”37.19%“?
y

:1;-
§,)_
7

....e
. : 1.5.;
. A21:

:uufi‘. ..
«hflltlt§l
. 1 37

.7. I. .
!:u.u.l.r
\I‘ ¢Q~ .

....W'WWMI:%1AU ($1.1. ‘ ‘.
wggrd u. .« ‘ LE r!“ 73..

aggaamwmwm.

 

a1"... a". ..h‘ .r.._x9l...J.wr

$8.?! . ...Ir
...Awimnv. \t

. M.h....mn Hair

Jam»

7 .

34:: »
filai‘
x

...
23
. A .
flux. $5.?! 4.9“
... ......C ..flaofils .3

. . .
, ...:i 3:73...
...? V. . AW .
u . J . v .3.) .152. L

.2. .
fink?
344?”?

:3

. -

k 5.
12.1.1!
5(3‘ ..I

Z .

hazis. ...
5‘ .15 .

.5... zany.
. rm

1

V

333...

irruusu

..
by .3.

:.§:

.t.

".‘2'1

 

5‘

u. .172
:VCtaw...
s...» v.‘ i.

t. v .3:
ll

X. \\||-nv‘¢.¥.|

v loft}:
‘t‘.

:SIJS.
4“. y

. 1x41.
...? 3. vv
I, “\

J‘s.
. r ‘11:» A .-
\. t . I“ ‘

”EC/s

7

J _ LIBRARY
109,6 Michigan State
University

This is to certify that the
dissertation entitled

IMPACTS OF GLOBAL AGRICULTURAL TRADE REFORMS AND
WORLD MARKET CONDITIONS ON WELFARE AND FOOD
SECURITY IN MALI: A CGE ASSESSMENT

presented by

Kofi L. Nouve

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Agricultural Economics

 

 

W341;

Mflor Professor’s SWre

’chembuz 1‘4: 1005

 

Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

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

 

APPI_4ZOM

101909

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/C|RC/DateDue.indd-p.1

IMPACTS OF GLOBAL AGRICULTURAL TRADE REFORMS AND WORLD
MARKET CONDITIONS ON WELFARE AND FOOD SECURITY IN MALI:
A CGE ASSESSMENT

By

Kofi L. Nouve

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Agricultural Economics

2004

ABSTRACT
IMPACTS OF GLOBAL AGRICULTURAL TRADE REFORMS AND WORLD
MARKET CONDITIONS ON WELFARE AND FOOD SECURITY IN MALI:
A CGE ASSESSMENT
By
Kofi L. Nouve

The Malian agricultural sector faces a series of trade reforms originating from
within West Africa and from the rest of the world. These reforms are expected to change
income and food consumption levels in Mali, but the direction and the magnitude of these
changes are unknown. This dissertation contributes to the understanding of how, and the
degree to which, these reforms would affect welfare and food security in Mali. The
analysis is based on computable general equilibrium simulations using a 1997 social
accounting matrix that has been specifically built for the purpose.

The dissertation is organized in seven chapters. Chapter I reviews the theoretical
and empirical evidence supporting the need to investigate the impacts of trade reforms in
the specific context of the Malian economy. Chapter II discusses the use of the Hicksian
equivalent variation to measure welfare impacts, as well as the use of changes in
household food consumption as a proxy measure of food security. Chapter III presents
nine trade reforms scenarios, organized in four groups: (i) the FAPRI and OECD price
change scenarios of partial reforms of world commodity markets; (ii) the IF PRI-l status
quo and the IF PRI-2 full liberalization price change scenarios; (iii) the EEP effective
erosion of existing preferences and the DE A complete duty-free access preferential trade

scenarios; and (iv) government policy scenarios on applying a common external tariff

regime, banning cereals exports, and increasing investments in key sectors.

Chapter IV presents the analytical method, which is a single-country computable
general equilibrium (CGE) model in the neoclassical structuralist tradition. The Malian
model is based on a standard CGE model from the International Food Policy Research
Institute, which is itself based on the Dervis, de Melo and Robinson (1982) seminal work.
The CGE framework uses data from a disaggregated social accounting matrix (Chapter
V), and the simulation results represent counterfactuals the nine trade reform scenarios.

The results, presented in Chapter VI, reveal that Mali has as much to gain from
increased agricultural reforms in world markets, as it has to gain from deepened
commercial integration in West Africa. The gains would amount to an average of three
percent of initial income levels. Most of the gains would go to urban consumers who
would benefit from reduced prices, appreciated real exchange rate, and increased factor
incomes. In general, and in absence of productivity gains, rural producers would lose
reductions in world commodity prices. The negative effects are mitigated, and could even
be reversed, if Mali benefits from gains in productivity. The results also indicate that
expanding existing trade preferences would raise incomes and food consumption in Mali,
whereas reducing or eliminating these preferences would reduce incomes and weaken
food security in Malian households.

The last chapter of the dissertation discusses several implications and limitations
of these results. Mali would gain by allowing global market forces to work in the
economy, expanding regional trade, increasing investments in key sectors, and improving
agricultural productivity. While the first two options come at the cost of increased urban
inequality, the last two have the potential to deliver Pareto-compatible results. Overall,

the analysis may be refined by improving the underlying social accounting matrix.

To my family

iv

ACKNOWLEDGMENTS

Many people and institutions made the completion of this dissertation possible. I
feel blessed, and I am very grateful, for all the support I received throughout my doctoral
program at Michigan State University.

My greatest thanks and appreciation go to Dr. John Staatz, my major professor,
who mentored me through the program and supervised this dissertation. It was a great
privilege and inspiration working with him. He was very generous with his time and
patiently and carefully read several drafts of this dissertation. He offered invaluable
comments that contributed immensely to the quality of this work.

I also thank Dr. Richard Bemsten for serving as my temporary adviser and as an
active member of my dissertation committee. I am equally grateful to Dr. Robert Myers,
Dr. David Schweikhardt and Dr. Jay Choi for their constant availability and their
generous guidance as members of my dissertation committee. I should also mention the
valuable support of Dr. Rich Sexton at U.C. Davis who, during the final stages of this
project, facilitated my interaction with faculty and students and my access to the UC
Davis department resources, including the GAMS software I used for the analysis.

This dissertation and my entire doctoral training were made possible through
graduate research assistantship from the Department of Agricultural Economics under the
Sahel Regional Program of the Food Security II/III Cooperative Agreement between
Michigan State University and the USAID. I also received a Dissertation Completion
Fellowship and an Emergency Fellowship from the Graduate School. I thank everyone,
and particularly Dr. Eric Crawford, for facilitating the access to these resources. Many

department staff members went out of their ways to help with various logistics. I am

particularly thankful to Sheryl Rich and Debbie Conway for their valuable administrative
assistance, and to Josie Kiel for helping with travel arrangements.

Dr. Mbaye Yade served as the anchor of my research works in Mali. I want to
thank him and all his staff at the Institut du Sahel, Bamako, for their superb support. My
data gathering activities brought me to several institutions in Mali, including the
Direction Nationale de la Statistique et de I’Informatique (DNSI), the Institut
d ’Economie Rurale (IER), the Observatoire de I’Emploi, and the Observatoire
Economique et Statistique de l'Afrique Subsaharienne (AFRISTAT). I would like to
express my sincere thanks to the administrators, researchers and staff at these institutions
for sharing their expertise on various aspects of the Malian economy.

A unique atmosphere of mutual support and collaboration pervades the Agecon
student community at Michigan State. This invaluable social capital made my graduate
studies in East Lansing one of my most enjoyable experiences of my life. I particularly
want to thank Gerald Nyambane, Meeta Punjabi and Lourdes Martinez for their
friendship and all the many administrative arrangements they have to make on my behalf.

Finally I want to thank my family for their constant love, prayers, and
unconditional support. My wife Regina Geinignani was an exceptional companion whose
love and care carried me through the challenging years of my doctoral program. Her
enthusiasm and encouragement gave me the confidence I needed to stay focused on the

dissertation. I dedicate this work to Regina and to my entire family.

vi

TABLE OF CONTENTS

LIST OF TABLES .................................................... ix
LIST OF FIGURES ................................................... xi
LIST OF APPENDICES ................................................ xii
CHAPTER I: INTRODUCTION ........................................ 1
1.1. Overview ..................................................... 1
1.2. Problem Statement ............................................. 3
1.3. Objectives of the Study .......................................... 6
1.4. Organization of the Dissertation ................................... 7
CHAPTER II: INDICATORS OF WELFARE AND FOOD SECURITY 8
2.1 . Introduction ................................................... 8
2.2. Welfare ....................................................... 8
2.3. Food Security ............................................................ 11
2.3.1. The Concept .......................................................... 11
2.3.2. Food Security and Trade ............................................... 12
2.3.3. Measuring the Level of Food Security .................................. 14
2.4. Conclusion .................................................... 20
CHAPTER H1: DESIGN OF AGRICULTURAL TRADE REFORM SCENARIOS 21
3.1. Introduction ................................................... 21
3.2. Border Protection in the Malian Export Markets ....................... 21
3.2.1. Border Protection in the ROW .................................. 21
3.2.2. Border Protection in West Africa ................................ 28
3.3. Projected Changes in World Agricultural Prices ....................... 29
3.2.1. FAPRI Price Projection Model .................................. 31
3.2.2. IFPRI Baseline Projection: The IMPACT Model .................... 32
3.2.3. OECD Baseline Projection: The AgLink Model ..................... 33
3.2.4. Summary of Model Assumptions and Projected Changes in World
Commodity Prices ............................................ 34
3.4. Government’s Trade and Fiscal Policies ............................. 38
3.5. Scenarios Summary and Conclusion ................................ 44

CHAPTER IV: A COMPUTABLE GENERAL EQUILIBRIUM MODEL FOR

TRADE REFORM ANALYSIS ............................ 47
4 . l . Introduction ................................................... 47
4.2. The Malian CGE Model in the Context of the Literature ................ 47
4.3. Structure and Equations of the CGE Model .......................... 51

vii

4.4. Summary and Conclusion ........................................ 60

CHAPTER V: A 1997 SOCIAL ACCOUNTING MATRD( FOR MALI ......... 61
5.1. Introduction ................................................... 61
5.2. A 1997 Macroeconomic SAM for Mali .............................. 63
5.3. A SAM Balancing Cross-Entropy Method ........................... 66
5.4. Structure and Entries in the Malian Microsam ......................... 72
5.5. Conclusion .................................................... 91

CHAPTER VI: CGE MODEL IMPLEMENTATION AND RESULTS .......... 94
6.1 . Introduction ................................................... 94
6.2. Trial Runs and Re-specification of the Model ......................... 94
6.3. Model Calibration .............................................. 97
6.4. Results ....................................................... 100

6.4.1. Partial Market Reforms: FAPRI and OECD Scenarios ............... 101
6.4.2. No or Full Market Reforms: IFPRI-l and IFPRI-2 Scenarios .......... 109
6.4.3. Duty-Free Access and Preference Erosion: DFA and EEP Scenarios . . . . 112
6.4.4. Government Policy Reforms: BAN, CET and GINV Scenarios ........ 114
6.4.5. Food Security ............................................... 119
6.4.6. Global Trade Reforms with Productivity Gain in Mali ................ 132
6.5. Sensitivity Analysis ............................................. 135
6.5.1. Sensitivity to Model Parameters ................................. 136
6.5.2. Sensitivity to Macroeconomic Closures ........................... 140
6.6. Summary and Conclusion ........................................ 145
CHAPTER VII: SUMMARY, CONCLUSION AND RECOMMENDATIONS 149
7.1. The Research Question ........................................... 149
7.2. Trade Reform Scenarios .......................................... 150
7.3. Construction of a Detailed Social Accounting Matrix for Mali ............ 152
7.4. A CGE Analysis of Welfare and Food Security Impacts of Trade Reforms
in Mali ....................................................... 153
7.5. Achievements and Limitations of the Dissertation ..................... 158
APPENDICES ....................................................... 162
BIBLIOGRAPHY .................................................... 208

viii

Table 3.1:
Table 3.2:
Table 3.3:
Table 3.4:

Table 3.5:

Table 5.1:
Table 5.2:
Table 5.3:
Table 5.4:
Table 5.5:
Table 5.6:
Table 5.7:
Table 5.8:

Table 5.9:

LIST OF TABLES

Average Import Tariffs per Commodities and Importers .............
Key Assumptions in the Price Projection Models ...................
Projected Changes in World Commodity Prices with Four Price Models
Common External Tariffs on Imports from the Rest of the World ......

Summary of Trade Reform Scenarios ............................

Structure of a Social Accounting Matrix ..........................

A 1997 Unbalanced Macroeconomic Social Accounting Matrix for Mali .

A Balanced 1997 Macroeconomic Social Accounting Matrix for Mali . .
Structure of the Malian Microsam ...............................
Distribution of the Value-Added Across Sectors ....................
Distribution of the Value-Added Within Sectors ....................
Distribution of Factor Incomes to Households and Enterprises ........
Sources of Income to Households and Enterprise ...................

Regional Composition of Household Groups (and Composite Rural
Farming Production Factors) ...................................

Table 5.10: Home Consumption of Activities by Household Groups ............

Table 5.11: Budget Shares by Household Groups ...........................

Table 5.12: Commodity Shares in West Afiican and ROW Trade in Mali, 1997 . . .

Table 6.1:

Table 6.2:

Mapping the Original 37 Commodities Into 20 Aggregated Commodities

Projected Changes in World Commodity Prices: FAPRI and OECD
Scenarios ..................................................

ix

27

36

38

41

45

62

65

71

74

80

8O

82

82

83

86

87

89

96

103

Table 6.3: Equivalent Variation : FAPRI and OECD Scenarios .................
Table 6.4: Disaggregated Factor Income Distribution: FAPRI and OECD Scenarios

Table 6.5: Projected Changes in World Commodity Prices: IFPRI-l and IFPRI-2
Scenarios ..................................................

Table 6.6: Equivalent Variation : IFPRI-l and IFPRI-Z Scenarios ...............

Table 6.7: Disaggregated Factor Income Distribution: IFPRI-l and IFPRI-2
Scenarios ..................................................

Table 6.8: Equivalent Variation: EEP and DFA Market Access Scenarios ........

Table 6.9: Disaggregated Factor Income Distribution: EEP and DFA Market
Access Scenarios ............................................

Table 6.10: Equivalent Variation under Government Controlled Scenarios .......

Table 6.11: Disaggregated Factor Income Distribution Under Govemment-
Controlled Scenarios ........................................

Table 6.12: Food Consumption Under Govemment-Controlled Scenarios ........

Table 6.13: Ratio Food Imports/'1‘ otal Exports Under Global, Bilateral, and
Government Trade Policy Scenarios ............................

Table 6.14: Welfare Impacts of a 10% Productivity Gain Under the IFPRI-2 Global
Commodity Price Change Scenario .............................

Table 6.15: Food Security Impacts of a 10% Productivity Gain Under the IFPRI-2
Global Commodity Price Change Scenario .......................

104

106

110

111

112

113

114

116

117

120

130

132

134

LIST OF FIGURES

Figure 1.1: A Framework for Analyzing the Impact of WTO and West Afiican
Trade Reforms on Welfare and Food security in Mali ............... 2

Figure 2.1: Conceptual Framework for Food Security . . . . . . . ................. 13

Figure 3.1: Three Border Protection Regimes with Infinitely Elastic World
Demand .................................................. 25

Figure 4.1: Commodity Flows in the CGE with Regional Differentiation of the
Export and Import Sectors .................................... 5 2

Figure 5.1: A Political Map of Mali Showing Administrative Regions and
Circles .................................................... 84

Figure 6.1: Change in EV in Response to Change in Factor Substitutability (e. g.
of OECD Price Change Scenario) .............................. 137

Figure 6.2: Change in EV in Response to Changes in Export Tradability (e. g. of
OECD Price Change Scenario) ................................ 139

Figure 6.3: Change in EV in Response to Changes in Import Tradability (e. g. of
OECD Price Change Scenario) ................................ 140

Figure 6.4: Changes in EV When Nominal Exchange Rate is Flexible (Initial
Simulation and Fixed (e. g. of the OECD Price Change Scenario) ..... 142

Figure 6.5: Changes in EV with Alternative Closures of the Saving-Investment
Account (e. g. of the OECD Price Change Scenario) ................ 144

xi

Appendix 4.1:
Appendix 5.1:
Appendix 5.2:
Appendix 5.3:

Appendix 5.4:

Appendix 6.1:
Appendix 6.2:

Appendix 6.3:

Appendix 6.4:

Appendix 6.5:

Appendix 6.6:

Appendix 6.7:
Appendix 6.8:

Appendix 6.9:

LIST OF APPENDICES

Mathematical Model: Sets, Parameters, Variables and Equations . . .
Justification of Entries in the Malian Macrosarn .................
Justification of Entries to the Malian Microsam ...... ' ...........
Ratio of Expenditures to Incomes in the Unbalanced Microsam . . . .

A 1997 Balanced Social Accounting Matrix for Mali ............

Base Production and Trade Elasticities ........................
Base and Calibrated Data for Selected Indicators ................

Final and Intermediate Demand Prices Under Price and Market
Access Scenarios .........................................

Summary Macroeconomic Indicators Under Price and Market
Access Scenarios .........................................

Real Imports Under Price and Market Access Scenarios ..........

Disaggregated Real Household Consumption Under Price and
Market Access Scenarios ..................................

Implementation of the CET Scenario .........................
Real Imports Under Govemment-Controlled Scenarios ...........

Real Exports Under Govemment-Controlled Scenarios ...........

Appendix 6.10: Final and Intermediate Demand Prices Under Govemment-

Controlled Scenarios .....................................

xii

162

170

173

186

187

198

199

200

201

202

203

204

205

206

207

CHAPTER I: INTRODUCTION

1.1. Overview

This study investigates the impacts of global and West African agricultural trade
reforms on welfare and food security in Mali. The research consists of designing various
trade reform scenarios and simulating their impacts on welfare and food security in Mali
using a static, single-country computable general equilibrium model. The reform
scenarios are constructed from three sets of assumptions on (i) the direction and
magnitude of change in world agricultural prices following the implementation of the
World Trade Organization’s (WTO) agreements on agriculture, (ii) the change in the
level of border protection facing Malian exports, and (iii) the choice of fiscal and trade
policies by the Malian government. These criteria are the most important elements
determining the welfare and food security impacts of ongoing trade reforms in Mali. The
study is expected to contribute to the empirical literature on linkages between agricultural
trade reforms and food secruity in Sub-Saharan Africa.

In this study, trade reforms are modeled as scenarios which are jointly determined
by the three sets of criteria mentioned above. A scenario is defined based on either (i) a
magnitude of change in world agricultural prices, (ii) a specific level of tariffs and non-
tariff barriers (in tariff equivalent) imposed on Malian exports, or (iii) a given set of fiscal
and trade policy choices by the Malian government (Figure 1.1). Plausible changes in
prices and border protection, as well as feasible fiscal and trade policies by the

government, are identified through a critical review of the existing literature (see Chapter

III).

 

 

 

Possible Changes in Possible changes in Possible trade and

world agricultural border protection investment policies of the
prices [FAPRL OECD, facing exports Malian government
IFPRI-l and IFPRI-2]* [DFA, EEP]** [CET, BAN, GINV]***
Obtain from existing Review the conditions Review policy options of the
global trade models a set under which the Malian government, including import
of scenarios for changes in exports enter the world tariffs and export taxes or bans.
world prices resulting and West African regional Possible policy choices include
from the implementation markets and evaluate how full implementation of the
of WTO agreements. This these conditions would common external tariff regime.
is done by reviewing the change as a result of the Exports bans, though illegal, will
assumptions and the increased market access also be allowed as there is
robustness of the projected brought about preferential evidence that they have been
changes, based on the trade agreements. used in the past. Finally, a public
nature and the degree of investment scenario will allow
implementation of WTO comparison of the impacts of
agricultural agreements. trade and non-trade policies.

 

 

 

 

 

 

 

form 806

General Equilibrium
Model

 

 

 

Model Outputs

Equilibrium prices; Productions at national and household levels; Factor
incomes and incomes by household type

 

 

 

 

 

Welfare performance
indicator: measure of the
equivalent variation at
national and households levels

 

 

 

Food security indicators at national and
household levels: changes in food
consumption, and share of food imports
in total exports (national level only)

 

 

 

 

 

Notes: *, ", ***: Chapter III identifies four different world price change scenarios (FAPRI, OECD,
IFPRI-l and IFPRI-Z). The chapter also discusses two alternative market access regimes, based on whether
the Malian exports will have a Duty-Free Access (DFA) to major markets, or whether there will be Erosion
of Existing Preferences (EEP). Finally, government policy choices include a full implementation of the
West African Common External Tariff (CET) regime, a ban on the Malian cereal exports (BAN), and
increases in public investments (GINV).

