percent increase in real income would push average at-home
calorie availability above the recommended dietary allowance (RDA) of 2200 kcal.
However, the increase in the amounts of protein, calcium, vitamin A, and iron availability
induced by the change in real income would not be enough to meet the RDA of 63 grams,
1000 milligrams, 600 micrograms, and 59 milligrams, respectively. Furthermore, the
seasonally pooled data results indicate that only households in the high-income group are
able to meet the RDA for calories, protein, and vitamin A. However, once the amounts
of nutrients available from away-ﬁom-home foods are taken into account, the results
indicate that all income groups would be able to meet the minimum calorie requirements.
However, the low and middle-income groups’ consumption of protein, calcium, vitamin
A, and iron would remain below the recommended levels. High-income households

would be able to meet the RDA for protein and vitamin A.

176

Table 4-11: Effect of a 20 Percent Increase in Real Incomes on the Amounts of
Nutrients Available by Season and By income Group.
Nutrients from at-home foods I Total Amounts of Nutrients

 

 

 

 

 

 

 

 

Nutrient L H PH P PooledI L I H | PH | P Pooled
Total Calories
Low 2182 2166 2163 2082 2150 2378 2360 2357 2269 2343
Middle 2182 2102 2207 1976 2118 2425 2336 2453 2196 2354
High 2688 2581 2616 2413 2576 2914 2798 2837 2616 2794
MP?! ............................. 2. 15.1..2.235?"9.32.9...213?.-..2.2.§.1-..2§Z3..3.4.?!3..2??2.3.3.§9..3§2?..
Calories From Staples
Low 1465 1553 1596 1586 1549 1597 1693 1740 1728 1689
Middle 1540 1434 1478 1391 1460 1712 1594 1642 1546 1622
High 1720 1678 1688 1663 1686 1865 1819 1830 1804 1828
Mean ............................. 1.?15...!.5.5.§...1.5.§.7...!.5.4.7.."1.5.9.5...1.7.2.5..1.7.9.2...1??l..l§2§...l.7.1.§..
Calories From Others-
Low 720 603 554 494 595 785 657 604 538 648
Middle 642 665 724 586 657 714 739 805 651 730
High 975 907 926 753 893 1057 984 1004 816 969
Man-779725735611 ..... 7.1.5.....§€?....??§.-..§.Q‘!....§§? ..... 7. M...
Protein
Low 58 57 58 53 57 64 63 64 57 62
Middle 57 52 53 48 53 63 58 59 54 59
High 74 70 70 62 69 80 76 76 67 75
Mean .............................. 99......62 ..... 9 9 ..... 5.4. ...... 6. 9 ...... 92.....62 ..... M......5.2......6.5....
Calcium
Low 497 391 417 369 421 541 426 454 402 459
Middle 443 390 347 329 380 492 433 386 365 422
High 599 487 592 470 541 649 528 642 510 586
Mm513422452389 ..... 2M .7.....29.1...3.6.2....:124....42.5.....4M...
VitaminA
Low 371 283 313 189 294 405 308 342 206 320
Middle 519 389 295 289 379 577 432 328 321 421
ngh 630 532 763 546 627 683 577 827 592 680
Mean507401457341 ..... 49.2....252...432...9.92....3.z.3....AM...
Iron
Low 23 23 24 22 23 25 25 26 24 25
Middle 22 21 19 20 20 25 23 21 22 23
High 28 26 24 26 26 30 28 27 28 28
Mean 24 23 22 22 23 27 25 24 25 25

 

Note: Nutrients are expressed in mean daily availability per adult equivalent; Calories are

expressed in kilo-calories/AE/day; Protein in Grams/AE/day; Vitamin A in

Micrograms/AE/day; and Calcium and Iron are in Milligrams/AE/day. L = August = lean

season, H = November = harvest, PH = February = post-harvest and P = May = planting.
The seasonal results show that the impact of such increments in real incomes

Would push average at-home calorie availability above the recommended dietary

177

allowance (RDA) of 2200 kcal in all seasons except the planting season. However, these
results mask the fact that only the households in the high-income group would be able to
satisfy the minimum calorie requirements in all seasons. Average calorie availability
would remain below the requirement levels for low and middle-income households in all
season, except the post-harvest for the middle-income group. High-income households
would be able to satisfy the RDA for protein in all seasons, except the planting, and for
vitamin A during the lean and post-harvest seasons. Households in the low and middle-
income groups would be unable to meet the RDA for protein, calcium, vitamin A, and
iron in all seasons.

Once the amounts of nutrients available from away-ﬁ'om-home foods are taken
into account, the results indicate that all households, on average, would be able to meet
the RDA for calories in all seasons, except the planting for the middle-income group.
Low and high-income households would satisfy the RDA for protein in all seasons,
except the planting for the low-income group. However, the amounts of iron, calcium,
and vitamin A, with the exception of the high-income group in the lean and post-harvest
seasons, available for household consumption would still remain below the recommended
levels in all seasons. It should be remembered, however, that these ﬁgures are upper-end
estimates, as they make no allowance for nutrient wastage or loss during food

preparation.

4.4. Conclusion
In this essay, the relationship between real income, relative prices, and households’
demand for nutrients in Bamako, Mali, was examined by season and annually using

Engel functions. First, the results indicate that the price of major food commodities, real

178

income, and household size factors account for a substantial part of the observed
variation in the amounts of nutrients available for household consumption at any given
season.

Second, the study ﬁnds that Bamako households’ demand for nutrients are
responsive to changes in their real incomes and relative prices and that the magnitude of
the nutrient income and price elasticities will change from one season to another. The
null hypothesis of stability in the nutrient income parameters across seasons was rejected
for all the estimated coefﬁcients, except for calcium, suggesting that there is a
statistically signiﬁcant shift in the estimated nutrient-income elasticities across seasons.
Moreover, the Chow test results indicate a certain degree of non-constancy of many price
parameters across seasons, implying that the impact of a uniform food policy on the
quantity and quality of food available in Bamako households will vary by season.

Third, the results indicate that improvements in Bamako households’ real incomes
will have a positive, but less than proportionate, impact on household demand for
nutrients in any given season and for the pooled data. More speciﬁcally, the results
indicate that increases in households’ real incomes will have a positive impact on calorie
purchases but that households will increase their consumption of other foods more
rapidly than that of staples as their real incomes increases. The results clearly indicate
that households are upgrading the quality of their diets, substituting less expensive
sources of calories for cheaper sources, as their income increases. In addition, the
seasonal results suggest that Bamako households try to maintain calorie consumption
more or less constant across seasons at the expense of foods that contribute essential

vitamins and minerals, such as calcium and vitamin A, to households’ diets.

179

Fourth, the results on the estimated nutrient price elasticities indicate that (1) the
price of rice and beef have no statistically signiﬁcant effect on the availability of any of
the nutrients considered in all seasons, except the lean season, and for the pooled data; (2)
the price of millet sorghum has a statistically signiﬁcantly negative and less than
proportionate effect on the effective demand for nutrients in any given season; (3) the
price of dry ﬁsh has a negative and statistical signiﬁcant effect on the amounts of calcium
demanded by households; and (4) the price of green leaves has a negative statistically
signiﬁcant impact on the demand for calcium and vitamin A.