Figure 1.1: A Framework for Analyzing the Impact of WTO and West African

Trade Reforms on Welfare and Food security in Mali.

Once trade reform scenarios are designed, they are fed into a computable general
equilibrium (CGE) model to generate a series of results. These include equilibrium prices
and outputs as well as incomes to production factors and to various household groups.
Results from the CGE models are then used to characterize the welfare and food security
impacts of trade reforms at national level and household levels. While subsequent
chapters of this dissertation elaborate on each of the steps outlined in Figure 1.1, the
remaining part of this chapter presents the rationale and the objectives, as well as the
organization of this dissertation.

1.2. Problem Statement

There is an ongoing debate about how agricultural trade reforms under the World
Trade Organization (W TO) would affect welfare and food security in developing
countries, including Mali. Current figures indicate that more than one-fifth of the Malian
population is undernourished (FAO, 2003), while about 65% live under the poverty line
(U SAID-Mali, 2003). Reducing poverty and undemourishment are key policy objectives
for the Malian government, and there is a general understanding among policymakers
that ongoing trade reforms may have important effects on welfare and food security in
Mali.

In this study, trade reforms refer to the WTO agreements on market access,
domestic price supports and export subsidies. The general goal of these agreements is to
encourage reduction of tariffs, eliminate non-tariff barriers, and reduce domestic price
supports and export subsidies to agriculture. In Mali, trade reforms have an additional

West Afiican regional component whereby the country participates in region-wide trade

liberalization policies, characterized by a common external tariff and the implementation
of a free-trade area within West Africa.

The ongoing debate is generated from a general uncertainty regarding the impacts
of trade reforms on welfare and food security. This uncertainty has both theoretical and
empirical components. Theoretically, it is well-known that trade reforms have ambiguous
impacts on welfare. This ambiguity generally has two origins. First, the theory of second-
best holds that trade reforms in an economy that is not perfectly competitive may
generate net welfare gains or losses (Lipsey and Lancaster, 1956).1 No economy in the
world is perfectly competitive. Thus, all economies are prone to the uncertain second-best
effects of trade reform (Suranovic, 1999). Second, this initial uncertainty may be
exacerbated by the distributional impacts of trade reforms. It is also well-known from the
Stolper-Samuelson theorem that trade reforms usually generate factor income gains in
some production sectors and losses in other sectors (Dixit and Norman, 1980; Khan,
1998). Theoretical evidence on the impacts of trade reforms on food security can be
inferred from the welfare impacts of these reforms. Welfare or income changes may be
' translated into changes in effective demand for food, which is an important indicator of
national and household food security.

In addition to the theoretical uncertainty, empirical evidence also suggests that
trade reforms have ambiguous effects on welfare. This is evidenced in existing studies in
Sub-Saharan Africa (e.g., Mbabazi, 2003; Lofgren, 2001; 2002; Dorosh and Sahn, 2000)
and elsewhere (Baustista and Thomas, 1997). In general, the empirical evidence depends

on individual country characteristics. For example, Wobst (2002) shows in the context of

 

' Results from other trade theories, such as the new trade theory (based on imperfect competition and
economies of scale) also suggest that trade reforms have an ambiguous impact on the direction of change in
welfare (Krugman, 1981; Helpman and Krugman, 1985).

five Southern African economies that common policies measures may yield dramatically
different welfare impacts. It follows that the structure of the Malian economy will be a
major determinant of the way in which WTO and West African trade reforms are likely
to affect welfare in the country.

In terms of food security, there are few systematic analyses of the impacts of
ongoing WTO trade reforms on food security in Sub-Saharan Africa, and particularly in
West Afiica. However, there are countless case studies, usually found in online reports by
non-governmental organizations, which generally come to the conclusion that agricultural
trade reforms are harmfirl to food security in poor small-scale farm communities.
Examples of such studies include Posner (2001), Charles, Longrigg and Tugend (2001)
and Madeley (2000). The Charles, Longrigg and Tugend study is a Consumers
lntemational’s assessment of the reforms in 13 developing countries. The study
concludes that, in many poor countries, increased exposure to agricultural trade reforms
‘ weakens food security through increased dependence on food imports and reduced
employment opportunities. Madeley (2000) reviews the experience of trade reforms in 27
developing and least-developed countries and reaches a similar conclusion. The main
limitation of these case studies is that they rely heavily on isolated situations of a specific
community in a given country. These isolated situations rarely reflect the economy-wide
impacts of trade reforms. Thus, they constitute a poor proxy of what could be the overall
effects of trade reforms. Economy-wide effects must be captured within a economy-wide
analytical framework, which is the approach taken in this study.

Given the theoretical and empirical uncertainties outlined above, it is impossible

to establish a priori the direction of change in welfare and food security in Mali following

the implementation of global and West African regional trade reforms. This is primarily
an empirical question, and thus requires further investigation. The need for this
assessment also arises because improving welfare and food security is one of the most
important concerns for poor countries, when it comes to global trade reforms (Diaz-
Bonilla et al., 2000) .2
1.3. Objectives of the Study

The main objective of this study is to assess empirically the welfare and food
security impacts of the WTO-led agricultural trade reforms on the Malian economy while
accounting for regional integration dynamics in West Africa. This main objective is
broken into the following four specific objectives:
(i) To design trade reform scenarios based on expected changes in agricultural prices,
border protection facing Malian exports and government trade and fiscal policy choices;
(ii) To build a social accounting matrix for the Malian economy, differentiating between
the West African and rest-of-the-world imports and exports;
(iii) To simulate the impact of trade reform scenarios in a general equilibrium setting,
maintaining the regional differentiation of trade; and
(iv) To translate the CGE results into welfare and food security indicators at country and
household levels.

The proposed study is expected to contribute to the empirical literature on the

linkages between agricultural trade reforms and food security primarily in Mali, but with

 

2 Improving welfare and food security has also been recognized as an explicit objective of the WTO
agreements. This is acknowledged not only in the preamble of Marrakesh Decision that established WTO
in 1994, but also in the WTO Decision regarding the possible negative impacts of the reform agenda on the
least-developed and Net-Food Irrrporting Developing Countries (NF IDC). The Doha Development Agenda
also stressed on the need to use the trade engine to support development (including improved welfare and
food security) in poor countries.

possible implications for other Sub-Saharan African countries. The contribution will be in
the form of conditional statements on the impacts of specified trade reforms on welfare
and food security in Mali, given a series of reform outcomes depending on changes in
world agricultural prices, border protection in the Malian export markets, and the
government trade and fiscal policy choices.
1.4. Organization of the Dissertation

The dissertation is organized in seven chapters, including the present introductory
chapter. The next chapter offers a smnmary discussion of the concepts of welfare and
food security, in the context of global and West African agricultural trade reforms.
Welfare is measured by equivalent variation at both national and household levels, and
food security is captured by changes in the level of national and household food
consumption (see Figure 1.1). Trade reform scenarios are discussed and presented in
Chapter III, and Chapter IV lays out the analytical model, which is a computable general
equilibrium model in the neoclassical structuralist tradition. Chapter V documents the
construction of a disaggregated, agriculture-oriented, social accounting matrix (SAM) for
the Malian economy. The SAM is the main database used in simulating the welfare and
food security impacts of trade reforms in Mali. Chapter VI deals with the implementation
of the SAM-based CGE model and discusses the simulation results in terms of the

welfare and food security indicators. The last chapter concludes the dissertation.

CHAPTER II: INDICATORS OF WELFARE AND FOOD SECURITY

2.1. Introduction

There are several indicators of welfare and food security in the literature. This
chapter does not attempt to make an exhaustive survey of welfare and food security
literatures. It simply aims at presenting a limited number of indicators that may be used to
characterize the impacts of various scenarios of agricultural trade reforms on welfare and
food security in Mali. In particular, the chapter discusses briefly the equivalent variation
(EV) concept, which is one of the most common indicators used to measure changes in
welfare. Most of the chapter elaborates on the concept of food security at the national
level, and its linkages with regional and global trade. The chapter also discusses how
changes in food consumption may be used to measure changes in the state of food
security under various trade reform scenarios. Both the welfare and food security
measures may be computed under each scenario of trade reforms, allowing a comparison
of impacts between scenarios and with respect to the base scenario.
2.2. Welfare

As mentioned earlier, this section does not conduct an exhaustive review of
welfare indicators. Instead, the aim is to identify an appropriate indicator able to capture
welfare changes under different trade reform scenarios. The equivalent variation (EV) is
considered to be such an indicator.

While there is still an ongoing debate on the appropriate measurement of welfare,
many CGE-based studies rely on the EV measure. For example, Decaluwé, Dissou and

Patry (1998) used EV to measure changes in welfare in seven West African countries.

Anderson and Martin (1996) review several measures of change in welfare and conclude
that EV dominates other measures. EV is particularly attractive for the purpose of this
study because, as explained in Varian (1978; p. 162), this will keep the baseline scenario
as fixed and compare all the other scenarios with respect to this status quo. In addition,
EV may be computed not only for the country as a whole, but also for various household
groups within the country, allowing welfare impact assessment across households.

In order to compute EV, the Malian economy in the base period (1997) may be
viewed as supported by an initial commodity price vector, say p0. Each trade reform
scenario 3 corresponds to new price vector p5. A consumer (or household group) with
income m enjoys an initial utility u0 at price p0 and a new utility us at price p‘. Define an
expenditure function e(p, u) to be the amount of money that a consumer spends in order
to achieve utility level u given the price vector p. EV is defined as follows3:

(2.1) EV = e(p", u‘) — e(po, u°)

EV represents the net change in welfare that causes the consumer to get the utility level uS
at price [70. A consumer would be better off if EV is positive, and worse off if it is
negative.

In addition to EV, further insight into welfare change may be gained through
direct examination of changes in national and household incomes and in the functional
distribution of income to factors of production. This study will make use of this insight to
complement conclusions drawn for the EV measure.

This choice of EV as a measure is not without problems. In the strict sense, EV

measures changes in individual welfare. Thus, the use of EV to evaluate welfare changes

 

3 Varian (1992, p. 167) defines EV as e(po, u‘) - e(p‘, u‘), which is slightly different from the definition in
Mas-Colell, Whinston and Green (1995, p. 82), shown in Equation (2.1). The two expressions are identical,
however, because e(po, u°) = e(p’, u') = m, since m is the wealth needed at any given price and utility level.

in a country or household group raises the traditional aggregation problem. Central to the
aggregation problem is the representative agent assumption. It is clear that for any
household group, individuals are likely to be heterogeneous within that group. Yet, the
representative agent assumption considers individuals to be homogenous within the
representative group, so that any decision made by the representative group is deemed
equivalent to the aggregate choice of individuals acting independently. Kirrna (1992)
offers a detailed critique of the representative agent concept, but falls short of proposing
any meaningful alternative other than calling for “a paradigm in which individuals
operate in diverse subsets of the economy, [and] are diverse both in their characteristics
and the activities that they pursue” (p.134). This is precisely what modern applied
general equilibrium models emphasize in grouping individuals in household groups, so as
to achieve some reasonable level of within-group homogeneity. But even with such a fix,
the fundamental aggregation problem remains regarding the use of EV as a measure of
welfare change. For most CGE-based analyses, however, household grouping is
indispensable if one is to keep a model tractable and relevant for policy applications.4
This study opts for the latter approach, acknowledging that EV may not be a perfect
measure of welfare changes in household groups or for the country at a whole. It is,
nevertheless, a reasonable measure of such changes and is maintained for the analysis in

this study.

 

‘ There is an emerging CGE literature that emphasizes the use of “real households”, instead of the standard
use of “representative households” (e.g., Bourguignon, Robilliard and Robinson, 2003). The application of
the real household approach has delivered results that are encouraging, but the linkage between the macro
model and real household is not yet satisfactory. Furthermore, the new approach requires individual
household data from national household surveys. Such data were not available for this study, hence the use
of the standard representative household approach.

10

 

2.3. Food Security
2.3.]. The Concept

“F ood security” is a very broad and complex concept, which makes it very
susceptible to misuses. In order to avoid misunderstandings, this study adopts the widely
used definition put forward by the 1996 World Food Summit. It will be considered that
“food security exists when all people, at all times, have physical and economic access to
sufficient, safe, and nutritious food to meet their dietary need and food preferences for an
active and healthy life” (F A0, 1996). Thus, food security implies three key conditions.
First it requires adequate availability of food, i.e., a sufficient, sustainable and stable food
supply, as argued by Maxwell (1996). Second implies a sustained access to food by each
individual. And finally, it requires effective utilization of the available food (for details,
see Staatz, 1990; Tweeten, 1999; Barrett, 2002; PCHA, 2002.)

It is clear that insufficient food supply at any level of aggregation implies
undemourishment, hence food insecurity. It is also true that sufficient food availability in
a given geographic area does not assure adequate food access to all individuals living in
that area. This has been the essence of the entitlement literature pioneered by Sen (1981).
It is finally clear that while food availability and access are preconditions for adequate
utilization, they “do not determine unequivocally the more substantive issue of
malnutrition and nutrition security at the individual level” (Diaz-Bonilla et al., 2000; pp.
4-5). The utilization dimension of food security, which depends on intra-household food
allocation and use, goes beyond the scope of this study and is therefore not addressed
here. Instead the study focuses on the first two dimensions of food security. Physical food

sufficiency is measured through nationwide food supply indicators, whereas economic

11

access is captured by both households’ effective demand for food and the country’s
capacity to finance its current food imports through trade.
2.3.2. Food Security and Trade

The idea that trade is an instrument for achieving food security is best understood
with the concept of “self-reliance”. For a country or a region, a self-reliance strategy
consists of securing a sufficient, safe, and nutritionally adequate food for its population,
using a combination of domestic production and imports. The self-reliance concept is
widely discussed in the literature and appears in trade and food security-related works,
such as Martin (1988), Sumner (2002) and Beghin, Bureau and Park (2003). Self-reliance
is also the building block of food security strategies in Mali and other West African
countries.5

Trade has a direct link with national food security, allowing direct food purchases
in the world market with incomes that are generated from exports. Global and regional
trade allows a country to buy or sell food in the world market, adjust its production to
terms of trade shocks, generate government revenues, and achieve overall economic
growth, all of which have direct or indirect impacts on the nutritional status of people in
the country. Figure 2.1 depicts the various components of a food security system at

various levels of aggregation.6

 

5 Nouve and Staatz (2003) review food security dynamics in West Africa and highlight the key linkages
between global and regional trade among West African countries.

6 This figure appears in various forms in several publications. Diaz-Bonilla et a1. (2000) adapts it from
Smith (1998), who adapted from UNICEF (1990) and Frankenberger et al. (1995). Also, FAO (2002)
exposes a similar framework.

12

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

REGIONAL g g
Regional & 8‘ GLOBAL % 51
Global Food TRADE a ,3
Availability
X ....«r‘wflfl 4.x" GrOWth, 11
National Net National Food Employmsntg Government ._1
Food Imports Production Income Distribution Revenues <Zt
\ / if! I if .‘I 0
National Food “kc \
Availability \\.
3" ,3. i". \‘l g
\ 4 .’i’ if \‘i.
.R.
Income to household \
members Incomes ,’ \
,.-' .-" : l ,_]
Household / Z
Food Access ' 1,! . Q
a
/ c:
, f E
/ ,3
Food Security Care Health Other Basic Needs I
and Non-necessities g)
Qua-household distribution ruD :
E 2‘
2 D
Q

 

Note: Solid lines represent actions (at individual, households, country and multi-country levels) that

directly affect national food availability, household food access, and individual food and nutritional
security. Dotted lines represent actions beyond the control of households or individuals that indirectly
influence household and individual food and nutritional security.

Figure 2.1: Conceptual Framework for Food Security (Based on Diaz-Bonilla et al.,

2000).

13

 

Government expenditures and investments in productivity-enhancing
infrastructures have both direct and indirect effects on national food production. This
feedback is shown in Figure 2.1, with arrows from “Government Revenues” to “National
Food Production” either directly or via the government financing of “Other Basic Need
and Non-Necessities”. These linkages are important, and are modeled as shocks to
government investments in the Malian computable general equilibritun model.

2.3.3. Measuring the Level of Food Security

There are several indicators of food security in the literature. This section first
discusses a selected number of these indicators, and identifies the indicators to be used to
evaluate the food security impacts of global agricultural trade reforms. Commonly used
indicators include: (i) per capita food production (in cereal equivalent or per food
category), (ii) per capita caloric availability, (iii) per capita protein availability, and (iv)
ratio of food imports in total exports. The frrst three indicators measure food security at
both national and household (group) levels, and the larger are the values of these
indicators the higher is the level of food security. The last indicator is exclusively a
national food security measure. Diaz-Bonilla et a1. (2000) and FAQ (2001), among
others, discuss in detail the appropriateness of these indicators as measures of national
and/or household food security.

(i) Per Capita Food Production

Per capita food production measures the level of food production in the economy.
It is the total national or household production divided by the relevant population.
National food production is usually calculated by food category and as an aggregate

index for all food products. The aggregate Food Production Index (FF!) is computed

14

following the F AO production index. For any given trade reform scenario .9, this index
can be defined as the ratio of value of food production in scenario 3 to the base food
production value. Algebraically, the ratio is expressed as follows:

c
P:

(2.2) Food Production Index (FPI) =°—g‘— *100

210.022
c=l

9.

where the subscript c indicates commodities and the superscripts and 0 represent prices

(P) and quantities (Q) in the base period and under scenario 3, respectively. C is the total
number of food products, and all PC0 and P: are expressed in real terms.

A FPI that is greater than 100 would indicate an average increase in the value
food production under s, relative to the base situation. An index value less than 100
would indicate an average reduction in the value of food production as a result of the
implementation of a trade reform scenario 5. A potential major limitation of this index is
that it can change in response to changes in relative prices (brought about by the trade
reforms), even if physical quantities of production remain unchanged or change in the
opposite direction. An index based on physical quantities would therefore be more useful
in describing changes in food production under various trade reform scenarios.

One of the criticisms of the WTO is that it shifts the mix of agricultural
production away from domestically consumed staple food towards production of export
crops, thus lowering domestic food availability. Also important to domestic food
availability are food imports, so that a more useful food security indicator would consider
changes to overall food availability that accounts for both domestic production and

imports. This is what is done in this study by focusing on changes in physical quantities

15

of national and household food consumption. Changes in consumer food demand depend
on changes in domestic food production and in food imports, and these two sources of
changes are accounted for in the Malian CGE model.
(ii) Per Capita Caloric Availability

This indicator characterizes the average consumption levels in a given country.
The indicator can be calculated for the country as a whole, and for individual household
groups. As a national (or household) average, caloric availability or dietary energy supply
(DES) may be viewed as an imperfect indicator of the state of individual food security.
But empirical evidence, such as Smith and Haddad (2000), suggest that there is a strong
correlation between this national (household) average and more individual-based
indicators of food security (e.g., anthropometric indicators of children’s nutritional
status). In particular, Smith and Haddad (2000) show that national caloric availability was
responsible for more than a quarter of reductions in child malnutrition in developing
countries over the period 1970-95.7

DES has been widely used in the literature as one of the main determinants of
national food availability (Smith and Haddad, 2000). It has also been widely used in
connection with health and nutrition-related studies, such as Anand and Ravallion (1993),
Subbarao and Raney (1995), Prichnett and Summers (1996), Frongillo, de Onis, and
Hanson (1997) and Osmani (1997). The DES level is often calculated from food
composition tables and a highly disaggregated food consumption basket. While a food

composition table is available for Mali, products in the Malian social accounting matrix

 

7 Women’s education and status accounted for nearly 55% of these reductions, whereas the health
environment of the household explained the remaining 19%. However, most of these effects were found to
be concave in time, with diminishing marginal contribution to reduction in child malnutrition over time.

16

 

are not sufficiently disaggregated to produce a reliable DES measure for the counfiy.
Thus, the DES indicator will not be used in this study.
(iii) Per Capita Protein Availability

Similar to the caloric availability (DES), per capita protein consumption indicates
the average levels of national or household protein consumption. According to Diaz-
Bonilla et a1. (2000), this indicator is usefirl in addressing traditional criticisms (e.g., by
von Braun et al.(l992), Smith (1998), and Bouis (2000)) that food consumption measures
tend to ignore protein and micronutrient consumption. Data on national micronutrient
consumption are generally said to be scarce (Smith and Haddad, 2000), but Mali could be
an exception. The Malian food composition table could allow a computation of changes
in the levels of micronutrient consumption under various trade reform scenarios. The
problem is once again the insufficient level of commodity disaggregation, making it
impossible to obtain accurate estimates of micronutrient levels, both in the base year and
under the trade reform scenarios.