F ifth, the sensitivity analysis revealed that increases in real incomes would
improve average calorie availability but the effects of such increments would be larger on
the amounts of calories obtained from other foods than on those obtained from staples.
Furthermore, the sensitivity analysis suggests that, once the availability of nutrients from
away-from-home foods is taken into account, households are able to meet minimum
calorie requirements in all seasons. However, the results indicate that households need to
achieve substantial income gains in order to be able to meet the RDAs for protein,
calcium, vitamin A, and iron in all seasons and annually. The results also showed that
substantial variability remains among and probably within households, suggesting that
improvements in income alone may not be enough to reduce malnutrition in Bamako
households.

The ﬁndings of this essay have several implications for policy design in Mali.
First, the positive nutrient-income elasticities imply that increasing households’ real
incomes will improve the quantity (i.e., calories) and the quality (i.e., protein, minerals,

and vitamins) of food available in those households and thereby will be an effective

180

mechanism in reducing malnutrition. Hence, the policies that aim at increasing
households’ real incomes will also improve their nutrition. Better nutrition outcomes
can, in turn, translate into improved worker productivity (Straus and Thomas, 1998).

Second, the fact that households will increase their consumption of other foods
more rapidly than that of staples as their real incomes increases suggests that households
will diversify their diets as their income grows. This ﬁnding suggests that greater
allocation of resources and investment in the production and marketing of horticultural
commodities offer the potential to substantially reduce malnutrition, especially vitamin A
deﬁciency, increase employment, and reduce poverty in urban areas.

Third, the ﬁnding that the demand for nutrients are responsive to changes in the
price of millet-sorghum, dry ﬁsh, and green leaves suggests that food prices can be used
as policy instruments to reduce malnutrition in households. Increased investments (public
and private) in the production and marketing of horticultural commodities can yield the
productivity gains necessary to substantially reduce the price of horticultural goods so
that low-income households can readily access these foods.

Finally, the ﬁnding that many of the estimated nutrient income and price
parameters are not stable across seasons imply that the effectiveness of food policies will
be contingent upon whether or not they are systematically synchronized with the short-
run response of households’ consumption patterns to income and price changes. The
results suggest that the impacts of a uniform food policy on the quantity and quality of

food available in Bamako households will vary by season.

181

REFERENCES

Behrman J .R. and AB. Deolalikar, (1987). Will Developing Country Nutrition Improve
with Income? A Case Study for Rural South India. Journal of Political Economy, 95,
pp.492-507.

Behrman, J ., A. Deolalikar, and BL. Wolfe, (1988). Nutrients: Impacts and
Determinants. The World Bank Economic Review, 2.

Behrman, J. R. (1995). Household Behavior and Micronutrients: What We Know and
What We Don't Know? International Food Policy Research Institute, Working Papers on
Agricultural Strategies for Micronutrients, No. 2.

Behrman, J .R. and Deolalikar, AB. (1987). Will Developing Country Nutrition hnprove
with Income? A Case Study of Rural South India Journal of Political Economy, 95, 108-
38.

Bouis, HE. (1994). The Effect of Income on Demand for Food in Poor Countries: Are
Our Databases Giving Us Reliable Estimates? Journal of Development Economics, 44(1),
pp.199-226.

Bouis, HE. and Novenaio-Reese, M.J.G. (1997). The Determinants of Demand for
Micronutrients: An Analysis of Rural Households in Bangladesh. International Food
Policy Research Institute, Food Consumption and Nutrition Division Discussion Paper

No. 32.

Dawson J .P., and R. Tifﬁn, (1998). Estimating the Demand for Calories in India, Amer. J.
Agr. Econ. 80 (August 1998): 474-481.

World Bank. 1981. World development report 1981. Oxford: Oxford University Press for
the World Bank.

Deaton, A. The Analysis of Household Surveys: A Microeconomic Approach to
Development Policy. World Bank. The John Hopkins University Press. Baltimore,
London: 1997.

Bhargava, A., (1991). Estimating Short and Long Run Income Elasticity of Foods and
Nutrients for Rural South India. Journal of the Royal Statistical Society, A, 154, 157-74.

Bouis, H. and Haddad, L.J. (1992). Are Estimates of Calorie-Income Elasticities too

High? A Recalibration of the Plausible Range. Journal of Development Economics, 39,
333-364.

182

Deaton, Angus (1997). The Analysis of Household Surveys: A Microeconometric
Approach to Development Policy, World Bank, The Johns Hopkins University Press,
Baltimore and London.

Food and Agriculture Organization of the United Nations and International Life Sciences
Institute, 1997. Preventing micronutrient malnutrition: A guide to food—based approaches
- A manual for policy makers and programme planners. ILSI PRESS. Library of
Congress Catalog Number: 97-71096.

Krishnaswamy K, K. Vijayraghavan, J . G. Sastry, H. Rao, G.N.V. Brahman, G. Radhiah,
K. Kashinath, V. Rao (1997). 25 years of National Nutritional Monitoring. Board Report
of the National Institute of Nutrition, Indian Council of Medical Research, Hyderabad

Lipton, Michael (1983). Poverty, Undernutrition, and Hunger. World Bank Staff
Working Paper Number 597, The World Bank, Washington, DC, USA.

Ravallion, M., (1990). Economic Effects on Undemutrition. Economic Development and
Cultural Change, 38, No. 3, 489-515.

Strauss, J. and Duncan Thomas (1990). The Shape of the Calorie Expenditure Curve.
Economic Growth Centre Discussion Paper No. 595, Yale University.

Subramanian, S. and Angus Deaton (1996). The Demand for Food and Calories. Journal
of Political Economy, Vol. 104, No. 1.

Ullah, A. and N. Roy (1998). Nonparametric and Semiparametric Econometrics of Panel

Data. In A.Ullah and D.E.A. Giles (eds.), Handbook in Applied Economic Statistics,
Marcel Dekker, Ch. 17, 579-604.

183

APPENDIX 4

184

Table A4-l: Summary Statistics

 

 

Variables Lean Harvest Post-Harvest Planting Pooled
Mean SD Mean SD Mean SD Mean SD Mean SD
Total Calories 2259 689 2233 658 2249 713 2083 614 2206 675
Calories From Staples 1533 404 1563 402 1559 353 1509 409 1541 390
Calories From Other Foods 727 285 670 256 690 360 575 205 665 285
Protein 60 15 58 15 58 16 52 13 57 15
Calcium 498 182 414 151 435 252 367 141 429 181
Vitamin A 443 298 365 315 399 360 304 364 378 334
Iron 23 7 23 7 22 7 22 7 23 7
Rice price 269 21 275 27 260 22 264 16 267 21
Millet-Sorghum price 147 40 146 43 136 34 180 47 152 41
Beef price 980 294 101 1 274 959 258 945 230 974 264
Dry ﬁsh price 1308 760 1482 571 1392 705 1427 386 1402 605
Green leaves price 299 206 307 174 518 1074 348 353 368 452
Calorie Price 111 85 108 53 116 67 103 60 109 66
Household sizeinAE 13 7 13 7 13 8 13 8 13 8
Total expenditure/AB 9521 6086 7155 3172 6818 3898 6076 4287 7392 4633

 

Note: Nutrients are expressed in mean daily availability per adult equivalent; Prices are in
CFA Francs per kilogram. Calories are expressed in kilo-calories per adult equivalent
(AE) per day; Protein in Grams/AE/day; Vitamin A in Micrograms/AE/day; and Calcium
and Iron are in Milligrams/AE/day. Prices are expressed in CFA Francs per kilogram.
Total expenditures are expressed in CF A Francs per adult equivalent per week.