This study will emphasize changes in physical quantities of protein-rich food
products consumed at both national and household levels. Evidence from Bouis (2000),
and reported in Diaz-Bonilla et a1. (2000), suggest that availability of animal proteins is
more directly correlated with measures of nutritional security than is availability of total
proteins. While themselves using the total per capita protein availability, Diaz-Bonilla et
a1. (2000) implicitly recommend the use of per capita availability of animal protein. Thus,
this study will approximate the changes in nutritional security using the average national

and household consumption of livestock and fish products.

17

(iv) Ratio of Total Exports to Food Imports

This indicator is often thought to be a measure of the capacity of a country to
finance its food imports using revenues from exports of goods and services. This measure
has been used, among others, by Diaz-Bonilla et a1. (2000). Also widely used is the
inverse of this ratio, that is, the percentage of food imports in total exports (e. g, F A0,
2001). Diaz-Bonilla et a1. (2000) argue that their measure, which they traced back to
earlier works such as Valdés and Konandreas (1981), is easier to interpret because it is
directly related to food security. However, the interpretation of the inverse ratio is equally
straightforward, the target being to achieve a ratio (food imports/total exports) that is as
small as possible.8 So the inverse ratio will be used in this study.

Whether food imports or total exports are put in the denominator of the ratio, it is
misleading to treat this ratio as a measure of a country’s capacity to import food, let alone
as a measure of food security. The ratio will decline when food imports fall faster than
total exports, which may result in reduction in the level of overall food consumption. This
point is clearly made in FAQ (2003). What is measured by the ratio is the capacity of the
country to pay for current food imports, and this has little or no relationship with changes
in overall food consumption. Therefore, the ratio needs to be interpreted in light of
observed changes in food consumption. This is what is done in this study.

It is clear that the food imports/total exports ratio is an imperfect food security
indicator. There is, however, a concern among developing nations and activist

movements that WTO agricultural trade reforms would increase food import dependency

 

3 Even Diaz-Bonilla et a1. (2000) proceed with their discussion of this indicator, using the (inverse) ratio of
food irrrport divided by total exports. For instance, they compared “A country. . .for which the total food bill
takes a larger percentage of total exports) with “a country. . .whose food bill takes only a small percentage
of its total exports” (p. 6).

18

in these countries, which is often equated to a situation of worsened food security. Thus,
computing the ratio and comparing it with expected changes in food consumption will
shed light on the tradeoffs between reducing food import dependency and increasing food
consumption in Mali.
(v) Measuring Changes in Food Consumption Levels

The change in food security is proxied by changes in the level of food
consumption between the base and the counterfactual scenarios (Equation 2):
(2.2) %Axi = 100* [mom - xs(p°.y°)1/xi(p°.y°)
where %Axi is the percentage change in the consumption of food item i between the
initial and new scenarios; x,(p,y) represents the demand of food product i; y° and y‘
represent the initial and new household incomes, respectively; and p0 and p5 are, as
defined earlier, the initial and new price vectors, respectively. One of the general
equilibrium relationships that is key in agrarian economies is the well-known endogenous
income effect. Because households are endowed with consumption goods it that they
produce, their income ys under each scenario depends on the equilibrium price vector p5.
Thus, as domestic prices change under each trade reform scenario, the net consumer
demand for goods it changes in response to direct price changes and indirect income
effects. It is argued later in the paper that while the direct price effects are important in
terms of welfare, they are less so in terms of food consumption. A substantial part of
households’ food demand comes from home consumption, which is very price-inelastic

in rural Mali.

19

I, , , , . . . _ - . . .. .
“’ 11"" " "‘ ' 'u ‘-"‘ -vn-*--" .o'IIov""'_ . Al'”"'" ‘ ", u v" . Iy‘l'” ‘
one. 2. Juuasaumarflumr Lam . ".mmfimfi, t,b-:,~:~:‘,2,", “- ..«feng'rg‘, "I: ...mt'n‘f, Pimp. . -'l‘::".“ I‘... . , , ‘
. _ , ,

 

2.4. Conclusion

This chapter discusses the concepts of welfare and food security and proposes
indicators to be used to measure welfare and food security impacts of global agricultural
trade reforms in Mali. Welfare is measured by equivalent variation at national and
households’ levels. Food security is characterized by physical quantities of food
consumed at the national and household levels, as well as by the ratio of food imports to
total exports. The comparison of the two food security indicators will elucidate tradeoffs

involved in increasing food consumption and increasing the level of food self-sufficiency.

20

CHAPTER III: DESIGN OF AGRICULTURAL TRADE REFORM SCENARIOS

3.1. Introduction

This chapter discusses the design of various scenarios of global agricultural trade
reforms that are relevant for Mali and other agriculture-dependent Sub-Saharan African
countries. The scenarios are designed based on expected changes in (i) the level of border
protection applied to Malian exports, (ii) world agricultural prices, and (iii) domestic
trade and fiscal policies in Mali. The first three sections of this chapter elaborate on each
of these sources of changes, and the last section summarizes and concludes the chapter.
3.2. Border Protection in the Malian Export Markets

The Malian export markets can be differentiated into West Africa and the rest of
the world (ROW). This section discusses how border protection is treated in these two
differentiated markets.
3.2.1. Border Protection in ROW

Commodity exports from Mali to the rest of the world (ROW) in the base year
(1997) were limited to three products, including two agricultural products (groundnuts
and cotton) and one mining product (gold). Thus, this study will assume that
counterfactual changes in border protection facing the Malian exports to ROW are
restricted to these three commodities. A further restriction will apply to gold, and this is
discussed in the next section. Since 1997, gold and cotton have contributed more than
80% of export receipts in Mali. Therefore, focusing on the three commodities will

capture most of the export activities in the country.

21

 

Agricultural and mining exports originating in Mali have benefited from
preferential access to developed countries’ markets. These benefits include the General
System of Preferences (GSP) in the Quad markets (European Union, the United States,
Canada and Japan). In addition, Mali has preferential access to the European Union’s
markets under the Lomé Convention (and the subsequent Cotonou Agreements) between
EU and the African, Caribbean and Pacific (ACP) countries. The EU, together with other
Quad countries, absorbs most of the Malian exports of primary products. But Mali is
increasingly directing its exports (particularly cotton) to other countries, such as China.

As a preference beneficiary country, Mali is concerned with potential erosions of
trade preferences following the implementation of the WTO agricultural trade reforms. In
terms of the magnitude of the preference margins, Hoekrnan, Michalopoulos and Winters
(2003) draw on existing studies and evaluate that these reforms would cut average tariffs
of preference commodities from 4.3 to 2.5 percentage points, which is equivalent to more
than a 41% erosion. Comparable estimates suggest that the preference margins extended
to the least-developed countries vary with countries and average 50% of the Most
Favored Nation (MFN) rates (IMF and World Bank, 2001; 2002). 9 Thus, a removal of
these preferences under the ongoing WTO reforms would amount to 50% erosion. This
magnitude of erosion is accounted for as one scenario of erosion of existing preferences
(EEP) on the Malian groundnuts and cotton exports. Being a traditional monetary unit,

gold enters most countries duty-free, and this is maintained in the study.

 

9 EU’s preferential margins tend to be the highest for Sub-Saharan African countries, although their
relevance is ofien mitigated by a combination of low MFN rates and restrictive rules of origins (IMF and
World Bank, 2001; 2002). This is evidenced by much lower preference rmrgins for sensitive products on
which high tariffs (tariff peaks) are imposed.

22

In addition to the traditional GSP and ACP-EU preferences, new preference
initiatives have emerged with the new millennium. These include the US African Growth
and Opportunity Act (AGOA) and the EU’s Everything But Arms (EBA). Preliminary
investigations of these initiatives suggest that they are unlikely to improve the already
liberalized access the least-developed countries (such as Mali) have to markets in the
Quad. For example, Brenton (2003) argues that for many poor countries, the current EU
trade preferences are of no real value because more than 99% of LDC exports are
concentrated in products for which the EU’s MFN external tariff is zero. The EU tariff
databaselo indicates that the Malian exports of cotton and groundnuts were qualified for a
duty-free access to the Union in 1997, which is the base year used for simulations in this
study. In addition, existing simulation-based studies show that the impact of the EBA
initiative is likely to be concentrated on a narrow set of sectors, in particular, sugar and
rice (Ianchovichina et al., 2001, UNCTAD 2001, Trueblood and Somwaru 2002). Mali
imports both sugar and rice, and is therefore unlikely to be affected by improved access
to export markets of these commodities. Similarly, AGOA does not offer any additional
market access opportunity for products Mali exports to ROW, such are cotton and
groundnuts. AGOA benefits seem to be concentrated in textile apparels (Hoekman,
Michalopoulos and Winters, 2003). Textiles do not constitute major exports from Mali to
ROW. Also, Nouve and Staatz (2003b) find no evidence of change in agricultural exports
from Sub-Saharan Africa to US in the post-AGOA period.

In light of the preceding, this study acknowledges the possibility that the Malian

groundnut and cotton (in addition to gold) exports would have effective duty-free access

 

'0 It is the TARIC database, available online at
http://europa.eu.int/comm/taxation_customs/dds/en/tarhome.htrn, viewed April 16, 2004.

23

to major export markets. Such an access would be equivalent to zero effective duties on
these commodities. The effective duty-free entry of Mali’s exports to ROW’s markets,
called duty free access (DF A), is also accounted for as a scenario in this study.

The implementation of both the 50% preferential access and the duty-free access
to export markets is based on Figure 3.1. The figure is drawn from the perspective of a
small country that faces an infinitely elastic world demand at the prevailing world prices.
Three border protection regimes are depicted in the figure. They correspond to the access
of Malian export to ROW under duty-free, preferential, and EEP regimes. For the
exporter, the infinitely elastic world demand at the prevailing world price represents a
schedule of net marginal benefit accruing from each additional unit of exports. This not
benefit is the highest when the exporting country has duty-free access to importers’
markets. From Figure 3.1, Malian exporters would receive the net price DFAPW for each
unit of exports, and the corresponding marginal benefit is represented by the vertical
distance O-DFAPw. When exports are subject to the EEP rate, tag), the net marginal
benefit drops to its lowest level (O-EEPPw), as the not price received by exporters falls to
EEPPw. Under preferential border protection regime, the net price paid to exporters (PPw)
and the corresponding net marginal benefit (O-PPw) will lie between the two extremes.

In general, the level of border protection in export markets is not treated explicitly
in applied general equilibrium models, and this study does not make an exception to this
tradition. Export market protection is often thought to be part of the world price of
commodities sold in a hypothetical global market. Thus, the net price received by
exporters is considered to have already taken into account the border protection level in

importing countries. In implementing border protection scenarios, this study will assume

24

that the prevailing conditions in the base year (1997) corresponded to some level of
preferential access that lied between the extreme DFA and EEP levels. This is because
Mali was a legitimate beneficiary of trade preferences in 1997, which would make the
average protection facing Malian exports less than the EEP protection. But because
eligibility for preferences does not necessarily translate into utilization of these
preferences, the effective access level would be different from a duty-free access, even in
regions such as EU where zero-duties were recorded in 1997.11 In terms of Figure 3.1,
preferential export prices in the base year would correspond to PPw. This price level will

be obtained by calibrating the computable general model to the base year data.

 

 

 

 

 

. 1)
Net Price
ROW’s demand at
DFA rates
DFAPw
‘ A ROW’s demand at

 

 

 

ROW’s demand at
*
TEE? EEPPW EEP rates

 

 

 

preferential duties
PPw aim /

EEPPW

 

 

 

 

Quantity 7

 

 

 

Notes: DFAPw: World prices under a duty-free export regime; PPw: World prices under a preferential
export regime; EEPPW: World prices under an export regime of erosion of existing preferences; to”: level
of protection when existing preferences are effectively eroded; and t9: average preference rate.

Figure 3.1: Three Border Protection Regimes with Infinitely Elastic World Demand

 

" This is because there may be other duties besides tariffs (e.g., custom regulatory duties) that are imposed
on the exports. In addition, exports may face non-tariff barriers, which will usually have a tariff equivalent.
There is also the possibility for the exporting country to incur costs associated with conrpliance with
technical regulations, and these costs may conceptually be treated as tariffs. So even if the EU data show
zero duty on Malian exports of groundnuts and cotton in 1997, it is still possible that there were hidden
tariffs. Under the DFA regime, both the apparent and hidden tariffs are assumed to be eliminated.

25

The implementation of the two remaining scenarios of border protection in
Malian export markets requires knowledge of average EEP tariffs applied to commodities
of export relevance to Mali (tggp in Figure 3.1). The EEP tariffs are applied to all
countries having the most-favored nations (MFN) status in the importing country. Given
previously cited evidence, average preferential tariffs (tp) will be set to half the EEP or
MFN tariff level (tp = 0.5tggp). It follows from Figures 3.1 that DFAPW = (1 + t,,)"'PPw =
(l + 0.5t55p)*PPw. Similarly, DFAPw = (l + tEEp)*EEPPw, which implies that EEPPW =
DFAPw/(l + tEEp) = [(1 + 0.5tEEp)/(I + tEEp)]* PPw. The price level DFAPw defines the
duty-free access scenario, whereas EEPPw defines the EEP access scenario. Both
scenarios depend on the calibrated level of preferential access price (PPw) and on the
MFN tariff rate, I551).

The question that arises is how to determine tEEp. The average effective tariffs
faced by Malian cotton and groundnuts exports to ROW is not known, due to the lack of
data documenting the tariffs effectively paid by Malian exporters in various export
markets. A trade-weighted approximation could have been used, but country-specific
information on the destinations of these exports was not available at the Malian national
statistical office, the Direction Nationale de la Statistique et de I’Informatique (DNSI).
Though these two commodities were in theory eligible for duty-free entry to markets in
many developed countries (under the GSP and ACP-EU preferences), effective access to
these markets may not be duty-free, due to technical restrictions and other regulations.
Given the lack of data on effective rates applied to the Malian commodity exports
markets, and due to the lack of information required to approximate these rates, this study

relies on simple averages of MFN tariffs applied to imports of these commodities in

26

ajor markets. The database of the Global Trade Analysis Project (GTAP) at Purdue
riversity (McDougall and Elbehri, 1998) documents these rates for several regions in
3 world.

Table 3.1 presents the average MFN tariffs on groundnuts and cotton in the Quad
anada, EU, Japan, and USA) and in the rest of the world (ROW) for the mid-nineties.
1e table shows varying rates per country/region and commodities. Cotton appears to be
bj ect to the highest MFN rates, due to high ad valorem rates in Japan (42%), EU
9.6)% and ROW (23.7%). MFN rates on groundnuts lies between the relatively high
:es on cotton and the close-to-zero rates on gold.

rble 3.1: Average Import Tariffs per Commodities and Importers

 

 

 

 

 

Country and Region Mean Mean

mmodity Canada EU Japan USA ROW (Quad (All)
only)

ar (for tariff data) 1996 n.a. 1996 1996 n.a. Mid-nineties

getables, fruits and nuts (groundnuts) 0.2 2.1 5.9 1.3 26.1 2.4 7.1

.[nt-based fibers (cotton) 3.2 39.6 42.0 0.1 23.7 21.2 « 21.7

 

tes: EU is European Union (excluding Germany and Northern European countries); ROW: rest of the
rld in GTAP 4 database; n.a. not available.

lurce: Based on GTAP 4 Database, Table l (McDougall and Elbehri, 1998).

The average MFN rates used in this study could be determined in two ways,
own in the last two columns of Table 3.1. The first approach assumes that MFN rates
3 equivalent to a simple mean of the Quad rates. This yields average MFN tariffs of
1% on groundnuts and 21.2% on cotton. In the second approach, the MFN tariffs are
:asured as a simple average of the rates in the five countries/regions (Quad plus ROW).
is results in MFN tariffs of 7.1% on groundnuts and 21.7% on cotton. As indicated
rlier, Mali directed most of its commodity exports (except cotton) to Quad countries, so

3 Quad mean MFN rates will be used in the simulations to determine how the MFN and

27

 

duty-free scenarios would deviate from the (calibrated) preferential scenario.12 The
average MFN tariff rate on cotton is nearly identical under the two approaches, so
choosing the Quad mean tariffs will not discount the fact that Mali increasingly exports
its cotton to Asia, including China.
3.2.2. Border Protection in West Africa

In addition to changes in border protection facing the Malian exports to ROW,
there are also changes in the protection level facing the Malian exports to the West
Afiican region. These changes are made possible by ongoing inter-government efforts to
liberalize regional markets within West Afiica. Such efforts are supported by two
overlapping regional integration projects in West Afiica. The first is the Economic
Community of West African States (ECOWAS), which aims at creating a free trade area
among 15 West African countries, with a potential to evolve into an economic, monetary,
and political union in the image of the European Union. The second field of integration is
the West Afiican Economic and Monetary Union (WAEMU), which creates an effective
economic and monetary union as well as a customs union among seven francophone and
one lusophone (Guinea-Bissau) countries in West Africa. WEAMU has developed a
common external tariff (CET) regime that affects non-member imports, whereas within-
WAEMU imports are essentially free of customs duties. The WAEMU’s CET regime has

been in force since January 2000. Thus, the Malian exports to regional markets over the

 

‘2 Recall that DFAPW = (1 + 0.5tm.;p)“'PPW and EEpr =[(1 + 0.5t55p)/(1 + tEEp)]"'PP,,, where tggp is the EEP
or MFN average tariffs assumed above; DFAPW, EEPPw and PPw represent the export prices under DFA,
EEP, and preferential market access regimes, respectively. Consider for example the Quad mean EEP tariff
of 2.4% on groundnuts. The deviations from the calibrated preferential regime would be 1.012*PP,, for the
duty free access and 0.988"‘PPw for access when existing preferences are eroded (EEP). In other words, the
marginal revenue from each additional unit of duty-free groundnut exports would be 1.012 times the
marginal revenue under the status quo (corresponding to the prevailing regime in the base year). This
marginal revenue drops to 0.988 times the base year’s value, when the trade regime switches to EEP tariffs.
Similar computation can be made for groundnut under the alternative EEP protection, and for cotton and
gold under the two assumed EEP protection in Table 3.1.

28

simulation horizon will be considered as entering these markets free of duties.
Commodities affected by such measures, based on the Malian social accounting matrix
(Chapter V), include grains (millet, sorghum, maize and rice), beans, livestock products
(cattle, sheep and goats, poultry, and hides and skins), fishery products (fresh and dry
fish), and other agricultural products.

In summary, there are three border protection regimes under which Malian
exports enter markets in West Africa and the rest of the world. For exports to both West
Africa and ROW, the base scenario is calibrated to the country’s data in 1997. The two
other scenarios correspond to duty-free access (DFA) and preference erosion (EEP)
regimes.

3.3. Projected Changes in World Agricultural Prices

This study is based on the premise that the implementation of WTO agricultural
trade reforms will change world prices of agricultural commodities. The extent of the
change is unknown, but a number of world commodity models routinely publish
estimates of future trends in commodity prices. These estimates are based on a wide
range of assumptions regarding expected macroeconomic and trade policy changes in
major producing, consuming and trading countries, and assumed changes in other
exogenous factors such as agricultural productivity and weather conditions. Projected
future prices also vary with the theoretical framework used in projection models, which
in turn vary in commodity and country coverage. This study considers four different price
projections that are available from public sources. These include models from (i) the
Food and Agricultural Policy Research Institute (F APRI), (ii) the International Food

Policy Research Institute, assuming current commercial policies will remain unchanged

29

(IFPRI-l), (iii) IFPRI with full trade liberalization assumption (IFPRI-2), and the
Organization of Econorrric Cooperation and Development (OECD).

The rest of this section highlights the key features of each of these four
projections and summarizes the expected changes in prices over the projection period.
The degree of details available on these models varies greatly and most of these models,
such as FAPRI and OECD, are constantly updated to account for new developments in
world economic conditions. Thus, no attempt will be made in this study to present a
detailed description of these models. Studies such as Tongeren and Meijl (1999) offer
detailed technical discussions of many applied international trade model used in
agriculture and related resources. The four price projection models retained for this study
are also reviewed in Tongeren and Meijl (1999), although these models may have
evolved in terms of expanded regional and commodity coverage since the review was
completed in 1999. One of the common threads linking all the projections models
considered in this study is that they represent the best conjecture stakeholders can make
about the future, given a set of assumptions regarding expected dynamics in economic
and policy variables over a medium run. One exception to the “medium-run” assumption
is IFPRI’s long-rim projections to 2020. But since such projections do not diverge greatly
from what they would be in 2010 (Rosegrant et al., 2001), they are included as plausible
scenarios of price changes over the medium run.13 It is an expectation of this study that
by focusing on four different price projection models with different conjectures about the
future, one can effectively capture an array of future possibilities, in terms of- the

direction and magnitude of changes in world commodity prices.

 

'3 See grains baseline projections in Rosegrant et a1. (2001), Figures 5.1-5.4, pp. 104-106.

30

 

 

 

3.3.1. FAPRI Price Projection Model

The Food and Agriculture Policy Research Institute (F APRI) projects the state of
US and world commodity markets ten years into the future. Projection assumptions are
spelled out in detail and are available online. 14 These assumptions relate to the expected
general macroeconomic conditions, agricultural policies, the weather, and technological
change. In general, the projections assume a normal continuation of the comnritted trade
and agricultural policies in the United and other third countries. The FAPRI projections
also assume average climatic conditions and technological change that are based on
historical rates. A summary of key projection assumptions is presented later in the
section. By assuming average conditions in the world commodity markets, and allowing
for the continuation but not the completion of agricultural and trade policy reforms,
FAPRI baseline projections may be viewed as representing an intermediate trade
liberalization scenario.

The FAPRI model is a multi-country, multi-market partial equilibrium model of
world agricultural markets. The model generates initial baseline projections, which are
then reviewed and amended by a panel of experts. The panel includes extension
specialists and industry experts, as well as scholars from various US agencies,
international organizations and academia. As mentioned earlier, the FAPRI model is
being upgraded, with the aim to expand both country and commodity coverage. Thus,
existing documentation of the technical description of the model are out of date, though

the basic structure of the model is available in Tongeren and Meijl (1999).

 

" Projection assumptions (and other information) on the 2004 outlook are available online at the following
address: www.fapri.iastate.edu/outlook2004/ (viewed, May 3, 2004).

31

Several commodities covered by the FAPRI model are relevant to the Malian
agricultural sector. These include: wheat, rice, sorghum, maize, cotton, sugar (component
of the agro-industrial product category), beef, poultry and milk.

3.3.2. IFPRI Baseline Projection: The IMPACT Model

The International Model for Policy Analysis of Agricultural Commodities and Trade
(IMPACT) is IFPRI’s simulation laboratory for alternative global food demand, supply
and trade scenarios. It is a multi-region and multi-product, competitive markets, partial
equilibrium model, covering the quasi-totality of world food production and
consumption. A summary discussion of the IMPACT model is available in Rosegrant et
a1. (2001). There are two IFPRI price projections, one assuming the status quo in terms of
trade and agricultural policies (IFPRI-l) and the other assuming a firll liberalization of
the world commodity markets (details are available below, and in a summary table later
in this section).

Nine out of the 16 IMPACT commodities are relevant for the analysis undertaken
in this study: wheat, maize, other coarse grains (sorghum and millet), rice, beef (cattle),
sheep and goats, chicken, eggs and milk. Baseline assumptions in the IMPACT model
generally refer to modest population and income growth. The model also hypothesizes a
structural shift in food consumption from main staples to more processed and high-value
livestock products, due to a combined effect of urbanization and improved incomes.
These changes are accounted for in assumed income and price elasticities in the IMPACT
model.

Like in the other projection models, world commodity prices are endogenously

determined. Trade and domestic agricultural policies are modeled in the form of

32

producers and consumers subsidy equivalents, expressed as a percentage of world prices.
These policies are maintained at their base level in the IFPRI-l (status quo) scenario. The
producer and consumer subsidy equivalents are effectively set to zero in the IFPRI-2 (full
liberalization) scenario. The IMPACT model also accounts explicitly for expected change
in agricultural productivity, given historical changes in the use of productivity-enhancing
inputs and in the level of public investments in agricultural research and rural
infrastructures.
3.3.3. OECD Baseline Projection: The AgLink Model

Projections are made to 2008 and represent what the Organization calls “a
plausible medium-term future for the markets of key commodities” (OECD, 2003. p. 3).
The projections are said to reflect the combined knowledge and expertise of experts in
OECD member and non-member countries. The projection process is highly interactive,
involving several exchanges between OECD member countries and the Secretariat
(Uebayashi, 2000). Like in previous baseline projections, OECD projections are based on
several assumptions regarding: (i) the existing or expected agricultural and trade policies
in OECD and major non-OECD countries, (ii) the existing and expected evolution in
underlying macroeconomic environment, and (iii) other developments in members and
non-member countries. Like the FAPRI scenario, the OECD projections may be viewed
as representing some level of increased (but incomplete) liberalization of the world
commodity markets.

OECD baseline projections are based on the AgLink model. As in previous cases,
AgLink is a multi-country, multi-commodity partial equilibrium model, with focus on

temperate agricultural products relevant to OECD countries. Prices are the equilibrating

33

variables in the model, allowing the global market of each commodity to clear.
Commodities covered in the OECD projections that are relevant to the Malian study
include: wheat, coarse grains (maize, millet and sorghum), rice, sugar and vegetable oil
(key components of agro-industrial products), beef, poultry, sheep and goat and milk.

OECD analysts acknowledge the existence of uncertainties associated with any
outlook of the agricultural markets. Thus, the Organization’s projections may be subject
to some important uncertainties in the short run. OECD (2003) identifies these
uncertainties to include: (i) unfavorable geopolitical environment of low growth in
Europe and Japan, (ii) the growing twin deficits (budget and trade) in the US, and (iii) the
slow down of global agricultural trade liberalization with the maintenance of high level
of trade distorting support and protection to agriculture. All these factors are expected to
reduce trade volume and keep agricultural prices at lower levels, with a possibility for
increasing price volatility. Though OECD did not account for these uncertainties in their
2003-2008 projections, they may be reflected to some extents in IFPRI-1’s conservative
projections discussed earlier.

3.3.4. Summary of Model Assumptions and Projected Changes in World
Commodity Prices

Table 3.2 highlights key differences between the four scenarios of change in
world commodity prices. Underlying assumptions in the FAPRI and OECD price
projection models are very similar. Both models are based on the expectations that the
world economy at projection horizon (i.e., end of the 2000 decade) will grow on average
by more than 3% annually, which is slightly higher than the average growth rate of the
world economy in the base year (1996-1998). The two models also assume more than 1%

grth in the world population in 2007-2009, about 20% lower than the population

34

growth rate in the base period. The two IFPRI scenarios also assume nearly 3% annual
economic growth and 1% growth in the population, but these rates remained unchanged
between the base and projection periods. In addition, the rate of urbanization is expected
to accelerate over time, and this would shift consumption toward livestock and high-
value food products. The accelerated urbanization assumption is not explicit under the
FAPRI and OECD scenarios, but the two models expect that major world currencies will
appreciate against the US dollar (the IFPRI models have no explicit assumptions on the
exchange rate). All the four models assume that the climatic conditions will remain at
their historical averages, although only the F APRI model states this explicitly.
Agricultural and trade policies assumed under the FAPRI and OECD are also
substantially similar. The two models assumed that current policies will remain in place
over the projection horizon. In particular, they assume no change to the provisions of the
2002 US farm bill, the 2003 reforms of the European Union (EU) Common Agricultural
Policy (CAP), the changes brought about by the enlargement EU to new members, and
agricultural policies in other parts of the world. Regarding trade policies, the models are
based on the trade reforms committed to by countries under the WTO’s Uruguay Round
Agreements on Agriculture (URAA). The commitments relate to limitations in the use
export subsidies, expansion of tariff rate quotas, reduction of import duties, and cuts in
domestic support programs. The main difference between the F APRI and OECD models
lies in the difference in commodity coverage and the magnitude of committed policies.
Also, the OECD model explicitly assumes increased production efficiency over time, but
this is not explicit in the FAPRI model. The two IF PRI models also assume that farm

productivity will grow at a declining rate over the projection period.

35

Table 3.2: Key Assumptions in the Price Projection Models

 

 

 

 

 

 

 

 

 

 

Criteria Key Projection Assunrptions
FAPRI OECD IFPRI-1* IFPRI-2
World economic growth
2007-09 annual average
growth rate 3.2% 3.2% 2.9% Same as
Change in annual growth rate in IFPRI-1
From 1996-98 to 2007-09 + 4.2% + 12.2% 0.0%
World population growth
2007-O9 annual average
grth rate 1.1% 1.3% 1.0% Same as
Change in annual growth rate in IFPRI-1
from 1996-98 to 2007-09 - 19.7% - 24.8% 0.0%
Urbanization Not explicit Not explicit Accelerated Same as
Urbanization in IFPRI-l

Currencies Most currencies will Same as in Not explicit Same as

appreciate against USD FAPRI in IFPRI-1

Climatic conditions Average Not explicit Not explicit Not

explicit

Agricultural policies Include: Similar to Maintain Producer

- Provisions of the 2002 US FAPRI’s producer and and

farm bill are maintained agricultural consumer consumer
(provisions set to expire in policies subsidy subsidy
2007 are extended through equivalent equivalent
projection period); (PSE and are

- Provisions of the 2003 CSE) of the eliminated
reform of EU common base year

agricultural policy; also (1997)

account for EU

enlargement;

- Maintain other ag policies

WTO Agreements (Uruguay In developed countries, Continuation No explicit Same as

Round Agreements on maintain the 2000 level of of existing reference to in IFPRI-l

Agriculture) current policies (export policy WTO

subsidy limits, tariff rate reforms and agreements;
quota expansion, import URAA however,
duty and domestic support commitments trade policy
reductions); in developing reforms are
countries, continue implicit in
inrplementation of WTO the PSE and
commitments through 2004, CSE

and hold them fixed measures.
afterward.

Productivity Not explicit Increased Positive, but Same as
farm declining in IF PRl-l
productivity crop yield

growth

Reference to further details FAPRI (2004) OECD Rosegrant et Rosegrant

(2004) al. (2001) et al.
(2001)

 

" The GDP growth rates for the two IF PR1 scenarios indicate the 1997-2025 yearly averages. There is no
change between the base year (1997) and the projection period (2025). The population is projected to grow
from 5,788 in 1997 to 7,456 million in 2025, an equivalent of 1% yearly increase.

Source: Based on references indicated in the table.

36

Regarding the two IFPRI models, agricultural policies are captured through the
aggregate producer and consumer subsidy equivalents. These subsidies, which create a
price wedge between domestic and world prices, are completely eliminated in the IFPRI-
2 full liberalization scenario. IFPRI trade policies are less explicit in terms of WTO trade
reforms, but they are implicit in the implementation of the agricultural policy reforms
(i.e., maintaining or eliminating the producers and consumer subsidies).

Projected price changes from the five models are compiled in Table 3.3. The
changes are expressed with respect to the base year (1997). The projection year is 2009
for FAPRI, 2008 for OECD, and 2020 for IFPRI projections. As explained earlier, price
projections to 2020 in IFPRI models do not deviate substantially fi'om what these
projections would be around 2010. For both the base (1997) year a three-year average
price was used to reduce year-to-year variability in the base year prices. To be consistent
with the base year average prices, a three-year average was also used for the projected
prices. Price variability was not an issue with the projections, which assmne
parsimonious changes in underlying economic conditions. Even though the models do not
assume any change in price variability with the implementation of the trade accords,
policy makers in Africa do continue to be concerned about the volatility of world prices
and are concerned about how the trade accords may affect that volatility. Unfortunately,
that issue cannot be addressed with these models.

Table 3.3 suggests that most prices, particularly grain prices, are projected to
decline. The magnitude of projected changes varies depending on commodities and
projection models, but the direction of change for most commodities was consistent

across projection sources.

37

Table 3.3: Projected Changes in World Commodity Prices with Four Projection
Models (percent changbetween 1996-98 and 2007-2009*)

 

 

 

Commodities Price Projection Models
FAPRI OECD IFPRI-1 IFPRI-2
Cattle 9.7 14.6 19.8 40.7
Chicken 1.3 7.5 ..
Cotton -20.3 .. .. ..
Maize -24.3 0.6 -23.4 -l6.7
Milk ~29.9 -0.9 -3.2 -3.8
Millet -16.6 -31.1 -25.5
Rice -25.7 -1 8.7 -25.1 -14.7
Sheep and Goat .. .. 17.4 39.6
Sorghum -16.6 .. -3 l .1 -25 .5
Wheat -18. 4 -3. 6 -31.2 -25.6
_groeindustrial products -25. 9 -23. 8 ..

 

Note: * Changes are expressed with respect 2020 (IFPRI) and the 2007-08 average (OECD)

Source: Based on projection data from various sources (discussed m the text and available online for each
of the three institutions that publish price projection data).

Source: Based on references indicated in the table.

Cereal prices are projected to fall under all scenarios, except in the OECD
scenario, where the price of coarse grains will remain practically unchanged. Livestock
prices would increase under all scenarios, with higher increase in the IFPRI (particularly
the IFPRI-2) scenarios. Cotton prices are expected to fall by 20% under the FAPRI
scenario. Like the border protection regimes (Section 3.2) and the govemment policies
(next section), each of the four prices projections would represent a scenario of changes
in world commodity prices.

3.4. Government’s Trade and Fiscal Policies

Several trade and fiscal policy instruments are available to the Malian
government. Among them, import tariffs constitute the single major trade policy
instrument available not only to the Malian government, but also to ahnost all
governments in the developing world. In addition, the Malian government may use export
taxes and export bans on selected commodities. The main fiscal policy instrument is the

value-added or domestic sales tax, which is applied to both imported and domestically

produced commodities. Other duties and taxes are also possible. These include

38

production factor taxes and direct taxes on domestic institutions. Finally, government
policies may involve changes in the level of public investments. As discussed in Chapter
II (Figure 2.1), such investments are critical in financing public goods necessary to
improve agricultural productivity.

Although one can simulate the effects of changes in indirect domestic sales taxes,
their application in the Malian context is not realistic because most of the sales of
domestic output take place in hard-to-tax channels of the economy. Also possible are
simulations involving direct taxes on factors and domestic institutions, such as
households and firms. But these taxes are less relevant for this study, which is primarily
concerned with commodity-based policies. Thus, the study will focus on three
instruments: changing import tariffs, imposing export bans on cereals, and changing the
level of public investments. Changes in import tariffs are a direct result of ongoing trade
reforms in West Africa and the rest of the world. Cereal exports may be banned, as these
bans are politically attractive in Sahelian countries, such as Mali, that are highly prone to
sudden production shortages due to drought. Imposing export taxes on cereals is not
explicitly examined in this study because the idea of such taxes would be inconsistent
with the Malian Cereal Market Reform Program (PRMC). The PRMC is a nationwide

initiative aimed at improving farmers’ incomes thrOugh a strengthening of the workings

39

of cereal markets in Mali.15 Finally, changes in productive public investments will affect
production possibilities, and thereby incomes and food security in the country.
Import tariffs

Since January 2000, Mali has switched to a common external tariff (CET) system.
The CET is implemented within a customs, monetary and economic union that includes
Mali and 7 other West African countries that are members of the West African Economic
and Monetary Union (Benin, Burkina Faso, C6te d’lvoire, Guinea-Bissau, Niger,
Senegal, and Togo).

The CET regime harmonizes the rate of tariffs applied to extra-WAEMU imports,
while eliminating intra-WAEMU tariffs. Imports from the rest of the world are classified
into four categories, denoted 0, l, 2 and 3. Tariff rates of 0%, 5%, 10% and 20%
respectively, apply on each of these four categories. The average CET rates applied to
imported goods and services considered in this study are different from the four rates
indicated above, due mainly to differences in commodity aggregation.

Decaluwé, Dissou and Robichaud (2000) computed the average CET rates
applicable to non-WAEMU imports in Mali, based on insights from the expert’s

commission of the union. A summary of these tariffs is shown in Table 3.4. The original

 

'5 In a recent review Dembélé, Staatz and Weber (2003) show that the PRMC had led to major
improvements in farmers’ incomes in Mali. Imposing export taxes would most likely erode the benefits
derived from improved market efficiency associated with the program since the 19805. Thus, the Malian
government is unlikely to impose such a regressive tax, which, besides the economic costs, could also hear
high political costs both nationally (with the farmers’ constituency) and internationally (with the donor
community, which has lent a tremendous support to the Malian market policy reforms). Some may argue
that banning cereal exports is equally regressive, and that such a scenario does not constitute a plausible
policy choice in Mali over the next decade. While this argument has some merits, this study takes the stand
that cereal exports may be banned in periods of national grain shortage following adverse climatic shocks.
The ban has the clear objective to redirect all exports towards domestic consumption, which is the desired
goal of the government in period of shortage. Conversely, grain export taxes will still maintain some level
of exports, unless the tax is set at a level high enough to make exports economically not profitable
compared to selling the products in the domestic markets. Such a prohibitive tax would, therefore, be
equivalent to a ban on exports.

4o

figures appear by broad commodity groups, such as food crops, industrial agriculture,
livestock, fish and forestry, mining, food and textile industries, and public utilities. Each
of these broad categories corresponds to one or several commodities covered in this
study. For example, the CET on food crops imports was 8.7%. This rate was maintained
for all commodities produced from food cropping activities and which were traded
outside West Afiica (ECOWAS) in 1997. Based on the Malian Social Accounting Matrix
(see Chapter V), eligible products include maize, roots and tubers (such as potatoes, yarn
and cassava), beans, other crops (mainly fruits and vegetables), and rice.

Table 3.4 Common External Tariffs on Imports from the Rest of the World (%)

 

Commodities Common External Tariff (CET)
Maize 8.7
Roots and tubers (potatoes, yams, cassava) 8.7
Beans 8.7
Other crops 8.7
Rice 8.7
Tobacco 5.1
Wheat 5.1
Eggs 5.0
Milk 5.0
Firewood 10.1
Furniture wood 10.1
Fresh fish 10.1
Agro-industry & beverages 14.7
Textile 19.0
Other manufacturing goods 12.9
Electricity, water and energy 8.7
Transport & telecom services 8.7
Bank and insurance services" 15.0

 

Notes: CET rates were available for broad categories of commodities, such as food crops (maize, roots and
tubers, beans, other crops and rice), industrial agriculture (tobacco and wheat), livestock (eggs and milk),
fishery and forest products (woods and fish), food industry (agro-industry and beverages), textile industry
(textile), metal industry (construction and public works), public utility (electricity, water and energy, and
transport and teleconnnunication), and other industry (other manufacturing goods).

" Rate taken from IMF (2004)

Source: Adapted from Decaluwé, Dissou and Robichaud (2000)
The CET reform scenario is implemented in this study by setting the rates on
relevant non-West African imports to those indicated in Table 3.4. These rates are set to

zero on West African imports. With one external sector, the Malian CGE model uses an

41

aggregate CET tariff, which is computed as a weighted average of the West Afiican and
non-West African tariffs under the CET regime (the weight being the respective shares of
West African and non-West Afiican imports in total imports.) The CET simulation is run
by substituting the aggregated import tariffs to the base tariffs that calibrated from the
Malian data.

The CET rates available in Decaluwé, Dissou and Robichaud (2000) are
exclusively based on trade activities between Mali and other WAEMU members. In this
study, the definition of “West Africa” is not restrictive to WEAMU. Instead it refers to
the larger group of the 15 country members of the Economic Community of West African
States (ECOWAS). The WEAMU CET is assumed to be representative of the entire West
African region.

Cereals Export Ban

The Malian data suggest that there were exports of millet, sorghum and maize to
West Africa in the base year. One of government trade policy options is to ban these
exports. One is interested in knowing how such a ban would affect welfare and food
security in the country. The question is investigated by implementing a scenario of
cereals export ban, in which exports of millet, sorghum and maize are effectively set to
zero.

Change in Public Investment

A final possible action of the government is to allow for changes in public
investments. As discussed earlier, such investments may improve agricultural
productivity through a possible reduction in the prices of intermediate inputs. The SAM

(Chapter V) shows that food production activities consume some government services

42

(such as agricultural research and extension) as intermediate inputs. Thus, increasing
public investments in these services would positively affect their intermediate use in food
production activities.