185

Table A4-2: Nutrients Contributed by Major Food Groups (%) by Season

 

 

Seasons Calories Protein Calcium Vit A Iron
Lean
Rice 39.2 27.6 13.3 0.0 18.9
Other Staples 28.7 30.3 11.2 3.7 44.6
Meat and Fish 4.8 19.9 22.8 3.4 7.0
Vegetables 4.5 8.9 40.7 50.7 20.8
Oil 8.3 0.0 0.0 36.4 0.0
Sugar 7.3 0.0 0.0 0.0 0.0
All others 7.1 13.4 12.0 5.8 8.8
Harvest
Rice 41.2 29.7 16.6 0.0 19.8
Other Staples 29.0 29.8 13.9 11.9 45.6
Meat and Fish 4.7 18.4 21.5 4.0 6.8
Vegetables 3.4 6.6 34.6 55.4 17.5
Oil 6.2 0.1 0.0 20.9 0.0
Sugar 7.2 0.0 0.0 0.0 0.0
All others 8.3 15.5 13.3 7.9 10.2
Post-Harvest
Rice 42.2 30.9 16.9 0.0 21.7
Other Staples 27.4 28.7 12.1 2.1 43.8
Meat and Fish 4.5 18.0 22.8 2.6 7.1
Vegetables 3.8 7.3 31.5 73.4 18.0
Oil 7.1 0.1 0.0 17.2 0.0
Sugar 7.2 0.0 0.0 0.0 0.0
All others 7.8 15.0 16.7 4.7 9.4
Planting
Rice 41.6 31.4 17.8 0.0 19.9
Other Staples 31.0 32.2 13.8 3.1 48.4
Meat and Fish 3.9 16.3 20.3 3.5 5.8
Vegetables 3.3 6.2 34.5 62.9 17.7
Oil 5.5 0.0 0.0 27.9 0.0
Sugar 7.8 0.0 0.0 0.0 0.0
All others 7.0 13.9 13.6 2.5 8.3
Average
Rice 41.1 29.9 16.2 0.0 20.1
Other Staples 29.0 30.2 12.8 5.2 45.6
Meat and Fish 4.5 18.1 21.8 3.4 6.7
Vegetables 3.7 7.2 35.3 60.6 18.5
Oil 6.8 0.1 0.0 25.6 0.0
Sugar 7.4 0.0 0.0 0.0 0.0
All others 7.6 14.4 13.9 5.2 9.2

 

186

Table A4-3: Nutrients Contributed by Speciﬁc Food Items (%) by Season

 

 

Commodities Calories Protein

Aug Nov Feb May Avg Aug Nov Feb May Avg

__Shues ‘ '— "'
Rice 38.96 40.97 42.13 41.49 40.89 27.33 29.38 30.67 31.23 29.66
Millet-Sorghum 23.58 24.75 22.80 26.54 24.42 24.80 25.62 24.07 27.24 25.43
Maize 3.35 2.42 2.50 2.82 2.77 3.42 2.50 2.50 3.01 2.86
Wheat 1.52 0.94 1.28 1.17 1.23 1.82 1.17 1.62 1.52 1.53
Other Cereal 0.15 0.00 0.08 0.19 0.11 0.12 0.00 0.06 0.16 0.09
Atieke 0.04 0.01 0.03 0.05 0.03 0.01 0.00 0.01 0.02 0.01
Cassava 0.06 0.25 0.07 0.02 0.10 0.04 0.16 0.05 0.01 0.07
Potato 0.07 0.02 0.43 0.11 0.16 0.06 0.02 0.38 0.10 0.14
Sweet Potato 0.10 0.64 0.02 0.04 0.20 0.06 0.36 0.01 0.02 0.11

Meat and Fish
Beef 3.01 3.34 2.98 2.77 3.03 8.62 9.78 8.84 8.51 8.94
Mutton 0.11 0.00 0.36 0.03 0.13 0.28 0.00 0.93 0.09 0.32
Poultry 0.08 0.01 0.05 0.02 0.04 0.40 0.05 0.27 0.12 0.21
Dry Fish 1.27 1.00 1.10 0.89 1.07 8.34 6.72 7.52 6.28 7.22
Fresh Fish 0.44 0.37 0.11 0.23 0.29 2.50 2.13 0.65 1.42 1.68
Vegetables
Leaves 0.76 0.49 0.37 0.38 0.50 1.73 1.22 0.88 0.97 1.20
Okra 0.76 0.51 0.45 0.56 0.57 1.42 0.79 0.67 0.97 0.96
Onion 0.58 0.61 0.87 0.77 0.71 0.83 0.90 1.24 1.13 1.03
Tomato 0.25 0.35 0.39 0.39 0.34 0.43 0.60 0.68 0.71 0.61
Other Vegetable: Fresh 0.41 0. 46 0.61 0. 34 0. 45 0. 56 0.71 1 .09 0. 53 0.72
All Other Vetegable 1. 72 1.03 1.08 0. 87 1.17 3. 96 2.44 2. 65 1 9.1 2. 74
Oil
Peanut Oil 6.19 4.14 5.23 3.89 4.86 0.00 0.00 0.00 0.00 0. 00
Palm Oil 1.08 0.51 0.36 0.64 0.65 0.00 0.00 0.00 0.00 0. 00
Sheanut Oil 1 .06 1.65 1. 52 0. 94 1.29 0. 05 0. 07 0. 07 0. 04 0. 06
Sugar 7. 37 7. 22 '7. 29 7. 83 7. 43 0. 00 0. 00 0. 00 0. 00
All others