The implementation of the change in public investments scenario is based on real
investment data from the World Development Indicators (WDI) (World Bank, 2004). The
WDI data indicate that real investments in Mali increased on average by 5% over the
period 1990-2002. A cumulative increase between the base year (1997) and the
projection year (end of the 2000 decade) may be assumed to be approximately 50%. This
rate is therefore used in this study to investigate the effects of potential change in public
investments on welfare and food security in Mali. Given the general equilibrium
framework used in this study, the increased public investments will be generated through
either increased government savings, or increased foreign transfers, or both. Details on
this type of macroeconomic closure are discussed in Chapter IV on the computable
general equilibrium model.

From the Malian social accounting matrix (Chapter V), public investments were
made in four goods (other manufacturing goods; buildings and public works; transport
and telecommunication; and nontraded services). Since there were also private
investments in the first two goods, the 50% increases in public investments will
materialize into lesser increase in total investments on these two goods (22.5% for other
manufacturing goods, and 15.1% for buildings and public works). The increased
investments will be fully transmitted to transport and telecommunication and non-traded

services into which private institutions did not invest in the base year, according to data

43

in the Malian social accounting matrix. The various magnitudes of changes in public
investments will be accounted for in the simulation.
3.5. Scenarios Summary and Conclusion

All the scenarios discussed in this chapter are presented in Table 3.5. In addition
to the base scenarios, there are four possible changes in the border protection facing key
Malian exports; four possibilities of change in world agricultural prices; and three
possible scenarios of government trade and fiscal policies. The combination of the three
sources of changes results in 48 possible reform scenarios, which is clearly intractable.
This number can double or even triple, if one takes into account the usual sensitivity
analyses performed with computable general equilibrium models. It appears that there is
a need to trim the number of scenarios, so as to facilitate the communication of the
simulation results. One obvious way to reduce the scenarios is to focus on only two
(instead of four) scenarios of border protections, one for the duty-free access (DFA) and
the other for the erosion of existing preferences (EEP) market access regimes. This is
because the difference between the “Qua ” and “All” DFA and EEP scenarios lies in the
assumed level of the EEP (or MFN) tariffs. Using the “Quad” or “All” EEP rates would
change the quantitative estimates, but not the qualitative effects. In any case, quantitative
estimates depend on many other factors (e.g., model parameters), so the emphasis should
be put on qualitative effects. These effects will be captured a single rate, in this instance
the “Quad” EEP tariff rates for DFA and EEP market access regimes.

Another way to trim the number of scenarios is to focus on individual scenarios
and to compare the results with the base year scenario. Focusing on the individual

scenarios will not only reduce the number of simulations to nine (EEP, DFA, FAPRI,

IFPRI-l, IFPRI-2, OECD, CET, BAN, and GINV), it will also allow cross-scenario
comparisons and tests of some key hypotheses regarding trade reforms in Mali. At least
four comparisons can be identified.

Table 3.5: Summary of Trade Reform Scenarios

 

 

Changes in Border Protection in Exports Changes in World Prices Changes in Trade and Fiscal
Markets Policies in Mali

1. Base (calibrated) 1. Base (calibrated) 1. Base (calibrated)

2. Duty-free (based on mean EEP, Quad) 2. FAPRI 2. CET

3. Duty-flee (based on mean EEP, All) 3. IFPRI-l 3. Export bans

4. EEP tariffs (using mean EEP, Quad) 4. IFPRI-2 4. Increase in public

5. EEP tariffs (using mean EEP, All) 5. OECD investments (50%)

 

Source: Based on the discussions in Sections 3.2, 3.3 and 3.4 of the present chapter.

First, the base scenario, which calibrated from the Malian data, roughly represents
a situation of no implementation of WTO trade reforms. Second, the base scenario may
be compared with prices changes scenarios under partial liberalization of world
agricultural markets (F APRI and OECD) or firll liberalization of these markets (IFPRI-2).
These two comparisons would inform on the impacts of the so-called North-South trade
expansion on welfare and food security in Mali. This is because trade reforms under these
price change scenarios would originate from the rest of the world, but the changes in
world commodities prices will be transmitted to the Malian economy (Mali being a
country in the South). Third, the base scenario may also be compared to the CET regime,
which will inform on the potential impacts of South-South trade liberalization. Fourth,
the impacts of South-South trade liberalization (the CET regime) may be compared with
the impacts of the North-South trade expansion, whether the expansion comes through a
partial (FAPRI and OECD) or a full trade liberalization of the world commodity markets.
A test of the impacts of a combined South-South and North-South trade liberalization is

also possible.

45

As indicated earlier, it is impossible to exhaust all possible scenarios, so the study
has emphasized the use of the most plausible scenarios among a nearly infinite number of
possibilities. There is tremendous room for adding more scenarios, particularly as they
are tailored to specific needs. For example, many more world commodity price projection
models are available in the literature (for a review, see Tongeren and Meijl, 1999). Also,
we have selected a limited number of government trade and investment policies. While
the selected policies capture plausible government choices, alternative policies remain
possible. For example, domestic sales taxes/subsidies can potentially play an important
role in the Malian economy. They can be used to strike a balance between changes in
consumer and producer prices as world commodity prices change. '6

The reliability of trade reform scenarios presented in the chapter depends on the
reliability of assumed changes in border protection and government policies, and in
projected changes in world agricultural prices. All price projections are based on the best
conjecture experts can make about the future. But not all future outcomes can be
projected. Thus, the scenarios will still be characterized by some elements of uncertainty.
Nevertheless, they will help see clearer through the future, even if they do not offer a

perfect image of that future.

 

'6 This dual role of commodity (food) prices is the source classical “food price dilemma” (Timrner, Falcon
and Pearson, 1983). The dilemma arises when a single price instrument is used to try to secure
remunerative revenues to producers and maintain affordable prices for consumers. Facing falling
commodity prices (e.g., for cereals), the government may use indirect taxes on domestic sales to raise the
consumer price, generate tax revenues that can be used to increase producer prices. Similarly, the
government may reduce the level of domestic sales taxes when corrrmodity prices rise (e.g., for livestock
products), partially shielding consumers from the price rise effect, thus maintaining a demand for the
domestic production. These various price stabilization mechanisnrs could be investigated in the Malian
context with an appropriate database that isolates sectors (often modern) where domestic sales
taxes/subsidies can be realistically irrrplemented. This current database does not allow such
irrrplementation.

46

CHAPTER IV: A COMPUTABLE GENERAL EQUILIBRIUM MODEL FOR
TRADE REFORM ANALYSIS

4.1. Introduction

Among the numerous tools available for economy-wide policy analyses,
computable general equilibrium (CGE) models have proven to be one of the most
valuable (Dervis, de Melo and Robinson, 1982; Sadoulet and de J anvry, 1995; Wobst,
2001; 2002). It is now well accepted that CGE models “have become a standard tool of
empirical policy analysis” (Lofgren, Harris and Robinson, 2002; p. 1). In particular, trade
is one of the policy areas that have received a tremendous amount of attention in terms of
CGE modeling (Devarajan and Robinson, 2002).

This study uses the standard CGE framework, which follows the usual single
country, open economy trade model developed in the classical work of Dervis, de Melo
and Robinson (1982). This chapter presents the characteristics of the CGE model applied
to the Malian economy. The next section offers an overall discussion of the Malian CGE
model in the context of the literature. Section 4.3 elaborates on the structure of the model
and presents key equations, putting emphasis on the innovations introduced in this study.
This section also discusses the macroeconomic closures of the CGE model. A supplement
to Section 4.3 is available in an appendix which details all the model’s equations. Section
4.4 summarizes and concludes the chapter.

4.2. The Malian CGE Model in the Context of the Literature
The basic theoretical framework of the Malian or other CGE models is a

competitive market equilibrium that satisfies Walras’ law (Decaluwé and Martens, 1988).

47

Producers are assumed to maximize profit using a concave production technology;
consumers maximize utility, and factors are remunerated at the margin, with factor
payments equal to their marginal-value product. 17 Only relative prices matter, and market
clearing conditions allow instantaneous and simultaneous price-adjustment between
supply and demand. Thus, the model is not dynarrric in the classical sense that current
decisions affect future states of economic variables.

This theoretical foundation can be applied using either a classical approach to
trade with direct specification of production and consumption behaviors18 (e.g., Barnou,
1998; Wobst, 2001; 2002), or a dual approach to trade (e. g., Beghin, Bureau and Park,
2003) as systematized in Dixit and Norman (1980) and widely used in many applied trade
models such as Anderson and Neary (1996) and Anderson (1997). The dual approach
represents technology with a production cost function, and consumption by an
expenditure function. In practice, the classical approach allows for a more complete
mapping of the economy, with a richer set of institutional details and interactions, than a
typical dual representation.”

Following the modern use of the classical approach, the Malian model
incorporates specific country characteristics into the standard CGE model to deal with

structural rigidities. Incorporating these rigidities into the classical model transforms it

 

’7 This standard neoclassical technology excludes the existence of a scale economies in the Malian
production sectors, though cotton production and marketing in the country may present some features of
scale econonries. It is expected, however, that the standard technology will characterize the essential
features of all production in Mali.

‘8 This is the representation used in the Dervis, de Melo and Robinson’s (1982) seminal literature. This is
also the essence of substantial applied macroeconomics modeling efforts by the Trade and
Macroeconomics Division (TMD) of the International Food Policy Research Institute (IFPRI). For details,
see www.ifpri.org.

'9 It is, however, possible to incorporate these details into a dual representation, even if existing
applications, such as Beghin, Bureau and Park (2003) or Anderson (1997) tend to use a multi-market or a
CGE framework with no institutional details and typically impotent cross-sector interactions.

48

into a so-called “neoclassical structuralist” model. Structural rigidities of interest include
an explicit accounting of households’ home consumption, as in Wobst’s (2002) study of
five Southern African economies. This specific feature is relevant for agrarian
econonries, such as Mali, where many poor rural households may be unable to participate
in the global markets for food, due to a lack of effective demand. By altering the relative
incentive to produce for home consumption, trade reforms may still affect welfare and
food security in these poor rural households.

The neoclassical structuralist CGE model also incorporates imperfect Armington
constant elasticity of substitution (CES) on the demand side and imperfect Powell and
Gruen’s constant elasticity of transformation (CET) on the supply side.20 This allows for
substitution between domestically produced and externally traded goods. In addition, and
contrary to Wobst (2002) and following Wobst (2001) and several other studies, this
study uses a highly disaggregated sectoral representation of the Malian economy. The
highly disaggregated framework increases intra-sector homogeneity, permitting a more
accurate evaluation of the impact of trade reforms on each of the producer and consumer
groups.

Closely related to this study is Beghin, Bureau and Park’s (2003) investigation of
the food security implications of agricultural protection in South Korea. This study
covers three important aspects. First, the authors use the trade restrictiveness index (TRI)
to evaluate the welfare cost and trade implications of agricultural policies (tariffs, price
support, input subsidies, and consumption taxes) in the country using a multi-market dual

approach following Anderson and Neary (1996). Second, they measure the restrictiveness

 

2° Despite the price-taking behavior of countries, export supply will remain positively elastic, an essential
feature of the neoclassical structuralist specification based on the Powell-Gwen’s CET function.

49

of South Korean agricultural policies using Anderson and Neary’s (2000) mercantilist
indicator of trade restrictiveness index (MTRI). This indicator measures external costs of
trade restrictions, an objective not addressed in this study. Third, they estimate the
second-best optimal use of policy instruments (consumption taxes and production
subsidies) that can help achieve (i) self-sufficiency in each of the staple crops and meat,
and (ii) self-reliance, based on predefined targets of imports and productions of various
sets of commodities.

The model proposed in this study can also provide information on the cost (in
terms of welfare and food security) of the existing structure of protection in Mali, where
the range of available policy instruments is likely to be narrower than in South Korea. As
indicated earlier, input, factor and output tariffs or taxes seem to be the most likely policy
instruments that the Malian government can use, though there may be occasional
exceptions.21 In addition, it is also possible to obtain an optimal second-best structure of
tariffs that is necessary to achieve national food security, defined not as a self-sufficiency
target but as a self-reliance target where both domestic production and imports play a
critical role.

Despite the similarities, this study has some fundamental differences with the
Beghin, Bureau and Park approach. First, due to the relatively small contribution of
agriculture in the Korean economy (about 10% of GDP), Beghin, Bureau and Park have

emphasized details in the agricultural sector, which were assumed to be separable from

 

2' Other policies, such as quantitative restrictions, are illegal under WTO, although the Malian government
may occasionally use them. The use of input subsidies is not compatible with bilateral and multilateral
financing arrangements with, among others, the World Bank and the International Monetary Fund. WTO-
legal policies, such as the “green” and “blue” boxes policies are feasible, provided governments have
enough resources to invest in infrastructure, education, research, and market information systems, which
are sonre of the policy elements allowed under these boxes. The CGE model will capture the effects of
public investments through linkages with agricultural production (see illustration in Figure 2.1, Chapter II).

50

the rest of the economy. These linkages cannot be easily ignored in Mali, where
agriculture contributes more than one-third of GDP, with an even more important
contribution to labor employment. Agricultural production interacts significantly with the
rest of the economy, and these interactions must be captured in a model such as the
neoclassical structuralist framework adopted in this study. Second, their study is based on
the representative economy model, which is silent on distributional issues. The model
proposed here addresses this limitation by distinguishing several household groups in the
Malian economy. Third, while their research focuses on government policy instruments,
this study considers two additional sources of changes, namely, WTO-induced changes in
commodity prices and changes in border protection facing Malian exports.
4.3. Structure and Equations of the CGE Model

The neoclassical structuralist model is very standard and there is no need to
present it in detail in this chapter. Both the classical book by Dervis, de Melo and
Robison (1982), and other references such as Lofgren, Harris and Robinson (2002)
thoroughly discuss the structure of the model and document the complete set of
equations. This section presents the overall structure of the Malian CGE model (Figure
4.1) which is similar to standard model in Lofgren, Harris and Robinson (2002).
Appendix 4.1 provides a complete listing of equations, variables and parameters.
Equations are divided into 4 blocks: prices, production and trade, institutions, and system

constraints.

51

039cm»? :2 can €334 33> u<> ”nouns—cough. Co 36:35 ESEoO ”Hm—O ”nocacmnsm Co bmoumsm “53.50 ”mmolwo.ez

AN— ... ”Sea 58:33— cea mica: .5...»qu Ea... 18963
288m v.25: “:3 22:5 2: .3 5652.205: 72.89% .53 .252 MOO 2: E v.33..— onEEeO "—6 0.53...—

.SQE 0362585

 

 

 

um:
b36253:—
+
2.2582:
+
cote—cacao
EoEEu>oO
+
com-955.80

22.3.5:

 

 

 

Econ—Eco
8.3950

 

 

Eon—E—
n.amocmm<

 

 

a

 

 

 

 

 

 

L

 

 

cocaEameoO
oEo:

 

 

D

 

AWMUV

 

 

 

 

 

EEOC

 

 

 

atonem—
Saws—3"

 

 

 

 

559:0
cow—=2
59:0
vote—.32 .
Samocwm< Gun—UV .5930
3x32

 

 

 

moEvoEEoo Lo
cocoa—:28 oEo:

 

 

 

L

 

 

   

   
    

.. 3E3
Eoc 59:0
b69550

 

   

_ 33:3
88.. 59:0
Econ—Eco

 

     
 

 

3:9:
Defiancefiv
.6353
u E539:
DEB—Eu“:—
Sawocww<
a.
398.:
30 u €023
“ca—«i <>
N .93.—

:_< .<>.
mmU n .2920
— .95.—

"awe—2298
cocoa—3.:—
e832

 

 

 

 

52

Prices

The price block is represented by a system of 10 equations for imports (Equation
4.1), exports (4.2), the demand price of domestic non-traded goods (4.3), the absorption
(4.4), the value of marketed output (4.5), activity price (4.6), aggregate intermediate input
price (4.7), and activity revenue and costs (4.8); there are also equations representing the
consumer price index (4.9) and the producer price index for non-traded marketed output
(4.10). While all these equations are documented in Appendix 4.1, an example is shown

below:

(4.1) PMC = pwmc.(1+tmc).m+ ZPchjmcc'c
c'eCT

Equation (4.1)22 illustrates how the import price in local currency for commodity
c (PM,) is determined from the exogenous world prices of commodity c (pwmc) using
import tariffs (tmc), the exchange rate (EXR), and the marketing margins incurred for
moving the commodity from the border to the end user, represented by the last term in
the right hand side of Equations (4.1). Marketing costs depend on the units icmc'c of other
commodities (c’) used in moving the good c from the border to the demander, c ' being
valued at composite commodity prices PQc-.
Production and Trade

Commodities are produced using a two-level nested technology. At the base of
the technology tree, production factors are combined into value-added using a CBS
technology, whereas intermediary inputs from each activity are combined into aggregate
intermediary inputs following a Leontief aggregation function. At the top level,

commodities are generated from a CBS combination of value added and aggregate

 

22 Equation numbers follow the indexing in Appendix 4.1.

53

 

intermediate inputs. While some of these commodities are home-consumed, the

remaining are marketed. The marketed output is then CBS-aggregated into an aggregate

marketed output, which is in turn CET-transformed into exports and domestic sales.

Domestic sales are CBS-combined with imports to generate the total domestic supply

(Figure 4.1).

There are 17 equations in the production and trade block, all of which are detailed

in Appendix 4.1. These include the following:

An activity production function represented by a CBS combination of aggregate
value-added and aggregate intermediate input quantities (Equation 4.11) and the
corresponding first-order condition, which is represented by the ratio of
quantities of aggregate value-added and aggregate intermediate inputs (Equation
4.12);

A CES combination of production factors into aggregated value-added (Equation
4.13) and the resulting first-order condition representing factor demand (Equation
4.14);

A Leontief demand for aggregate value added (Equation 4.15), and for aggregate
(Equation 4.16) and disaggregated (Equation 4.17) intermediate inputs;

A commodity production (from fixed unit of activities) and allocation of the
production to marketed output and home consumption (Equation 4.18);

A CES aggregation of marketed output from activities to an aggregate marketed
output for each commodity (Equation 4.19) and its corresponding first-order

condition (Equation 4.20);

54

 

A CET output transformation function of the aggregate marketed output into the
aggregated exports and domestic sales (Equation 4.21), the related first-order
conditions that show the optimal ratio aggregate exports to domestic supply
firnctions (Equation 4.22), and the requirement that domestic sales and aggregate
exports add up to the aggregate marketed output for commodities that are either
exported or sold domestically, but not both (Equation 4.23);

A CES composite supply combining aggregate imports and domestic supply
(Equation 4.24), the corresponding first-order condition indicating the ratio of
aggregate imports to domestic supply (Equation 4.25), and the condition that the
sum of domestic supply and aggregate imports equal the composite supply for
commodities that are either non-imported domestic sales or non-produced
imports, but not both (Equation 4.26);

Finally, there is a demand equation of marketing services defined as the sum of

marketing margins on domestic sales, imports and exports (Equation 4.27).

All equations mentioned in the production and trade block are directly taken from
Lofgren, Harris and Robinson (2002), and are therefore not outlined in this section.

Institutions

The CGE model allows for three domestic institutions (households, enterprises

and government) and one foreign institution (ROW). As in previous cases, all the
equations in the institutional block are fully documented in Lofgren, Harris and Robinson
(2002). These equations refer to: total factor incomes (4.28), factor incomes to
institutions (4.29), incomes to domestic non-govemment institutions (4.30), intra-

institutional transfers (4.31), household consumption expenditures (4.32), household

55

 

consumption spending on marketed commodities (4.33), households spending on home
consumption (4.34), investment demand (4.35), government consumption demand (4.36),
government revenue (4.37), and govemment expenditure (4.38).

The model assumes that households are the only non-govemment institutions with
final consumption of commodities. Preferences are modeled following a Stone-Geary
utility function, and the corresponding linear expenditure demand system (LES) is
derived for both marketed commodities (Equation 4.33) and home consumption (4.34).
System Constraints and Macroeconomic Closures

System constraints mainly refer to market clearing conditions and macroeconomic
closures. Equations in this block refer to market clearing conditions in the factor (4.39)
and the commodity markets (4.40). In the Malian model, most factors are location-
specific and are, therefore, assmned to be sector-specific. For these location-specific
factors, the quantity demanded of each factor will equate fixed supply of that factor, with
the equilibrating variable being a factor wage that adjusts to clear the market. There are
also flexible factors, such as wage labor, and such factors are allowed in the Malian
model to move across sectors and locations in response to relative factor returns.