Butter 0.04 0.04 0.00 0.03 0.03 0.00 0.00 0.00 0.00 0.00
Buttermilk 0.12 0.06 0.10 0.08 0.09 0.25 0.13 0.21 0.18 0.19
Fresh Milk 0.01 0.01 0.00 0.00 0.00 0.01 0.02 0.01 0.00 0.01
Condensed Sweetened Milk 0.12 0.01 0.00 0.00 0.03 0.11 0.01 0.00 0.00 0.03
Powdered Milk 0.52 0.46 1.05 0.48 0.63 1.02 0.94 2.16 1 02 1.28
Eggs 0.15 0.02 0.01 0.00 0.05 0.48 0.06 0.04 0.00 0.15
Peanuts 5.08 6.66 5.61 5.46 5.70 8.50 11.39 9.73 9.83 9.86
Seeds 0.46 0.49 0.49 0.47 0.48 1.24 1.32 1.30 1.25 1.28
Other Legume Nut and Seed 0.00 0.01 0.00 0.00 0.00 0.01 0.01 0.01 0.00 0.01
Coffee 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Tea Lipton 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Green Tea 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0 00 0.00
Quinqueliba 0.01 0.01 0.02 0.01 0.01 0.00 0.00 0.00 0.00 0.00
Other Beverage 0.02 0.00 0.06 0.00 0.02 0.02 0.00 0.03 0.00 0.01
Banana 0.03 0.07 0.07 0.00 0.04 0.01 0.03 0.03 0.00 0.02
Citronella 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Dates 0.01 0.01 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00
Lemon 0.02 0.07 0.01 0.00 0.02 0.01 0.05 0.01 0.00 0.02
Raisin 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Tamarind 0. 06 0. 00 0. 08 0. 09 0. 06 0. 07 0. 00 0. 09 0.11 0.07
Orange 0. 01 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00
Seansonings and Spices 0.42 0. 40 0.39 0. 39 0.40 1. 50 1.44 1 .53 1. 61 1. 52
Sum 100 100 100 100 100 100 100 100 100 100

 

 

 

187

Table A4-3: Nutrients Contributed by Speciﬁc Food Items (%) by Season

 

 

 

 

 

 

(continued)
Commodities Iron Calcium
Aug Nov Feb May As; Aug Nov Feb May
Staples
Rice 18.75 19.67 21.49 19.80 19.93 13.18 16.50 16.23 17.55
Millet-Sorghum 38.94 40.88 40.00 43.79 40.90 9.35 11.14 9.88 11.66
Maize 4.28 2.90 2.48 3.13 3.20 0.55 0.51 0.61 0.66
Wheat 0.98 0.57 0.77 0.70 0.76 0.87 0.66 0.84 0.90
Other Cereal 0.15 0.00 0.08 0.18 0.10 0.05 0.00 0.03 0.08
Atieke 0.05 0.01 0.04 0.06 0.04 0.08 0.02 0.08 0.13
Cassava 0.04 0.17 0.05 0.01 0.07 0.12 0.63 0.18 0 05
Potato 0.11 0.03 0.65 0.16 0.23 0.06 0.02 0.38 0.11
Sweet Potato 0.18 1.12 0.04 0.07 0.35 0.14 1.04 0.04 0.07
Meat and Fish
Beef 4.47 4. 95 4 68 4.08 4.54 0.64 0.85 0.72 0.74
Mutton 0.08 0. 00 0. 29 0. 02 0.10 0.02 0.00 0.07 0. 01
Poultry 0.06 0. 01 0. 04 0. 02 0. 03 0.02 0.00 0.02 0. 01
Dry Fish 2.24 1.76 2.06 1.56 1.91 21.80 20.45 21.59 19.13
Fresh Fish 0.23 0.19 0.06 0.12 0.15 0.44 0.43 0.12 0.29
Vegetables
Leaves 5.44 3.63 2.52 2. 84 3.61 22.54 16.52 12.08 14.08
Okra 5.27 5.86 5.92 5.69 5.68 9.53 9.17 7.99 10.08
Onion 1.37 1.45 2.11 1. 75 1.67 2.00 2.50 3.36 3.28
Tomato 0.81 1.03 1.21 1.13 1.04 0.49 0.83 0.88 0.98
Other Vegetable: Fresh 0.81 1. 39 1. 57 0.69 1.11 0.94 1.76 3. 00 1.07
All Other Vetegable 7.14 4.16 4. 66 5. 85 5.45 5.29 3.65 3. 76 5.37
Oil
Peanut Oil 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Palm Oil 0.00 0.00 0.00 0.00 0.00 0.03 0.02 0. 01 0.02
Sheanut Oil 0.00 0. 00 0. 00 0. 00 0. 00 0 00 0.00 0. 00 0. 00
Sugar 0.00 0. 00 0. 00 0. 00 0. 00 0. 00 0.00 0. 00 0. 00
All others

Butter 0.00 0.00 0.00 0.00 0.00 0.01 0.01 0. 00 0.01
Buttermilk 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0. 00 0.00
Fresh Milk 0.00 0.00 0.00 0.00 0.00 0.05 0.08 0. 03 0.00
Condensed Sweetened Milk 0.01 0.00 0.00 0.00 0.00 0.50 0.03 0.00 0.00
Powdered Milk 0.05 0.04 0.11 0.05 0.06 4.30 4.59 9.97 5.02
Eggs 0.21 0.02 0.02 0.00 0.06 0.22 0.03 0.02 0.00
Peanuts 5.12 6.74 5.96 5.55 5.84 2.49 3.89 3.12 3. 37
Seeds 2.56 2.61 2.47 2.09 2.43 1.61 1.95 1.89 1. 89
Other Legume Nut and Seed 0.01 0.02 0.01 0.00 0.01 0.00 0.01 0.00 0.00
Coffee 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Tea Lipton 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Green Tea 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0. 00
Quinqueliba 0.00 0.01 0.01 0.00 0.01 0.00 0.00 0.00 0. 00
Other Beverage 0.01 0.00 0.02 0.00 0.01 0.07 0.00 0.10 0. 00
Banana 0.02 0.03 0.03 0.00 0.02 0.01 0.01 0.01 0. 00
Citronella 0.00 0.00 0.00 0.00 0.00 0. 00 0.00 0.00 0. 00
Dates 0.00 0.00 0.00 0.00 0.00 0. 01 0.01 0.01 0.00
Lemon 0.04 0.14 0.02 0.00 0.05 0.05 0.21 0. 03 0.01
Raisin 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0. 00 0.00
Tamarind 0.01 0. 00 0.00 0.00 0. 00 0.25 0.01 0.36 0.46
Orange 0.00 0. 00 0.00 0.00 0. 00 0.02 0.00 0.00 0.00
Seansonings and Spices 0.63 0.56 0. 65 0. 64 0.62 2. 27 2.50 2.60 3.00
Sum 100 100 100 100 100 100 100 100 100

 

188

Table A4-3: Nutrients Contributed by Speciﬁc Food Items (%) by Season

 

 

 

(continued)
Commodities Vitamin A
Augr Nov Feb May AvL—

Staples
Rice 0.00 0.00 0.00 0.00 0.00
Millet-Sorghum l .22 1 .37 1 .18 1.38 1 .29
Maize 0.74 0.52 0.10 0.51 0.47
Wheat 0.00 0.00 0.00 0.00 0.00
Other Cereal 0.00 0.00 0.00 0.00 0.00
Atieke 0.01 0.00 0.01 0.01 0.01
Cassava 0.01 0.06 0.02 0.00 0.02
Potato 0.03 0.01 0.20 0.07 0.08
Sweet Potato 1.41 10.68 0.36 0.72 3.29

Meat and Fish
Beef 1.62 2.16 1.77 2.01 1.89
Mutton 0.02 0.00 0.08 0.01 0.03
Poultry 0.23 0.04 0.17 0.08 0.13
Dry Fish 0.00 0.00 0.00 0.00 0.00
Fresh Fish 1.66 1.66 0.46 1.18 1.24
Vegetables
Leaves 21.29 18.34 1 1.79 16.85 17.07
Okra 2.13 0.58 0.29 1.11 1.03
Onion 0.00 0.00 0.00 0.00 0.00
Tomato 4.22 7.52 7.65 9.66 7.26
Other Vegetable: Fresh 9.98 21.55 48.68 25.80 26.50
All Other Vetegable 13.51 8.32 6.77 9.04 9.41
Oil

Peanut Oil 0.00 0.00 0.00 0.00 0.00
Palm Oil 36.21 20.52 13.49 28.96 24.79
Sheanut Oil 0.00 0.00 0.00 0.00 0.00

Sugar 0.00 0.00 0.00 0.00 0.00

 

189

Table A4-3: Nutrients Contributed by Speciﬁc Food Items (%) by Season

 

 

(continued)
Commodities Vitamin A
Aug Nov Feb May Avg—
All others

Butter 0.77 0.94 0.00 0.72 0.61
Buttermilk 0.00 0.00 0.00 0.00 0.00
Fresh Milk 0.02 0.03 0.01 0.00 0.01
Condensed Sweetened Milk 0.16 0.01 0.00 0.00 0.04
Powdered Milk 1.48 1.60 3.34 1.86 2.07
Eggs 2.20 0.31 0.20 0.00 0.68
Peanuts 0.01 0.02 0.02 0.01 0.01
Seeds 0.00 0.00 0.00 0.00 0.00
Other Legume Nut and Seed 0.00 0.00 0.00 0.00 0.00
Coffee 0.00 0.00 0.00 0.00 0.00
Tea Lipton 0.00 0.00 0.00 0.00 0.00
Green Tea 0.00 0.00 0.00 0.00 0.00
Quinqueliba 0.00 0.00 0.00 0.00 0.00
Other Beverage 0.01 0.00 0.01 0.00 0.00
Banana 0.93 3.69 3.39 0.01 2.00
Citronella 0.00 0.00 0.00 0.00 0.00
Dates 0.02 0.04 0.02 0.00 0.02
Lemon 0.00 0.02 0.00 0.00 0.01
Raisin 0.00 0.00 0.00 0.00 0.00
Tamarind 0.00 0.00 0.00 0.00 0.00
Orange 0.09 0.00 0.00 0.00 0.02
Seansonings and Spices 0.01 0.02 0.00 0.00 0.01

Sum 100 100 100 100 100

 

 

190

Hum—m: E 8m 38 30:0 05 do.“ moigi :05. E023 2 05 we Becca—mu snowman; 03 3:2qu8
38833 05 35 8865 32?; Eon ”momoﬁqoaa E “coconut 8a muggé ”manage n .32 u m 98 5052-33 n beacon
.EzanOmaoEE u we ace—«33m ::e< be 2:85 _mox n msm ”882

n In doing M 598302 n E 5038 can: u «msmsax n A

 

 

 

 

 

 

 

some 33 83 eons 22 3nd :3 82 mm; $3: 82 82. $2 83 23 223a.
8: Sum 82 23 oz... 23 83 owed 9.3 2: 83 22 $2 $3 a: u
am: 33: Geo: $3: 5.9 83.: 3.: awn: a3: 83.: Gem: 5.9 9...: WW: and:
52 SS. 83 :2. $3: 2:: mad E; 23 ”3.2 was $3 8; some 2?: €228:
82: £2: 33:: 3.8 3:: so: 8.: 3.: 5.8 A8: E: am: 83: $98 Ea:
9.3 30.0. 23:. Sod :3: wood. tad- mac- :3. ES- e85- e3? 83. 83:- 8::- m<
«3.3 so: so: 84.8 :3: 8”,: 9.3: A8: 5.8 ::.8 an: ::.8 5.2 so: 64.8 3.98 an: 3.3:
on: Sod- Memo. :3 43¢- $3- 95.. on; 23:. 83 SS- 3.3. one. Rod- 82:. ”so 83. 32:. use:
«3S 2%.: ANS: :6: SS: 82: 3%: so: 33: 6...: 95.8 8&8 ass: 5.: 5.8 E: :3: :3:
So: 32:. 2::- £3- Sod- 9.3:- 23 9.0.0- 8:: Ed- 83. $2.. 83 $3. :3 2::- 32:. ES- SEE
32$ 8.: 62: em: 82: 5.3 ewe Geo: 3.8 :2: $2: $2: 63: ES: 8N8 AS: 82: 5,8
22 ::.o .86 £3 8:: $3 82 23:- _86 £3 $3: $3. 82 Sod :3: 83 ES omoo as:
3% Rod: see 82: so: :2: $3: 83: Gem: Red 5.: 5.: see: 92.0 so: an: ES: 8.3
83 mad :3 2::- :2 e3? 83. $3. $3. 32- 83. «N3- 83 3.2:- £3. 8::- 83. one- :5.
AS: 80.8 30.: $2: 25.: 39: 32: 34.: 2.: £2: 82: 8”: 3.3 E: 5.8 sod: 80.8 so:
32 33. E: 82: 82 $3- 03:: £2- 8::- $3 $3. 32. 82 $3. 82. 32 N23 82- more
see: 33: 5.: 33: ::..: $3: «3.3 ::..: Gm: so: :2: E: 3:: 93.: 5.: and Go: A2:
8? some 23 82 52: 22 8M: 23 3:: 83 48.0- 83 83. Ne; 83 ES 8:: 83 use
326 8.8: a _E _ z 4 >56 38: a _E _ : L 4 326 38: d _ E _ z _ n . mozeE>
muoom 550 E05 8:030 m2 Sm Eoi mote—m0 8:030 .30... Engage:—

 

eo we; UBES> walonzomoo

 

8383mm— ‘Eaﬁon— «:25: Z 313.. 03.2.

191

3:8. ... o... .8. 32.0 2.. 8.. ﬂag-.. ”.26. .582. o. o... ... .cwoc.:m.m 3.3.3.6.... 8.. 35.0508 BEEEM. o... ...... 83...... 32....-. 20..
”8855:... a. 3:82 o... gag-..- nwcgm... u .3.). u n. ...... ~85—Tao: n baa-50... n In. ...-6E9. N .38302 n I 6083 :3. n .392. u -. .882

 

 

 

 