Regarding commodities, markets clear when there is equality between quantity
supplied and quantity demanded of the composite commodity. As discussed earlier in the
production and trade block, the composite supply depends on domestic marketed output
and aggregates imports. Composite demand is the sum of demands for intermediate use,
household and government final consumptions, investment, stock change and trade input
use. With the exception of stock change, all the components of the composite demand are

defined earlier in the production and trade block.

56

The current account balance is given in Equation (4.41).

(4.41) 2 pwchMc + Ztrnsfi'mwj
ceCMW f 6F

= z pwecQEc + Ztrnsfrgmw + FSA V
ceCE W ieINSD

where PIS—Al; represents foreign savings. It is common to treat these foreign savings as
fixed, so that a variable real exchange rate adjusts to impose equilibrium in the current
account balance equations. An alternative treatment would be to fix the real exchange
rate and to flex the foreign savings, which adjusts so that the current accounts are
balanced. This study maintains real exchange rate flexibility as standard and performs
sensitivity analysis with the alternative specification.

The remaining system constraints in the CGE model refer to government revenue
and expenditure balance (4.42), a determination of direct institutional tax (4.43) and
saving rates (4.44), and the savings-investment balance (4.45). The saving-investment
balance requires that the sum of foreign, private and public savings must equate total
investments adjusted with the changes in stocks, both investments and changes in stock
being valued at the composite commodity prices. The government balance imposes
equality between government revenue, on the on hand, and the sum of government
expenditures and savings, on the other. This equation is cleared in the standard model by
making government savings endogenous, while fixing the direct institutional tax
components of the government revenue. Alternative specifications exist where
government savings are fixed whereas institutional direct tax rates vary so that
government revenues equate expenditures. Variations in direct institutional tax rates may
either be uniform across non-government institutions, or proportional to the initial rates

for each institution. In this study, the basic treatment of endogenous government savings

57

is maintained, as the government policies of interest do not involve changing direct taxes
on institutions.

Finally, the specification of non-govemment institutions’ savings (Equation 4.44)
is closely related to government savings. In the standard specification, the saving-
investrnent balance is said to be investment-driven in that the value of savings adjusts to
fixed levels of real investments. Institutional savings are unifonnly adjusted so that total
savings are in equilibrium with total investments. Lofgren, Harris and Robinson (2002)
discuss four alternative saving-investment closures that are commonly used in the
literature. These include: (i) another investment-driven closure similar to the standard
specification, with the difference that adjustments in institutional savings are not uniform
across relevant institutions, but proportional to initial savings of each non-govemment
institution; (ii) a saving-driven balance in which all non—govemment saving rates are
fixed whereas investments are proportionally adjusted to balance the fixed levels of
institutional savings; and two so-called “balanced” investment-driven closures, which
have the characteristics of the first two closures mentioned above (uniform and
proportional adjustment in savings). In addition, the latter two closures are balanced, in
the sense that they do not limit adjustment in absorption to the saving-investment
account, but extend the adjustment to public and private consumptions as well.

Lofgren, Harris and Robinson (2002) note that a macroeconomic scenario that
combines fixed real investment (i.e., investment-driven saving-investment balance), fixed
foreign savings (i.e., flexible exchange rates) and with fixed government consumption
(i.e., flexible government savings) “may be preferable to explore equilibrium welfare

changes of alternative policies” (pp. 15-16). This combination is consistent with the

58

objectives of this study, and will therefore be used as the underlying specification in all
simulations. Alternatively, sensitivity analyses will be performed with the saving-driven
and one “balanced” saving investment closure. Each of the saving-investment closures
(including the base closure) will be combined with flexible exchange rates and
government savings.

Implementing the balanced macro scenarios requires three additional equations,
which are defined in 4.46-4.48. Equation 4.46 defines total absorption as the sum of total
household consumption (including home consumption), government consumption,
investments, and the changes in stocks. Equation 4.47 and 4.48 define the ratios of
investments and government consumptions to absorption, respectively. The last three
equations are not necessary when implementing the model with alternative saving-
investrnent closures that are not balanced across all components of absorption. In that
case, the first 45 equations of the model form a square system that will lead to a unique
general equilibrium solution. The system remains square when the last three equations
are included, as they are associated with three additional endogenous variables. The CGE
system of equations is solved in GAMS as a mixed complementary programming (MCP)
problem, and therefore does not require a specification of an objective function.
Alternatively, it can be solved as a nonlinear optimization problem. It this latter case, a
dummy objective function is defined as solved subject to all the equations in the model.
Details on all the 48 equations are available in Appendix 4.1.

4.4. Summary and Conclusion
This chapter discusses the structure of a neoclassical structuralist computable

general equilibrium, which is applied to analyzing the impacts of global agricultural trade

59

reforms on welfare and food security in Mali. The model is substantially based on a
standard CGE framework developed at the International Food Policy Research Institute
(IFPRI), referenced in Lofgren, Harris and Robinson (2002). The chapter emphasizes the
description of IFPRI’s standard model, with some reference to the specificities in the
Malian economy. It also discusses the Malian structural specificities in the context of the
general CGE literature. The chapter finally discusses macroeconomic closures of the
CGE model, which will be executed using flexible exchange rate, flexible government
savings, and an investment-driven saving-investment balance. Other relevant macro
closures are retained for sensitivity analysis.

The CGE model described in this chapter may be viewed as a simulation
laboratory for investigating the welfare and food security impacts of various scenarios of
global agricultural trade reforms. The model is static, and thus, does not carry any
dynamic, inter-temporal effect. It is, however, able to capture the counterfactual impacts
of trade reforms over a medium horizon. The framework is adequate for addressing the
objectives of this study, but is also sufficiently flexible to accommodate, or be adjusted
to, other relevant objectives. As discussed in Chapter VII, further work could focus on

modifying the model to make it more dynamic.

60

CHAPTER V: A 1997 SOCIAL ACCOUNTING MATRIX FOR MALI

5.1. Introduction

This chapter describes the construction of a social accounting matrix (SAM) for
the Malian economy. A social accounting matrix is a square matrix that maps key
transactions between different accounts in a given economy (Table 5.1). It is a consistent
and comprehensive data framework with an economy-wide scope. The SAM is made up
of row and column accounts. By convention, row accounts receive incomes that are paid
by column accounts. The principle of double-entry accounting, which underlies the
construction of SAMs, dictates that the sum of receipts in a given account (the row total)
must equal to payments distributed along the corresponding column (the column total).

In practice, social accounting matrices are built from different data sources that
are rarely consistent. It is therefore almost always necessary to readjust initial entries to a
SAM, in order to make the row totals balance with the column totals. This readjustment
is known as “SAM balancing”, and standard techniques exist for optimizing that process.
This chapter presents one such technique, the cross—entropy method that was developed
and widely used by researchers at the International Food Policy Research Institute
(IFPRI).

The ultimate goal of this chapter is to produce the so-called “microeconomic
SAM” or “rnicrosam” that maps transactions between production activities, commodities,
production factors and the economic agents (including West Africa and the rest of the
world without West Africa) involved in the Malian economy. In order to construct a

microsarn that is consistent with standard macroeconomic aggregates of the economy, it

61

.358 aooB as 32:5 a 3.83 oomosom 2a was: .9233 x88 :25 aooa ”Goo: Es: 8 Ba soaooom soaom

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EB “moo Locum.“ 3996
8.0on 935a an 8:853:— EoEfioE. .6530 835530 226596 8 votes . 8 59.5 .395.
Q . DOEE
E mew—8:0 8 68¢ .83. Soc :60 Eosomzom omtEoEm 33> .88. $20
8.8% 8.8%
E owed—LO E 093:0 one—=5 seem
Box 8 30x 30x
”Mm—HM.” Echo—ab 9 30.35: c 3n”... 2 2:85 atone: o a "an“?
a m Sec 20:33: . _ m .28... a _ ~—
mess
me>8 . 9.38 macaw. «ea—58>:—
mwfigm cwscom :55 22.8.6: mEE .aznau
. -Eo>o0 .
aeoEmo>£ 22582: 32:32,:—
300 £60 .55538
3% 53 .826
$388 8 3.93 32088 30m tons: «cot—.93—
Eotaoo. m £8 2:85 66» 2 2:85 .85.; 30553.5
Soc EoEEo>o . is .28"—
05 was 33:33: 2 madam 2am
2.8... sienna .wofiomw as... hum.
223.8: 928.. 2 8.61.55 ESP—55 o=_a> 22.8.8:
2...... sown”... Mum
Sentim— vea u2a> flutes—.m—
38 88a 858:—
fi cocoa“ ”up? 9.82...—
o=_a> .a>
. :2. 53956
25:3 833» etaxm «5:582: «eczemoEm aEsmcoo £53.50 Ewan. ...8 89¢
names—on E omen—LO 0325 $00 is: Sow BE..— wczoxaz -253...— 83:38:30
393 swam—coo 593
see Bofioz 3...va 822:4.
.53. «den—.0 30¢ E2: 2.258.; «5.. Eeaom=e= meme—9.8:”.— fletam 82:. 8=$to<
398m .302: -5 250 ...—8.. 65:50 8563<
”saw :5

 

x232 wean-53¢. .325 a he 9.52:5 "—.m 03:.

62

is usually standard to assure the consistency of microeconomic entries in the SAM with
known macroeconomic aggregates. Therefore, the construction of a balanced
“macroeconomic SAM” (macrosam) becomes a necessary step towards building a
consistent microsam.

This chapter first describes the structure of, and entries to, a 1997 macrosam for
Mali. Second, the chapter briefly presents the cross-entropy SAM balancing method,
which is applied to the initial (unbalanced) macrosam. Third, the structure of, and entries
to, a 1997 microsam for Mali are presented along with data sources. The third section
also discusses characteristics of the balanced microsam, obtained after applying the cross-
entropy method to the unbalanced microsam. The last section in this chapter draws the
conclusions from this challenging exercise of constructing a consistent social account
matrix from the poor Malian data.
5.2. A 1997 Macroeconomic SAM for Mali

The year 1997 was chosen to present the Malian economic transactions within a
social accounting matrix framework for several reasons. First, keeping in mind that the
macrosarn is just a first stage towards the construction of a microsam, we have
considered general availability of survey data on microeconomic transactions. Most of
the available data (as will be discussed in Section 5.3) were from 1997 or earlier dates. A
second reason motivating the choice of 1997 is associated with the medium to long-term
forecast horizon that characterizes static general equilibrium models. Thus, it was
necessary to pick a base year that was close enough to the 1995 (when the WTO process
began) in order to capture the situation of the Malian economy in the early stages of the

WTO process. In addition, the base year should also permit the projection of the impacts

63

of market reform scenarios to a time horizon that reflects the current situation of the
Malian economy, bearing in mind the medium-to-long term forecast horizons associated
with CGE models.

National accounts for the Malian economy constitute the main source used for the
34 nonzero entries in the 1997 macrosam. These accounts are compiled and published by
the national statistical office, the Direction Nationale de la Statistique et de
l’Informatique (DNSI). Additional information is obtained from auxiliary (unpublished)
accounts available with DNSI, and from the Banque Centrale des Etats de I’Afrique de
l’Ouest (BCEAO), which is the central bank of 9 West Afiican states. Also, the Tableau
Economique Intégré (TEI), which is a preliminary data framework on the Malian
macroeconomic accounts, was important in finalizing the construction of the macrosam,
as the TEI provided key information regarding institutional saving rates, the distribution
of factor income to institutions, and transfers between institutions. The TEI, along with
other macroeconomic summary tables, are being compiled by DNSI with the assistance
of the Observatoire Economique et Statistique d 'Afiique Subsaharienne (AF RISTAT), an
international organization in charge of improving economic statistics in Sub-Saharan
Afiica in general, but with more visibility in francophone Afiica.

Table 5.] presents the structure of the Malian macrosam, with explanation of
transactions between the row (receipts) and column (expenditures) accounts; Table 5.2
shows the 34 nonzero entries in the matrix. The proposed structure of the macrosam has
10 accounts in the row and a similar number of accounts in the column. In addition to

activities, commodities and factors accounts, there are 6 institutional accounts (two of

64

...—av 25.. bass: 38:95.5 28 ”Slave
2630.. on 3.3%.... on .363 new NNEEMD wagon MGmZDv sawtoztohfi. on B mafiENSM 3. oh £38.32 :95me mocha—om

 

 

 

c: : $.88 88? 88. _ can? ”832 $3: 9.23 53% £88“ .38.

Eng—OD

2.. _ _ a: _ _ amuse seem

38% Ram v.82 8mm. so? 88%. 283

o... ..o 33.

8 am :9.” 332 $39 28: “8.53::

336

awe _ 82: _ 5:535

:66

8.3% 88a 8% am? «a: 88: sunbeam

:60

8%: _ 35m 3va mm 3S3. 2288:

£32 83 £32 agreea—

328 82 _ 985 328a

”25% a: _ ~ 223 BNRN 88. _ «8 SN $28 832 38$ 858558

$88" :SNM ESE 85%;.
.80... omen—.0 283 2.. 32582.. «5:526. .5588 $3.330: momteeam p.88... 83% 3:353.

33. xoefi ..e 63. .8130 :60 :60 58:80

 

3.90 .55.:— Eflaa—z ...... 5.5:—Z maze—593‘ 3.3m n.aeeeuoeuuaz 635.215 53 < "flm 03:.

65

which describe government recurrent and investment transactions), and one account
recording changes in stocks. The other four institutional accounts are households,
enterprises, saving-investment and the rest of the world. The explanation and sources of
these entries are discussed in Appendix 5.1.

5.3. A SAM Balancing Cross-Entropy Method23

Balancing a SAM using the cross-entropy (CE) method is now a standard
procedure in most SAM-based modeling. The method is formalized in two recent
publications by IFPRI researchers (Robinson, Cattaneo and El-Said, 2000; Robinson and
El-Said, 2000), but it has roots in both information theory (Shannon, 1948; Theil, 1967)
and maximum entropy econometrics (Golan, Judge and Robinson, 1994; Golan, Judge
and Miller, 1996). As explained in Robinson, Cattaneo and El-Said (2000), the CE
approach has been widely used to estimate social accounting matrices in Eastern and
Southern African countries (Botswana, Malawi, Mozambique, Tanzania, Zambia and
Zimbabwe), as well as in other countries, including Brazil, Mexico, North Korea and the
United States.

Traditionally, balanced social accounting matrices are estimated or updated using
the RAS method. This method particularly suits situations where new information
becomes available regarding the row and coltunn sums of a SAM, with no knowledge on
the new flows between various accounts of the SAM.

The RAS method can be represented as follows. Let T be an (n x 11) SAM with
each entry tg- representing the payment from column account j to row account i. By
definition, every row sum of a SAM must equal the corresponding column sum, and this

can be represented as follows:

 

23 This section is mainly based on Robinson, Cattaneo and El-Said (2000).

66

(5.1) yi = Zty' = thi
J' 1'

Define A to be a SAM coefficient matrix with each entry 0;; representing the ratio
Iii/)7. Let y be the (n x 1) vector of y,, it follows from the double-entry bookkeeping
convention underlying SAMs that:
(5.2) y = Ay

As y changes to y*, there is a need to obtain a new SAM coefficient matrix A*,
which is close to the original matrix A, such that the equation (5.2) holds:
(53) y* = A*y*

Robinson, Cattaneo and El-Said (2000) follow Bacharach (1970) and discuss how

the matrix A* can be obtained by means of biproportional row and column operations on

the original matrix A. Let R and S be diagonal matrices of elements r,- and s,-,
respectively. The biproportionality condition can be written as follows:
(5.4) A* = I“: A s

The form of this equation has lent its name to the procedure known as the “RAS”
method. By imposing the biproportionality condition, one effectively reduces a problem
of n2 unknowns (the n x n cells of a SAM) to (2n-1) unknowns, which are fully and
uniquely determined by the (Zn-1) independent adding-up restrictions corresponding to
the new row and column totals. It therefore is possible to frnd a unique set of positive
elements of A*, and the elements of R and S can be recovered by iteration.24

The cross-entropy method adds more flexibility to the traditional RAS approach.

The method can incorporate information regarding not only new row and column totals,

 

2‘ Bacharach (1970) discusses conditions under which a solution exists and Robinson, Cattaneo and El-Said
(2000) surmnarize some of these conditions, including the need that A is “connected”, that is, all rows and
columns must be nonzero.

67

but also prior knowledge on any part of the SAM. It can also incorporate errors in
variables and various forms of constraints to any part of the SAM. For example, a single
value of households? factor income can be disaggregated into F*H figures corresponding
to the incomes to H types of households from F different factors. Using the CE method,
one can impose the restriction that the sum of the F*H factor incomes to households be
equal to the aggregate value from the macrosam.

As indicated above, Robinson, Cattaneo and El-Said (2000) have explained that
the CE method is built on information theory, as developed by Shannon (1948) and
brought to economics by Theil (1967). The main idea is that the expected information
value of additional data can be expressed as a Kullback-Leibler (1951) cross-entropy
distance I(p:q) between the prior (q) and posterior (p) probability distributions of a set of
n events.

n
(5.5) —I(p:q)=—Zp.- 1an
i=1 1

The objective of the CE problem is to find the set of p,- that minimize (5.5), using
information on the prior and the data. With regards to SAM estimation or updating, the
problem is to find a new SAM coefficient matrix A* that minimizes the CE distance
between itself and the prior (or initial and probably unbalanced) coefficient matrix A. Let
ag“ and a.) be the respective elements of A* and A, the minimization problem can be

written as follows:

*

au-

aij ij aij
*

subject tozzag-y; =y:;za;,~ =1; and OSaJ-i $1
1' f

(5.6)

68

Problem (5.6) does not have a closed form solution, and it needs to be solved

numerically, afier setting up the Langragian. It is, however, possible to express the

optimal solution a; as a function of both the Lagrange multipliers ,1,- associated with the

row and column sums, and the initial coefficient 0;}:

.
a. _ 0;] exp(2.,-y j)

I] — t
Zayexpmyj)
i, j

(5.7)

 

Robinson, Cattaneo and El-Said (2000) discuss how (5.7) is comparable to Bayes’
rule, in which “the posterior distribution [. . .] is equal to product of the prior distribution
[. . .] and the likelihood function [. . .], dividing by a normalization factor to convert
relative probabilities to absolute ones” (p. 7). Thus, Equation (5.7) may be seen as an
efficient information processing rule that satisfies Zellner’s (1988) information
conservation principle. That is, it does not ignore any of the input information and neither
does it produce any false information. The authors also draw on Golan, Judge and Miller
(1996) to argue that the CE estimator is consistent and has maximum likelihood
properties under some distributional assumptions.

The basic minimization problem in (5.6) can be made richer by incorporating
aggregation constraints and measurement errors. For k restrictions, a typical aggregation

constraint may be expressed as follows:

k
(5.8) 22g; )tij = 70‘)
i j

where gij are elements of a (n x n), zeros-ones25 aggregator matrix G and y is the value of

the macrosam aggregate. Similarly, measurement errors are incorporated as follows:

 

25 The ones (zeros) correspond to the cells included (excluded) in the definition of a specific aggregate.

69

(59) y = x + e
where y is a vector of row sums and x, measured with error e, is the vector of known
column sums. The error is expressed as a weighted average of known constants viw:
31' = Z wiw-viw
W

(5.10)
Zwiw =1 andOSwiw 51

w

The weights are treated as probabilities, which are estimated together with the
elements of the matrix A*. The estimation procedure used in this study is based on five
weights that are symmetric about zero. The minimization problem (5.6) is solved, subject
to Equations (5.8), (5.9) and (5.10). The CE procedure is applied to the initial SAM in
Table 5.2 and the result is shown in Table 5.3, which is the balanced macrosam for Mali.
The figures in parentheses indicate the percentage change between the initial entries in
Table 5.2 and new estimates in Table 5.3.

In general, most estimates are within a 5% margin from initial figures. However, a
few higher adjustments were necessary, particularly regarding the incomes (+86%) and
expenditures (-7.3%) of enterprises as well as in the saving rates of households (+20.6%)
and the rest of the world (+12.4%). Also noticeable are a 10.5% decline in household
income from abroad and a 10.3% increase in the transfers from the rest of the world to
the government recurrent account. These various adjustments can be sources of many
speculations. For example, firms may be viewed as more profitable than the raw data may
suggest, as their incomes increase while their costs fell in the final SAM. Second,
households may be underreporting their real level of savings, which would be consistent

either with cultural valuation of diffidence in most parts in Sub-Saharan Afiica, or with

70

‘1'

And 2an Bob 5% no 8838 853‘ ”ooh—ow
Show 8 was 22.35; 8 31¢an th £86 8. 3262 33 73:38:80 803383 39:0 888—88
Samohwwa 05 63888 Sm .855 REE 2: 808 828, Be: 05 no Aowfiaoouoa as 55858 2: 328 835588 E moSwE ..