2.... ...-N... ....m... E... ...»... .8... «.3.. ....m... ..N... 2...... 03:3-..
8.... ...m. 2.... ..8... 8... EN... ...... 8a.. 2.... 8.... m
.8... .8... an... ...-...: .2.... 3...... .3.: 9...: 2...: .3.:
.5...- m...... .. ...m- 23 a... 8mm :3 m... .. .2... 2...: .2328.
..m... .3: .2.: ...... .8... .8... ...N... 6...: ...... .3.:
....N... 8.... 8.....- ......- E... 8.... 2.....- oNod- .8... 3......- m...
S... .32: .8... .8... .3.... E... ...... 5.: 3...... .8... 8...: .2....
...... SN..- .3....- Q . ...- ....N..- .... ...- ...... .8...- eo....- .8...- NN . ...- mm . ...- 3.3....
a. ... .2.: .3.: .3.... .... .... ...-...: 6...... ...-e: .. . .... 5.: .2.: ......
..m . ... ...... ......- 8. ...- 8.... m... ... .... . .N ... ...- 2.... 9mm...- Nm.....- 5...... 3.3...
.3.... .N..... ...... .3.... ...-m... .8... 3.... 3...: .2.... 5.: ...... .2.:
..mm... 8.... SN... .2...- S... ...... ..m... 2.....- 3.....- 3.... .8... 08...- ......
s. ... .2.: .N... : .2.. : .3.... .9... a... .3.: .8... : 2...: 3...: ...m:
..S... ......- ....:. o . .....- ... .... $2. ...-...... ... . ...- ...-m..- : .....- m . .... .2...- ......»a..-.....§.
.... ... 8...: .N. .... .2.. : .... .... 9...... am .... .3.: as... ...m... .o. . : 8...:
...... ...... a... 9.2 ...: ...- on . . . - .... . .. 2.....- ..m....- 8... 2...... ...... .- ......
.8 ... .8... an : .3.: ...m. : .8: s: ... .3.... .mm: 8...: 2...: .2.:
...... ...-5... 3m... ...... 3.... mm... ..m... 3.... 2.0.... no... ...-...... ...... 3.058533.
32.. 8.8. . A. . E . I T . 326 8.8. T a. . E . z . 8323>
< 58.35 580...“. 30:22.3...

 

 

3&3 no...m..m> Savannah.

305.559. 838.3,...— ...—«Eun— Eo.-:=Z ...-3. «En-_-

 

192

.mozm: E 03 “m8 26:0 05 Sm 853?; :26.
:5ch A: 05 “a 33$ch £33?me 08 356508 33850 05 $5 0329: 82“?“ Eon $085552“ E 83.53“ 03
mos—«>9 mwcucﬂm n >22 u m 93 5052.33 u Sagan u E €253 M 538302 n I £033. 52 u «mama/x u A M802

 

 

 

 

 

 

 

on; a; 82 ~85 5s 83 £3 33 2:3 :3 2.2.3.8
83 83 82 8,3 23‘ 8: 83 $3 83 82 u.
898 so: cam 858 Gui awe and and 3&8 $8
83 8.3 83 5a 882 ES 23 E; mi 33; 953208
£08 628 9:8 ::.8 5.8 398 $8 an: av: so:
88¢ Rod :3 83. 08.8 :3- :3 was- 3.3- «8.8 m2
::.8 am: A98 8:8 30: :9: «3.8 ::.8 3&8 so: 888 5.:
on: $3. ~88 Rod- Ed- 88.? $3 ::.8 ”So 33. $3- 2.3- 338$
«3.8 A38 A88 80.: 5.8 an: :88 aw: A8: $9.8 $88 3.:
8M: 82.. ”So 0:? ES- :3 on: 82- ES- 33- £3- 23. $.8an
5.8 an: $68 ::.8 3N8 §.: 668 8:8 A38 A88 8&8 5.8
82 :3- £3- 23. ~38 83- 33 :3. 08.8 $3 ”8.8 23 E:
888 5.8 $98 and ::.8 $38 888 5.: 80.8 5.: 888 93.:
83 $3- 83 :3- 23. 82. 82 33. ~88 M23- Sod- 32. 538538-385;
388 358 A8,: $8 $8 ::.8 $8 $8 $8 6.8 5.8 Go:
33 83- 3:- $3 82 OS... 83 82 5.8 02.0 $2 23. 3.5
:88 :88 an: $8 ::.8 and 3.8 and 3.8 so: 358 §.8
83 828 Ed :35 N88 33 23 Ma; 22 83 83. mm; msueSsEm
325 38“: a _ E; z _ A 326 8.8: m _ EL : _ 4 3325
.5: E2030 E2253:—
80 we: ”03mg; E08590

 

Aaoaatucuv 8383mm @5339 2.3.5: Z 34¢. «Emu.

193

CHAPTER 5

SUMMARY, POLICY IMPLICATIONS,
AND DIRECTIONS FOR FUTURE RESEARCH

5.1. Introduction

Mali’s market liberalization reforms, initiated in the 19808, have improved the
availability of cereals in most markets. However, economic accessibility remains a
problem for households, partly because the reforms resulted in higher and more variable
food prices. The seasonal variations in food prices translate into seasonal changes in
urban households’ real incomes, which in turn affect the quantity (level) and quality
(nutrition) of food available in these households.

This study tested the hypotheses that Bamako households’ consumption patterns are
responsive to changes in their real incomes and relative prices and that the income and
price response of demand for commodities and nutrients will change from one season to
another.

The primary objective of this study was to examine the impact of seasonal changes in
Bamako households’ real incomes and relative prices on their consumption patterns using
the complete demand systems approach and household-level panel data. The speciﬁc
objectives of the study were:

1. To describe households’ seasonal changes in expenditure/consumption patterns

and nutrient availability for households in Bamako;

2. To estimate income and price elasticities of demand for various commodities and

commodity groups for different seasons and test whether there is evidence of

seasonal changes in income and price responsiveness; and

194

3. To estimate price and income elasticities of demand for various nutrients across

seasons and assess the stability of the estimated parameters across seasons.

The impact of seasonal changes in Bamako households’ real income and relative
prices on their consumption patterns has not previously been investigated in Mali.
Therefore, this study, through the estimation of disaggregated consumption parameters,
attempts to make a signiﬁcant contribution to food policy formulation in Mali. The panel
data used in this study is from a 2000-2001 survey undertaken in Bamako by the
Direction Regionale du Plan et de la Statistique (DRPS) of the Direction Nationale de la
Statistique et de l’Informatique (DNSI) and the Projet d'Appui au Systeme d'Information
Décentralisé du Marché Agricole (PASIDMA) of Michigan State University (MSU), the
Assemblée Permanente des Chambres d’Agn'culture du Mali (APCAM), and the Centre
d’Analyse et de Formulation de Politiques de Développement (CAF PD). The same 40
households were interviewed in a survey that was conducted in four rounds over a period
of one year starting in August 2000 to May 2001 for the capital city, Bamako.

This chapter begins with a summary of the main ﬁndings of the study and the policy
implications and ends with a discussion of its limitations and scope for future research.
5.2. Summary of Main Findings

5.2.1. Essay I (Chapter 2): Seasonal Changes in Expenditure Patterns and Nutrient
Availability for Households in Bamako, Mali: A Descriptive Analysis

The descriptive analysis revealed that Bamako households’ mean real expenditures vary
considerably across seasons. Households’ mean weekly real expenditures per adult
equivalent are highest in August, the lean season, due in part to large remittances
received at that time and are lowest during in May, the planting season. They decrease by

36 percent between the lean and post-harvest season (August and November), increase by

195

4 percent between the harvest and post-harvest season (November and February), and
drop by 18 percent between the post-harvest and planting season (February and May).
Total real expenditures were disaggregated into food and non-food expenditures
in order to uncover the causes of the strong seasonal changes in expenditures. The results
indicated that much of the observed seasonal variation in expenditures could be attributed
to changes in non-food expenditures, as food expenditures remain fairly stable across
seasons. There are two possible explanations, which are not mutually exclusive, for the
observed seasonal variation in non-food expenditures. The ﬁrst is that households may
attempt to smooth their food consumption levels across seasons by incurring large
changes in their non-food budget. This can be explained by the fact that these
households, especially poor households, consume near subsistence levels of food, thus
are more likely to make large cutbacks in their non-food expenditures because this is the
only way for them to maintain their food consumption levels. However, the observed
seasonal variation in non-food expenditures could also be due to the seasonality of
demand for non-food commodities. For instance, households’ expenditures on clothing
and footwear are generally highest in August as they prepare for the school year, which
begins in September, and during periods of religious festivities, such as the Tabaski.
Hence, the issue for households could either be one of smoothing consumption in the face
of variable income and/or one of meeting seasonally high expenditure requirements in the
face of relatively stable income. One must keep in mind that, given the extreme level of
poverty that prevails in Bamako, households will have limited scope for discretion with

respect to their spending.

196

The results on nutrient availability, showed that Bamako households’ diets are
overwhelmingly based on starchy staples, in this case cereals, as average annual
carbohydrate availability exceeds the F AO recommended dietary allowance (RDA) by 36
percent. The results showed that the average annual calorie availability in urban
households slightly exceeds the F AO’s minimum daily energy requirement of 2,200 kcal
per adult equivalent but only households in the high-income group attain this
consumption level. In addition, there are some signiﬁcant micronutrient (vitamin A,
vitamin C, and calcium) deﬁciencies persisting in Bamako, even in high-income
households. Bamako households can only satisfy about 60 percent of the RDA for
Vitamin A, 73 percent of the RDA for vitamin C, and 42 percent of the RDA for calcium.
However, these estimates of nutrient availability were solely based on the at-home food
consumption data. Sensitivity analyses were performed to assess the effects of including
estimates of nutrient availability from away-from-home foods on average daily nutrient
availability. The results show that if away-from-home foods were taken into account,
average nutrient availability in Bamako households would increase by 9.5 percent. The
results also indicate that all income groups would now be able to meet minimum daily
calorie requirements, but only the high-income group would be able to satisfy the
recommended dietary allowance (RDA) for protein. Moreover, the increase in the
amounts of vitamin A, vitamin C, calcium, and iron will not be enough for households in
all income groups to meet the RDA for these nutrients.

Furthermore, the results indicate that the household availability of all nutrients
varies considerably across seasons. The greatest variation in nutrient availability is

observed between the post-harvest and planting seasons, and the smallest variations are

197

registered between the harvest and post-harvest seasons. A close examination of nutrient
availability in urban households revealed that food consumption smoothing was also
achieved through substitutions between and within food commodity groups. These
adjustments result in large variations in the quality of food available in the household, as
measured by protein, carbohydrate, and micronutrients’ availability. The results indicated
that seasonal variations in micronutrients (vitamin A, vitamin C, and calcium) are much
more pronounced than seasonal variations in calorie availability. Households maintain
their calorie availability somewhat constant during the year by making substantial
changes in the consumption of foods that contain essential micronutrients but few
calories (i.e., meat, ﬁsh, and vegetables). The results indicate that Bamako households
diversify their diets, through greater consumption of non-staple commodities, only during
periods of greater food availability in urban markets, when food prices are relatively low
(i.e., harvest and post-harvest seasons).

5.2.2. Essay 11 (Chapter 3): Estimating the Impact of Seasonal Changes in Real
Incomes and Relative Prices on Households’ Consumption Patterns in Bamako,
Mali, Using the Almost Ideal Demand System Model

In this essay, the Almost Ideal Demand System was applied to a three-stage demand
model for different seasons in order to estimate the impact of seasonal changes in
Bamako households’ real incomes and relative prices on their consumption patterns.
First, the results indicate that price, income, and household size factors account for a
substantial part of the observed variation in the budget share devoted to the commodities
considered in the Stage I (total expenditure allocation), 11 (food expenditure allocation),

and III (staple expenditure allocation) models.

198

Second, the study ﬁnds that Bamako households’ consumption is responsive to
changes in real incomes and relative prices in any given season and that that there are
seasonal changes in income and price responsiveness for all the commodities in the three
demand models. This implies that the impact of a uniform food policy on the quantity and
quality of food available in Bamako households will vary by season.

Third, the fact that the responsiveness of food and staples’ consumption to
changes in real income remains fairly stable across seasons compared to that of non-food
and non-staple commodities indicate that Bamako households engage in food
consumption smoothing from seasonal shocks in real incomes. Food consumption
smoothing was achieved at the expense of non-food commodities such as health and
durable goods (housewares and education), of non-staple foods, and through signiﬁcant

substitutions among and between broad commodity groups.

Fourth, the estimated price elasticities indicate that (l) the price of food has strong
and statistically signiﬁcant uncompensated effects on the demand for non-food
commodities, such as health and education; (2) the price of staples has striking impacts
on the demand for non-staple foods, which are sources of high-quality protein and
micronutrients and; (3) the price of rice has a positive effect on the consumption of

maize, meaning that rice and maize are net substitutes.

5.2.3. Essay III (Chapter 4): Estimating the Effects of Seasonal Changes in Real
Incomes and Relative Prices on Households’ Demand for Nutrients in Bamako, Mali

In this essay, the effects of seasonal changes in Bamako households’ real incomes and
relative prices on their consumption patterns were examined using Engel functions. First,
the results indicate that price, income, and household size factors account for a

substantial part of the observed variation in the demand for nutrients in Bamako

I99

households. The R2 ranges from 12 to 51 percent for calories, 13 to 40 percent for
protein, 12 to 30 percent for vitamin A, 16 to 40 percent for calcium, and 8 to 49 percent
for iron.

Second, the estimated results show that the demand for nutrients is responsive to
changes in real incomes and relative prices. Out of 35 estimated nutrient-income
elasticities, 22 are statistically signiﬁcant at least at the 10 percent signiﬁcance level. Out
of 175 price parameters, 40 are statistically signiﬁcant at least at the 10 percent
signiﬁcance level. Moreover, the null hypothesis of stability in the nutrient-income
demand parameters across seasons was rejected at the 10 percent level for all the
estimated coefﬁcients, except for calcium, suggesting that there is a statistically
signiﬁcant shift in the response of nutrient demand to income changes across seasons.
This means that the impact of changes in income on the demand for nutrients is not
constant across seasons. The Chow test results also indicate a certain degree of non-
constancy of many price parameters across seasons as the test of stability in the price
coefﬁcients was rejected at the 10 percent level for 13 out of 3S estimated coefﬁcients.
The results show that increasing households’ incomes will improve the quantity (i.e.
calories) and the quality (protein, minerals, and vitamins) of food available in those
households in any given season. In addition, the income and price responsiveness of
calories from staples remain stable across seasons while, that for calories derived from
other foods, calcium, and vitamin A varied quite substantially.