 

 

 

8.8 8.8 83 8.8 2.3 8.8 83 2.3 2.3 8.8 Boy

8.. : :88 88: 888 _ $88. $8 _ _ _ 2 2 m2 8888 88:8 8888 .835
8.8 8.8
c: : o: : 858 89m
83 2.3 8.3 83 8.8 8.3 233
:88 :5 8.8: 89.8 888. 88:. 23° “88
8.8 2.88 8.8 888 . 83 .8882—
883 838 8.8. 888 8:8 838
8.8 8.8 525822
888 _ 838 _ :60
2.8 8.88 8.8 83 8.8 8.8 .8588
$88. 888 $.88 88% 83. 82 22 .80
8.3 8.83 8.8-8 8.83 2.3
88:: 8on £88 8.. 828. 5838:
8.8 2.3 8.8
8:: 82m :88. maceuam
23 2.3 2.3
8888 38 _ 2.882 £988
88 8.8 8.8 8.3 8.8 8.3 8.8 8.8 2.8
8828 c: _ _ $88 8:88 888 _ $3: 8 83$ 88: 888 8889580
8.8 8.8 2.3
8888 8888 not: 3823.
230,—. Dag—«£0 2.53 «5:.— =55 «GP—.300.— man "”0th 390nm mote 82>
39m 33% 05 .802: -835 $00 -030: $ch 68800 -uo<

8 as. .388 $8

 

:52 new 5.5a: Mme—583‘ 7825 852899.32 h3— cooasam < "Qm nigh.

7l

the fear to pay taxes on declared incomes/savings. Third, foreign transfers to
government may be higher than reported, and some of the difference may be thought of
as being leaked into corruption channels. Finally, it may as well be that the observed
differences are a simple indication of measurement errors associated with some of the
flows in the SAM. It is obvious that none of these speculations are formally warranted
by the simple exercise of putting together macroeconomic data in a consistent SAM
framework. But for the Malian macroeconomic flows in 1997 to pass a basic
consistency test, the adjustments are necessary, even if the forces driving them remain
unidentified in this study.

All the 34 nonzero entries in the balanced macrosam should, in principle, be
used as the right-hand side figures in the aggregation constraints discussed in Equation
(5.8). Doing this would be necessary to ensure that entries in the disaggregated SAM
(see the next section) add up to their corresponding aggregates in the macrosam.
However, the prior disaggregated SAM that was obtained from the best available
information (on detailed transactions in the Malian economy) was so highly unbalanced
that the cross-entropy minimization problem would not converge, unless less binding
aggregate control conditions were imposed. This is further discussed in the next section,
which elaborates on the structure, entries and the balancing of the Malian microsam.
5.4. Structure and Entries in the Malian Microsam
Structure of the Microsam

The structure of the microsam is based on the 1997 input-output table for the
Malian economy, as well as on the 1997 survey of agriculture. Both the I-0 table and

the agricultural survey data were obtained from DNSI officials. There were 17 sectors in

72

the I-0 table, and each of them represents one activity in the microsam. There are four
cropping activities, two for food cropping (rice; and other food crops), and the other two
for industrial agriculture (cotton; and other industrial crops). In addition to the four
cropping activities, there are two other agricultural activities: fishery-livestock keeping,
and forestry-gathering. The last activity belonging to the primary production sector is
mining. The secondary or industrial sector is mapped into five activities (agro-industry,
beverage and tobacco; textile; other manufacturing goods; electricity, water and energy;
and construction and public works). The remaining five activities describe the service
sector. These activities include trade and four other services (transport and
communication; other traded services; banking and insurance; and nontraded services).
Table 5.4 shows all the 17 activities, along with other accounts in the microsam.

The second category of accounts in the microsam consists of 37 commodities, 25
of which are agricultural. Following the tradition of the Malian national statistical office
(DNSI), “agriculture” is broadly defined in this study to include not only crop
production, but also livestock keeping, gathering, forestry and fishery activities. Based
on the structure of the I-0 table, the food cropping activity (excluding rice) produces
seven commodities (millet; sorghum; maize; fonio; roots and tubers such as potatoes,
yarns and cassava; beans; and others). The industrial cropping activity (excluding
cotton) produces three commodities (groundnuts, tobacco and wheat). There is one-to-
one mapping between rice and cotton production activities and rice and cotton
commodities, respectively. The activity “livestock keeping and fishery” produces 10
commodities (cattle; sheep and goats; pork; donkeys and camels; poultry; eggs; milk;

hide and skins; fresh fish; and smoked fish), whereas the activity “forestry and

73

Table 5.4: Structure of the Malian Microsam

 

 

 

Account Account Account Account Description
Category Number Name
Activities 1 AF OOD Food production, excluding rice production
2 ARICE Rice production
3 AINAG Industrial agriculture, excluding cotton production
4 ACOTT Cotton production '
5 ALIFI Livestock keeping and fishery
6 AF OGA Forestry and Gathering
7 AMINI Mining
8 AAIDT Agro-industry, beverage & tobacco
9 ATEXT Textile production
10 AOMGG Production of other manufacturing goods
11 AEWEN Production of electricity, water and energy
12 ABBTP Construction and public works
13 ATRAD Trade
14 ATTEL Transport and telecommunication
15 AOTSE Other traded services
16 ABISE Bank and insurance services
17 ANT SE Nontraded services
Commodities 1 8 CMILL Millet
19 CSORG Sorghum
20 CMAIZ Maize
21 CF ONT Fonio
22 CROOT Root crops (potatoes, yam, cassava)
23 CBEAN Beans
24 COTHE Other crops
25 CRICE Rice
26 CGNUT Groundnuts
27 CTOBA Tobacco
28 CWHEA Cotton
29 CCO'IT Wheat
30 CCATT Cattle
31 CSHGT Sheep and Goats
32 CPORC Pork
33 CDKCM Donkeys and Camels
34 CPOUL Poultry
35 CEGGS Eggs
36 CMILK Milk
37 CSKIN Hide & Skins
38 CFFIS Firewood
39 CSF IS Furniture wood
40 CFWOO Hunting/Gathering products
41 CWOOD Fresh fish
42 CGATH Smoked fish
43 CPGLD Gold
44 CAIDT Agro-industry and beverages
45 CT EXT Textile products
46 COMGG Other manufacturing goods
47 CEWEN Electricity, water and energy—

 

74

Table 5.4: Structure of the Malian Microsam (Continued)

 

 

 

 

 

 

 

 

 

 

 

 

Account Account Account Account Description
Caflry Number Name
Commodities 48 C BBTP Construction and public works
(continued) 49 C TRAD Trade services
50 C TTEL Transport and telecommunication services
51 C OTSE Other traded services
52 CBISE Bank and insurance services
53 C NTSE Nontraded services
Marketing 54 MMDOM Marketing margins on domestic sales
margins 55 MMIMP Marketing margins on imports
56 MMEXP Marketing margins on exports
Production 5 7 F C API Capital
factors 5 8 F WLAB Wage labor
59 F IENT Independent entrepreneurship
6O FTRAD Composite factor for trading services
61 FOUAC Composite factor for other urban activities
62 FF CKS Composite farming factor in Rural Kayes-Sikasso
63 F F CRV Composite farming factor in Rural River Region
64 FFOTH Composite farming factor in other rural
65 F ORAC Composite factor for other rural activities
Households 66 HBMKO Bamako residents
67 HOURB Residents in other urban zones
68 HRUSK Residents in Rural Kayes-Sikasso Region
69 HRURV Residents in Rural River Region
70 HORUR Residents in other rural zones
Enterprises 71 ENTRE Formal enterprises
Taxes 72 Value-added taxes for government recurrent
GRVAT account
73 GRTAR Tariff incomes for government recurrent account
74 GROTP Other taxes on products
75 GRFAC Factor taxes for government recurrent account
76 GRDTA Direct tax on institutions
77 GROTH Other incomes for government recurrent account
Government
Recurrent 78 GRECU Government recurrent account
Government
Investment 79 GINVE Government investment account
Saving
Investment 80 SINVE Saving investment account
Rest of the 81 WESTA West Africa
World 82 ROROW Rest of the world, excluding West Africa
Stock cham 83 STCHG Change in stock

 

Source: Author.

gathering” produces three commodities (firewood; furniture wood; and hunting and

gathering products). The mining activity produces essentially the commodity “gold”,

75

although some national figures show traces of phosphate production in 1997. All the
remaining activities each produce each one commodity.

Besides the 17 activities and the 37 commodities, the Malian microsam includes
three marketing margin accounts, one for each of domestic sales, imports and exports. It
also includes nine production factors, five household types and 6 tax accounts, in
addition to the private and public saving-investment accounts, the rest of the world
(West Afiica and rest of the rest of the world) accounts, and a last account for the
change in stocks.

Entries to and Balancing of the Microsam

The entries to each account in the microsam are documented in Appendix 5.2. In
order to complete the prior (unbalanced) microsam, it was necessary to use data from
various (and often inconsistent) sources. These include the 1997 input-output table, the
1997 national agricultural surveys, the 1997 employment survey, the 1989 and 1994
household consumption and expenditure surveys, the DNSI trade statistics and other
auxiliary information. The entries involve two steps. In the first step, the aggregate
figures from the unbalanced macrosam were disaggregated into the corresponding micro
accounts, based on detailed activity and commodity information in the DNSI database,
or on disaggregation assumptions that are outlined in Appendix 5.2. The second step
was to readjust the initial entries using the balanced macro aggregates, so that each entry
in the new microsam is proportional to the initial entry. The proportionality coefficient
is equal to the ratio of the aggregate entry in the balanced macrosam to the
corresponding aggregate entry in the unbalanced macrosam. The resulting new

microsam is balanced in the aggregate but remains substantially unbalanced in the

76

details. The disparities are illustrated in Appendix 5.3, which shows to the ratio of
column totals (expenditures) to row totals (incomes) for the 83 micro accounts.

The disparities are most acute with the commodity accounts, the highest being
observed with grains and livestock products. For instance, the total expenditures by the
millet commodity account (made of payments to activities, marketing margins and
taxes) cover only 16% of the incomes received in the same account (the incomes are
generated from the use of millet as intermediate input, final consumption good by
households, and exports while accounting for the changes in stocks). At the other
extreme, total expenditures by the gathering/hunting commodity make up more than 8
times the incomes received in the same account (Appendix 5.3). These disparities are
mainly explained by the existence of inconsistencies in the various sources of data used
to obtain commodities’ total supply (expenditures by the commodity accounts) and
demand (incomes received in these accounts). In particular, the supply data, which are
mainly based on the 1997 LC table and DNSI’s national agricultural surveys, may have
been underestimated because of the importance of the traditional sector in the Malian
agricultural economy. Most of the transactions in this sector may not be reflected in
existing national statistics available with DNSI. Conversely, the demand data that are
based on older household surveys, such as the 1989 Enquéte Budget Consommation
(BBC 1989) and the 1994 Enquéte Malienne de Conjoncture Economique et Sociale
(EMCES 1994), may have been overestimated in many instances. This may be partly
due to the use of the possibly inflated data from EMCES 1994, which took place less
than four months after a historic 50% devaluation of the Malian currency, the CFA fianc

(CFAF). Though efforts were made to correct for the possible upward bias in the

77

demand data, using more recent surveys (e. g. the 1999 urban consumption survey in
Mali, by the West Afiican Economic and Monetary Union), discrepancies remained
important.

There is no unique way of dealing with this problem. Given that the best
available information was used in constructing the prior SAM, the readjustment of the
disparities was left to the cross-entropy SAM balancing process. This process has the
merit of effectively optimizing the use of all information contained in the unbalanced
SAM. The drawback in applying the CE method to such an inconsistent base matrix is
that the balancing process may sacrifice valuable information on the structure of the
economy, a necessary cost for obtaining a feasible solution. The Malian SAM was no
exception. The structure of the final microsam (Appendix 5.4) presents structural
differences with the initial unbalanced matrix. But given that the initial matrix was itself
inconsistent in terms of the transaction flows between accounts, it is unclear which of
the balanced and the unbalanced matrices present a structure close to the “true”, but
unobserved, social accounting matrix for Mali in 1997. The premise throughout this
chapter was that whatever this true SAM may be, it is fairly well approximated by the
balanced SAM, which makes the best use of all information contained in the prior and
unbalanced SAM. The rest of this chapter further discusses the microsam’s accounts
while presenting the structure of key transactions in the 1997 balanced SAM.
Production Factors

There are nine factor accounts in the rrricrosarn (Table 5.4). Factors receive
income in the form of value-added from activities, and fi'om the rest of the world. They

pay this income to households and to government factor tax account. The choice of the

78

nine factors was based on the results of the EMCES 94, which provided basic
information of the relative shares of each of these factors in the production process of
most activities retained in the SAM. The EMCES data was completed with additional
information obtained fi'om employment surveys. Due to data limitations, it was not
possible to disaggregate factor returns exclusively into standard production factors, such
as capital, labor and land. Instead, we retained two standard production factors (capital
and wage labor) in addition to seven unconventional (but widely used) composite
production factors: (i) independent entrepreneurship, (ii) trading services factors, (iii)
other urban factors, (iv) other rural factor, (v) farming factor in Kayes-Sikasso, (vi)
farming factor in the River region, and (vii) farming factor in other rural areas. These
seven factors are consistent with the notion of “composite production factor”, which has
been widely used in the literature (e. g., Thurlow and Wobst, 2003). The rationale for
using composite factors lies in the impossibility of obtaining accurate information-of the
shares of each of the classical production factors (land, labor and capital) in the total
returns to production activities in either the small-scale farming or to other informal
sectors. Returns to such activities are best viewed as an aggregate return to the
household composite factors, mainly composed of family labor, but also including other
assets such as capital and land.

The seven composite factors included in the microsam inform on both the
geographical locations and the socioeconomic groups of active households.26 Though
conceptually different from the five households included in the matrix, they represent

unique regional and socio-professional characteristics, and this uniqueness could enrich

 

2° The location of these regions is indicated in the next section, which deals with household
disaggregation.

79

Table 5.5: Distribution of the Value-Added Across Sectors

 

 

 

 

 

 

 

 

FCAPI FWLAB FIENT FTRAD FOUAC FFCKS FFCRV FFOTH FORAC
AFOOD 0.0 0.0 0.0 0.0 8.6 21.1 21.0 18.7 20.2
ARICE 0.0 0.0 0.0 0.0 1.9 4.4 4.3 6.8 7.3
AINAG 0.0 0.0 0.0 0.0 0.6 2.7 1.6 1.4 1.8
ACOTT 0.0 0.0 0.0 0.0 0.0 21.6 6.3 1.0 10.6
ALIFI 12.4 0.0 0.0 0.0 6.7 0.7 26.1 33.9 1.1
AFOGA 0.0 0.0 0.0 0.0 0.0 2.4 14.1 18.4 0.0
AMINI 0.0 0.0 0.0 0.0 0.0 22.9 0.0 0.0 0.0
AAIDT 0.0 42.9 3.9 0.0 0.0 0.0 0.0 0.0 0.0
ATEXT 0.0 7.0 0.0 0.0 0.0 0.7 0.2 0.0 0.4
AOMGG 11.3 0.1 0.7 0.0 0.4 0.1 0.7 2.6 1.7
AEWEN 0.0 20.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0
ABBTP 18.6 1.4 19.5 0.0 5.0 2.0 3.0 0.9 7.2
ATRAD 19.9 10.4 19.1 84.6 15.9 16.3 13.7 5.3 28.1
ATTEL 1.7 3.3 29.3 0.0 30.4 0.9 2.9 5.0 5.3
AOTSE 0.0 4.0 20.7 3.1 24.3 4.0 4.8 3.6 13.0
ABISE 0.0 5.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0
ANT SE 36.2 5.3 6.8 12.3 6.3 0.2 1.3 2.3 3.3
Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0
Table 5.6: Distribution of the Value-Added Within Sectors

FCAPI FWLAB FIENT mo FOUAC FFCKS FFCRV FFOTH FORAC Total
AFOOD 0.0 0.0 0.0 0.0 2.7 36.1 30.3 18.0 12.9 100.0
ARICE 0.0 0.0 0.0 0.0 2.4 29.5 24.4 25.4 18.3 100.0
AINAG 0.0 0.0 0.0 0.0 1.9 47.9 24.5 13.6 12.0 100.0
ACOTT 0.0 0.0 0.0 0.0 0.0 68.6 16.9 1.9 12.6 100.0
ALIFI 14.5 0.0 0.0 0.0 2.5 1.4 43.3 37.5 0.8 100.0
AFOGA 0.0 0.0 0.0 0.0 0.0 9.8 48.3 41.8 0.0 100.0
AMINI 0.0 0.0 0.0 0.0 0.0 100.0 0.0 0.0 0.0 100.0
AAIDT 0.0 94.5 5.5 0.0 0.0 0.0 0.0 0.0 0.0 100.0
ATEXT 0.0 68.3 0.0 0.0 0.0 21.0 5.6 0.8 4.3 100.0
AOMGG 68.9 0.3 1.6 0.0 0.7 1.0 6.2 14.8 6.6 100.0
AEWEN 0.0 100.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 100.0
ABBTP 45.5 1.9 17.2 0.0 3.8 8.0 10.4 2.1 11.0 100.0
ATRAD 16.3 4.8 5.7 12.2 4.1 22.5 15.8 4.1 14.5 100.0
ATTEL 4.6 5.1 28.4 0.0 25.5 4.0 10.9 12.7 8.9 100.0
AOTSE 0.0 5.2 17.4 1.3 17.7 15.8 15.8 7.9 18.9 100.0
ABISE 0.0 100.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 100.0
ANTSE 69.3 5.7 4.7 4.2 3.8 0.7 3.5 4.2 4.0 100.0
Total 14.1 8.0 5.1 2.5 4.4 23.7 20.0 13.3 8.9 100.0
Notes to Tables 5.5 and 5.6:

Activities: AF OOD: Food cropping (excluding rice); ARICE: Rice production; AINAG: Industrial
agriculture (excluding cotton); ACOTT: Cotton production; ALIFI: Livestock keeping and fishery;
AFOGA: Forestry and gathering; AMINI: Mining; AAIDT: Agra-industry and beverage; ATEXT: Textile
production; AOMCG: Production of other manufacturing goods; AEWEN: Production of electricity,
water and energy; ABBTP: Construction and public works; ATRAD: Trade; ATTEL: Transport and
telecommunication; AOTSE: Other traded services; ABISE: Banque and insurance services; and ANT SE:
Nontraded services.

Factors: FCAPI: Capital; FWLAB: Wage labor; F IENT Independent entrepreneurship; FTRAD:
Composite factor for trading services, FOUAC: Composite factor for other urban activities; FF CKS
Composite farming factor rn Rural Kayes-Sikasso Region; FFCRV: Composite farming factor in Rural
River Region; FFOTH: Composite farming factor in other rural; FORAC: Composite factor for other rural
activities.

Source: SAM

80

the impact assessment of various trade reforms simulated in this study. Factor allocation
is likely to change in response to trade reforms, and it will be possible to depict these
changes in terms of households’ location and socio-professional orientations. The two
remaining factors—capital and wage labor—also capture other aspects of the production
structure in the Malian economy. The distribution of factor incomes is shown below in
Tables 5.5 and 5.6. Table 5.5 shows the distribution of what each factor earns across
sectors while Table 5.6 shows the distribution of value added in a given sector across
factors. Details of these distributions are available in Appendix 5.2 and in the balanced
microsam (Appendix 5.4). In general, returns to capital were restricted to six activities
in which investments were made, and these returns were assumed to be proportional to
the investments. This assumption is reasonable if factor (capital) price equalizes across
sectors. The share of wage labor in activities’ value added was obtained in most case
from employment surveys (e. g. Doumbia and Kamaté, 1997). Aggregate payments to
the seven composite factors were obtained as residuals, after paying the returns to
capital and wage labor. These aggregate payments are further disaggregated using
shares obtained for employment surveys. Essentially, value added was distributed in
proportion to employment, under the assumption that that labor productivity is equal
across activities.

The distribution of factor incomes to domestic non-govemment institutions
(households and enterprises) is also explained in Appendix 5.2. Following standard
practices, it was assumed that the capital factor belongs to enterprises, so that the entire
return to this factor goes to these enterprises (with the exception of direct capital

incomes to government received in the other government recurrent income account, i.e.