5.3. Policy Implications
The ﬁndings of this dissertation have several implications for development planning in

Mali. First, the fact that the empirical analysis substantiates Engel’s Law (i.e., the

200

demand for food is income inelastic), suggests that in the course of economic growth in
Mali, the focus of economic activity will shift away from the agricultural sector. The high
absolute level of the food and non-food income elasticities suggests that (a) this shift will
be slow and (b) policies that aim at increasing households’ real income will result in
substantial improvements not only in the quantity of food available in urban households
but also in the demand for non-food commodities. Rapid growth in the demand for non-
food commodities could translate into sizable rise in employment, to the extent that these
commodities can be produced domestically. In addition, the relatively high-income
elasticities for food suggest that in the initial stages of growth, the demand for food will
continue to grow rapidly, especially for vegetables and animal products, both of which
can generate substantial employment.

Second, the results indicate that changes in the price of food, which are mainly
driven by variations in the price of cereals, will substantially increase or decrease
households’ real incomes, as Bamako households allocate a sizeable proportion of their
budget to food. The resulting real income-induced impact of volatility of the price of
food has strong effects on the demand for non-food commodities, such as health and
education. These ﬁndings suggest that policies that affect food prices, more speciﬁcally
cereals’ prices, through their impact on households’ purchasing power, will have
repercussions on households’ access to health and education services. Thus, the
country’s development policies need to be based on a multisectoral approach, in that
these policies need to be designed to systematically take into account the linkages

between various sectors of the economy.

201

Third, the results revealed that rice dominates Bamako households’ diets despite
the fact that it constitutes a more expensive source of calories than other staples (millet,
sorghum, and maize) and that there are some signiﬁcant nutrient and micronutrient
(vitamin A, vitamin C, iron, and calcium) deﬁciencies persisting in Bamako. These
ﬁndings suggest that, given the importance of rice in Bamako households’ diets and the
presence of very signiﬁcant shortages of micronutrients, successful breeding strategies to
incorporate these micronutrients into rice might have a high payoff, in terms of reducing
malnutrition in these households.

Fourth, the empirical results for food commodity groups showed that as Bamako
households’ real income increases, they will increase their expenditure on non-staple
commodities (e. g., meat and ﬁsh and vegetables) more rapidly than on staple foods. As a
consequence, Bamako households will diversify their diets, through greater consumption
of non-staple commodities, as their income grows. The empirical analysis on nutrient
demand also showed that increasing households’ incomes will improve the quantity (i.e.
calories) and the quality (protein, minerals, and vitamins) of food available in those
households in any given season. This implies that policies that aim at increasing
households’ real incomes will also be an effective mechanism in reducing malnutrition.
These ﬁndings suggest that the pattern of production within the agricultural sector will
have to change with economic growth, as increased specialization in livestock and
horticultural production will be required. Hence, greater allocation of resources and
investment in the production and marketing of horticultural commodities has the potential
to substantially reduce malnutrition, especially vitamin A deﬁciency, increase

employment, and reduce poverty in urban areas. This is feasible because commercial

202

production of horticultural goods is largely concentrated in the peri-urban zones due to
the perishable nature of these commodities and underdeveloped market and road
infrastructure. Also, production and marketing activities, mainly performed by women,
are small in scale and are dominated by low-income households (Morant and Caldwell
(1998)).

Fifth, one of the main objectives of this study was to estimate income elasticities
of demand for various commodities and commodity groups for different seasons in order
to investigate whether there exists, in the Malian context, any self-targeting foods (i.e.,
inferior goods). The results indicate that there are no inferior goods in the commodities
studied. This ﬁnding, consistent with the ﬁndings of previous consumption studies
(Rogers and Lowdermilk, 1991 and Reardon et al., 1999), point to the fact that as
Bamako households’ income increase, the immediate concern is to increase the quantity
of food consumed. This underscores the fact that in Mali the consumption patterns of the
poor and rich are very similar. High-income households tend to consume more of the
same type of foods that poor households eat, even if some diversiﬁcation of the diet is
evidenced at higher income levels.

Finally, the ﬁnding that the response of households’ consumption patterns to
changes in real income and relative prices was not stable across seasons implies that the
impact of a uniform food policy will vary by season. Moreover, “ a government that
decrees uniform prices for the entire year usually ﬁnds itself handling the entire marketed
surplus rather than just a small margin to dampen high prices.” (Timmer et al., 1983, p.
68). Therefore, the effectiveness of food and nutrition interventions (e.g., food-for-work

programs, general food price policy) could be substantially improved through temporal

203

targeting mechanisms in order to reduce transitory food insecurity in households.
Examples of targeting mechanisms to increase low-income households’ access to food
include seasonal income transfers to low-income households and seasonal imports of rice.
“With successful temporal targeting of food subsidies, the signiﬁcant welfare gains from
improving the seasonal distribution of food consumption have both economic and
nutritional components.” (Timmer et al., 1983, p. 68).

5.4. Directions for Future Research

There are three main issues that can be addressed in future research. First, further
investigation of the ﬁnding that the demand for maize is highly elastic is needed.
Identifying the factors that are driving the high price elasticity of demand for maize has
implications pertaining to the expansion of maize consumption as a substitute for rice.
Future research in this area will help determine if there is indeed scope for focusing on
maize as a potential substitute for rice, since more technology exists to increase maize
production in the short to medium run than for millet—sorghum.

Second, this study was solely based on data collected in the capital city of
Bamako. How seasonal changes in real incomes and relative prices affect the
consumption patterns of households in other urban and rural areas is still uncertain at this
time. Therefore, similar studies of consumption patterns in areas of Mali with different
economic characteristics could provide substantial knowledge that can be instrumental in
the formulation of national development policies.

Third, the data used in this study was gathered at the household level and could
not be used to assess the effect of price and income changes at the individual level. The

present study was also not able to investigate the issue of food distribution within the

204

household. Future consumption studies can be systematically designed to obtain
information on the distribution of food within the household, with special emphasis on
food away from home consumption. This may allow the identiﬁcation of vulnerable
groups (i.e. children, pregnant and lactating women, and the elderly) in low-income
households. Strauss and Thomas (1995) have examined the allocation of food within the
household and have found that there exists gender bias in intra-household resource
allocation (e.g., males are favored over women). Gathering such information will also
permit the analysis of the nutritional status of household members, which require
information on nutrient intake, rather than focusing on nutrient availability. The effect of

income and price on individual nutrient intake could be useful in designing food and

nutrition policies.

205

REFERENCES

Morant, Mervalin and Caldwell, John S. (1998). Peri—urban Horticultural IPM Research
in the OHVN Zone of Mali: Bio-control Methods for Green Bean Seedbed Disease
Control and On-Farm Trial Monitoring. IPM CRSP University of Maryland and Virginia
Tech trip report.

Timmer et al. 1983. Food Policy Analysis. Published for the World Bank. John Hopkins
University Press.

206

   

NlllllllllllllTill?"