81

the GROTH account). Enterprises also received payments from other factors, including

the “independent entrepreneurship”, the “other urban” and the “other rural” factors

(Tables 5.7 and 5.8). These three factors, which contribute more than 50% of total

enterprise income, are central to production in the informal business sector.

Table 5.7: Distribution of Factor Incomes to Households and Enterprises

 

 

 

 

 

 

Other Rural K— Rural Rural Total
Bamako urban S River Other Entemrise
FCAPI 0.0 0.0 0.0 0.0 0.0 100.0 100.0
FWLAB 39.3 23.9 2.2 9.5 25.1 0.0 100.0
FIENT 18.4 14.6 1.6 16.6 7.7 41.1 100.0
FTRAD 31.9 33.1 5.3 18.5 11.3 0.0 100.0
FOUAC 0.0 67.1 0.0 0.0 0.0 32.9 100.0
FFCKS 0.0 0.0 100.0 0.0 0.0 0.0 100.0
FFCRV 0.0 0.0 0.0 100.0 0.0 0.0 100.0
FFOTH 0.0 0.0 0.0 0.0 100.0 0.0 100.0
FORAC 0.0 0.0 14.2 32.7 13.9 39.2 100.0
Table 5.8: Sources of Income to Households and Enterprise
Other Rural Rural Rural
Bamako Urban K-S River Other Enterprise
FCAPI 0.0 0.0 0.0 0.0 0.0 46.0
FWLAB 64.7 29.3 0.7 3.0 11.3 0.0
FIENT 18.4 10.8 0.3 3.1 2.1 15.2
FTRAD 16.9 13.0 0.5 1.9 1.6 0.0
FOUAC 0.0 46.9 0.0 0.0 0.0 11.5
FFCKS 0.0 0.0 93.6 0.0 0.0 0.0
FFCRV 0.0 0.0 0.0 80.4 0.0 0.0
FFOTH 0.0 0.0 0.0 0.0 77.7 0.0
FORAC 0.0 0.0 5.0 11.7 7.2 27.4
Total 100.0 100.0 100.0 100.0 100.0 100.0

 

Notes to Tables 5.7 and 5.8: Factors: FCAPI: Capital; FWLAB: Wage labor; FIENT: Independent
entrepreneurship; FTRAD: Composite factor for trading services; FOUAC: Corrrposite factor for other
urban activities; FF CKS: Composite farming factor in Rural Kayes-Sikasso Region; FFCRV: Composite

farming factor in Rural River Region; FFOTH: Composite farming factor in other rural; FORAC:

Composite factor for other rural activities.

Source: SAM

Wage labor is used in the formal production sector, and the returns to this factor

are distributed to households (Table 5.7 and 5.8). Bamako residents receive more than

60% of their incomes from wages, but other household groups also receive wage

incomes (Table 5.8). The combined wage incomes for urban households (Bamako and

other urban residents) make more than 60% of the total wage income (Table 5.7). Like

82

 

C00:

111C 0‘

:
.171?

C03.“

Fall?

the

3V5

Ha

(I),

wage labor, the composite trade and independent entrepreneurship factors also
contribute to incomes in the five household groups, with important contributions to
incomes in the two urban households (Table 5.7 and 5.8). The three types of composite
farming factors pay all their incomes to their corresponding household groups. Also, the
composite factor for other urban activities pays its incomes to other urban households
and to enterprises, whereas the other rural activities’ composite factor pays the three
rural households in addition to enterprises. Further details on production factors and
their interactions with other accounts are either discussed in the Appendix 5.2 or are
available in the balanced microsam (Appendix 5.4) and in Tables 5.7 and 5.8.
Households

The microsam contains five household types, which are defined based on the
place of residence of the population. This disaggregation, which originates from the
results of EMCES 94, distinguishes two urban households (Bamako and other urban
dwellers) and three rural (Kayes-Sikasso, River, and Other Rural dwellers). Regional
composition of these rural households is shown in Table 5.9, and their geographical
location is indicated in the political map of Mali (Figure 5.1).
Table 5.9: Regional Composition of Household Groups (and Composite Rural

Farming Production Factors)
Locations Composition
Rural Kayes-Sikasso Sikasso Region (plus Dioila Circle)
Kayes Region (minus Nioro Circle)
Rural River Koulikoro Region (minus Dioila and Nara Circles)
Segou Region (minus Niono Circle)
Mopti Region (minus Bandiagara, Koro, Bankass, Douentza Circles)
Rural Other Tombouctou Region
Gao Region
(plus Nioro, Niono, Bandgg , Koro, Bankass, Douentza Circles)

Source: Based on Charmes (1994; p. 19)

83

 

   
  
 

LEGEND

Rural Kayes- j
Sikasso

 

Koulikoro

    

Bamako

 
 

 

 

 

Figure 5.1: A Political Map of Mali Showing Administrative Regions and Circles
(Source: University of Texas)

Charmes (1994) originally gives the name “Southwest” to the Kayes-Sikasso
Region. This is because it is composed of the Southern Sikasso and the Western Kayes
regions (Figure 5.1). The term “Southwest” is confusing, however, so we explicitly call
the region “Kayes-Sikasso” in this study. It appears that Charmes (1994) was trying to
distinguish between households based on their core agricultural and pastoral activities.

One can note that Kayes and Sikasso households essentially live off an entirely rainfed

84

agriculture. Both rainfed and irrigated agriculture is practiced by households living
along the Niger River (Koulikoro, Ségou and Mopti), and the Other Rural households
live in predominantly pastoral and desertic regions (Tombouctou, Gao and Kidal).

One obvious limitation of the location-based household disaggregation is that the
socioeconomic characteristics of households are heterogeneous within each of these
geographic zones. As a consequence, the assessment of the impact of global agricultural
trade reforms (undertaken in this study) is likely to overlook this heterogeneity.
Nevertheless, as small number of households (five in this case) can still deliver a richer
result than will the standard single-household models.

Turning now to household demand, one can distinguish consumptions based on
households’ own productions from the market demands. Since home consumption is
based on production, the structure of that consumption was determined based on the
intensity of production activities in each of the five household groups. Home
consumption was concentrated in the three rural households (Kayes-Sikasso, River and
Other), with the home consumption of livestock products primarily important in the
River and Other Rural households.27 The share of some activities in households’ home
consumption in the balanced SAM changed substantially with respect to the unbalanced
SAM. For example, the relative share of Kayes-Sikasso in the home consumption of
food cropping the activity (excluding rice) increased to more than 53% in the balanced
SAM, from an initial value of 36% (Table 5.10). This increase may help compensate the

drastic decline in the share of commodities produced by that activity (i.e., food

 

27 Home consurrrption of livestock products in the Kayes-Sikasso households turned out much smaller
than anticipated. This was one of the drastic changes in the initial structure of the SAM, due to SAM-
balancing procedure. This problem could have been avoided, if the initial data presented greater
consistency between the demand and supply sides of the economy.

85

cropping) observed in the balanced SAM for the Kayes-Sikasso region. As suggested

earlier, this disparity may stem from the undervaluation of marketed output, or the lack

of valuation of home-consumed staples, due to the dominance of a traditional mode of

agricultural production and trade in Mali.

Table 5.10: Home Consumption of Activities by Household Groups (in percerm

 

 

 

 

Rural
Other Kayes/ Rural Rural,
Activities Bamako urban Sikasso River Other Total
Food production, excluding rice 0.36 0.52 53.22 44.72 1.19 100.00
Rice production 0.15 1.60 29.53 23.15 45.58 100.00
Industrial agric., excl. cotton 0.30 1.06 51.24 24.46 22.94 100.00
Livestock keeping and fishery 0.18 1.52 0.00 52.33 45.97 100.00
Forestry and gathering 0.00 0.00 10.10 47.41 42.49 100.00
Other traded services 1.68 3.23 27.31 25.37 42.42 100.00
Total 0.25 0.94 26.91 41.56 30.34 100.00
Source: SAM

Households’ final consumption is presented in Table 5.11, which shows the

budget shares for the 37 commodities in each of the five households. The initial budget

shares were based on the structure of households’ expenditure from the EMCES 94,

completed with selected figures from EBC 89 and from the 1996 and 1999 WAEMU

consumption surveys in Bamako. Details on the home and marketed final consumptions

are found in the Appendix 5.2.

Some anomalies arise from Table 5.11, which shows no consumption of roots

and tubers by the Rural Kayes—Sikasso households, or of beans in the River and Other

Rural regions. Optimal households’ consumption of these products was zero following

the cross-entropy balancing process. This reduction to zero, from a nonzero basis, was

the cost to pay in order to attain a feasible solution for the balancing problem. The

disappearance of such effective consumptions will necessarily alter the initial budget

shares, reducing substitution possibilities between commodities. In many cases,

86

f"

however, the initial consumptions were sufficiently small that they may be viewed as
negligible in terms of their welfare and food security impacts.

Table 5.11: Budget Shares byHousehold Groups
Other Rural, Rural, Rural,

 

 

 

Commodities Bamako Urban K-S River others
Millet 3.26 3.40 3.31 9.39 13.42
Sorghum 2.68 2.84 4.32 4.97 7.47
Maize 0.85 0.89 1.82 2.08 3.05
Fonio 0.10 0.10 0.28 0.26 0.50
Roots & tubers (potatoes, yam, cassava) 1.44 0.65 0.00 0.44 0.28
Beans 0.00 0.01 0.02 0.00 0.00
Other crops 4.97 5.25 6.53 6.06 6.26
Rice 9.06 9.38 4.58 5.23 7.31
Groundnuts 0.03 0.03 0.25 0.45 0.00
Tobacco 0.78 1.05 1.61 1.02 1.19
Wheat 2.37 0.89 1.52 0.85 1.77
Cotton 0.00 0.00 0.00 0.00 0.00
Cattle 5.05 5.41 3.94 1.62 1.62
Sheep and Goats 1.58 1.71 3.88 0.26 0.47
Pork 0.04 0.04 0.07 0.00 0.00
Donkeys and Camels 1.33 1.50 1.91 0.79 0.91
Poultry 0.46 0.52 0.98 0.41 0.40
Eggs 0.23 0.14 0.17 0.00 0.12
Milk 2.49 1.46 2.09 3.01 ' 3.62
Hide & Skins 0.14 0.16 0.00 0.00 0.00
Firewood 1.99 3.03 3.35 4.74 4.52
Furniture wood 0.60 0.74 0.43 0.59 0.31
Hunting & gathering 0.18 0.15 0.88 1.09 0.30
Fresh fish 0.03 0.03 0.00 0.00 0.00
Smoked fish 0.21 0.13 1.15 1.11 0.47
Gold 0.00 0.00 0.00 0.00 0.00
Agro-industry & beverages 15.43 14.86 17.62 14.06 15.31
Textile 5.31 6.68 7.74 8.50 8.92
Other manufacturing goods 8.75 10.36 15.22 12.84 10.59
Electricity, water and energy 2.66 7.97 3.93 4.84 4.85
Construction & public works 4.30 2.54 0.20 0.47 0.24
Transport & telecom services 19.38 13.89 8.61 11.52 2.67
Other traded services 3.83 4.01 3.58 3.39 3.42
Bank and insurance services 0.45 0.19 0.00 0.00 0.00
Total 100.00 100.00 100.00 100.00 100.00
Source: SAM
Trade

The data on external trade contained in the microsam originate from various

sources, including DNSI, FAO, BCEAO and some published analyses on the Malian

87

external trade (Diakite and Samake, 2002). As highlighted throughout this study, the
external sector was differentiated into regional and non-regional accounts. One of the
characteristics of trade statistics is that they are highly inconsistent, depending on
sources. For example, while the LO table showed no trade in cereals in Mali in 1997,
Diakité and Samaké (2002) documented the existence of trade in cereals between Mali
and its neighbors. These discrepancies are usually attributed to the existence of a
dynamic informal trade within West Africa, and this trade tends to escape official
channels of trade recording. These disparities are accounted for in the most pragmatic
way in the unbalanced microsam, and further explanations are available in Appendix
5.2.

Table 5.12 describes the structure of the regional and non-regional commodity
imports and exports, showing the share of each commodity in total exports and imports,
both at regional and non-regional levels. It appears from the table that imports are
dominated by the commodity “COMGG” (other manufacturing goods), which
constituted nearly 62% of total imports in Mali in 1997. Regarding exports, they are
dominated by cotton (about 42%) and gold (nearly 34%) in 1997. Currently, gold is the
main export from Mali, overtaking cotton since 1998. Table 5.12 also shows the West
African and non-West African shares in the commodity trade. All of Mali’s exports of
cereals, livestock and fish went to the West African region, and the quasi-totality of its
roots and tuber imports came from within West Africa. All rice and wheat imports and
most of the milk, tobacco and maize imports were from non-regional origins. Details are

provided in Table 5.12 and in the balanced rrricrosam (Appendix 5.4).

88

Table 5.12: Commodity Shares in West African and ROW Trade in Mali, 1997

 

 

 

 

 

Goods and Services Imports Exports
W. W.

Total Africa ROW Total Africa ROW
Millet 0.0 n.a. n.a. 0.3 100.0 0.0
Sorghum 0.0 n.a. n.a. 0.0 100.0 0.0
Maize 0.1 7.9 92.1 0.1 100.0 0.0
Fonio 0.0 n.a. n.a. 0.0 n.a n.a
Roots & tubers (potatoes, yam, cassava) 0.1 94.1 5.9 0.0 n.a n.a
Beans 0.0 16.3 83.7 0.1 100.0 0.0
Other crops 0.5 74.3 25.7 0.7 100.0 0.0
Rice 3.8 0.0 100.0 0.0 100.0 0.0
Groundnuts 0.0 n.a. n.a. 1.6 0.0 100.0
Tobacco 1.3 3 .2 96.8 0.0 n.a n.a
Wheat 1.3 0.0 100.0 0.0 n.a n.a
Cotton 0.0 n.a. n.a. 41.8 0.0 100.0
Cattle 0.0 n.a. n.a. 5.1 100.0 0.0
Sheep and Goats 0.0 n.a. n.a. 3.1 100.0 0.0
Pork 0.0 n.a. n.a. 0.0 n.a n.a
Donkeys and Camels 0.0 n.a. n.a. 0.0 100.0 0.0
Poultry 0.0 n.a. n.a. 0.2 100.0 0.0
Eggs 0.0 100.0 0.0 0.0 n.a n.a
Milk 2.0 3.0 97.0 0.0 n.a n.a
Hide & Skins 0.0 n.a. n.a. 1.0 100.0 0.0
Firewood 0.3 100.0 0.0 0.0 100.0 0.0
Furniture wood 0.1 100.0 0.0 1.3 100.0 0.0
Hunting & Gathering 0.0 n.a. n.a. 0.0 n.a n.a
Fresh fish 0.9 100.0 0.0 0.0 n.a n.a
Smoked fish 0.0 n.a. n.a. 0.0 n.a n.a
Gold 0.0 n.a. n.a. 33.6 0.0 100.0
Agro-industry & beverages 5.1 88.7 11.3 0.0 n.a n.a
Textile 4.5 37.8 62.2 0.0 n.a n.a
Other manufacturing goods 61.9 7.1 92.9 0.0 n.a n.a
Electricity, water and energy 13.3 99.5 0.5 0.0 n.a n.a
Construction & public works 0.0 n.a. n.a. 0.0 n.a n.a
Trade services 0.0 n.a. n.a. 0.0 n.a n.a
Transport & telecom services 2.3 1.7 98.3 10.0 100.0 0.0
Other traded services 0.0 n.a. n.a. 0.9 56.4 43.6
Bank and insurance services 2.5 51.9 48.1 0.0 n.a n.a
Nontraded services 0.0 n.a. n.a. 0.0 n.a n.a
Total 100.0 27.1 72.9 100.0 22.6 77.4

Source: SAM
Marketing margins

The main source of information about marketing margins on domestic sales and

imports (as percentage of marketed output) was the DNSI’s I-O table. Marketing

89

margins on the two main exports (gold and cotton) were assumed to be integrated into
the operating costs of gold mining companies and the cotton parastatal. These margins
would be captured through intermediate consumption of the marketing services. The
situation of these companies (which were operating at what may be called a meso-
economic level) may be viewed as different from that of micro agents who performed
independent trading activities involving commodities that were sold domestically,
imported or exported. Marketing margins on exports were, therefore, restricted to _
commodities traded on smaller scales, such livestock and cereals. Margins were taken
from the existing literature, such as Diakite and Samaké (2002), who evaluated the
marketing margin of livestock exports to be about 12% of the export values. Entries for
the marketing margins are explained in Appendix 5.2 and final marketing margins are
available in the balanced microsam (Appendix 5 .4).
Other entries to the microsam

All the remaining microsam entries not discussed above are fully explained in
Appendix 5.2, and their final values are indicated in the balanced microsam (Appendix
5.4). The main categories include (i) intermediate inputs, (ii) government receipts in the
form on taxes and tariffs, (iii) transfers between domestic and foreign institutions, (iv)

institutions’ saving rates, (v) public and private investments and (vi) changes in stocks.

5.5. Conclusion

This chapter was concerned with the work of building a detailed social

accounting matrix (SAM) for Mali, the first for the country. The matrix focuses on the

90

 

details of the agricultural sector. Constraints related to data availability and the medium
run impact assessment needs have dictated the choice of 1997 as the base year for the
matrix. The final matrix is composed of 83 accounts including accounts for 17 activities,
37 commodities, three marketing margins, nine production factors, five household
groups, one enterprise, seven taxes and tariffs (pooled into one government recurrent
account), one public and one private saving-investment account, two rest of the world
accounts, and one account for the changes in stocks.

Various data sources were used for entries of transactions between these
different accounts in the SAM. As result, the initial matrix was highly unbalanced, with
incomes on some commodities making up eight times the expenditures on these
commodities. It was therefore necessary to balance the initial matrix, using the best
available information on the transactions between the various accounts in the matrix and
controlling detailed entries with known aggregate values. The balancing process was
performed using the cross-entropy optimization procedure, which minimizes the entropy
distance between the unbalanced and balanced SAMs. The resulting SAM is the final
social accounting matrix for Mali, calibrated to 1997. Due to the high disparities in
income and expenditure data in the initial SAM, an optimal solution was only possible
at the cost of altering the structure of transactions in the initial matrix. But since the
initial matrix was in itself inconsistent, the alteration may be viewed as an improvement,
although it is impossible to ascertain this since balancing is not per se a virtue. Given
the limited information available, this study will proceed under the assumption the
balanced SAM represents an acceptable approximation of the true, but unobserved,

Malian social accounting matrix in 1997.

91

The exercise of building a detailed SAM for the Malian economy uncovers the
serious data limitations in Mali, which points to areas where efforts for future data
gathering may be concentrated. First, the discrepancies between incomes and
expenditures suggest a total disconnect between the supply and demand sides in terms
of data gathering. Perhaps the results of recent household surveys, once publicized, may
reduce the inconsistencies between supply and demand figures in the Malian economy.
Otherwise, future efforts should be aimed at narrowing the inconsistency gap between
the supply and demand sides of the economy. SAMs provide an excellent fiamework for
such a consistency check.

Second, though the structure of the Malian economy has evolved since the base
input-output table was developed, probably in the mid-eighties, this initial table
constitutes the basis for all projected I-O tables in Mali. There is an ongoing effort
between the Malian national statistical office in Mali (DNSI) and AFRISTAT
(Observatoire Statistique et Economique de I’Afrique Subsaharienne) to develop a new
I-O table as part of the implementation the 1993 United Nations’s Systems of National
Accounts (SNA). It will be useful if the new table emphasizes a detailed representation
of the Malian economy, and ensures among other things a one-to-one mapping between
key activities and goods and services.

Third, given the usefulness of social accounting matrices in policy analysis, and
given the latest interest on poverty assessment in Mali and elsewhere in Sub-Saharan
Africa (SSA), it will also be useful to implement the 1993 SNA so as to encourage the
construction of a poverty-focused SAM in Mali. Incorporating a fine household and

production factor disaggregation (in addition of a detailed representation of activities,

92

commodities and government policy instruments) will improve substantially the present
SAM that relies on both “composite” production factors and location-based grouping of
households.

Finally, it will be useful to encourage inter-departmental cooperation that, with
its pool of skills, could improve on the limitations associated with the 1997 SAM
presented in this chapter. Policy analyses will certainly benefit such an effort. The

current SAM will serve as a useful base for future improvements.

93

CHAPTER VI: CGE MODEL IMPLEMENTATION AND RESULTS

6.1. Introduction

The Malian CGE model was implemented using computer codes supplied by the
International Food Policy Research Institute (IFPRI), as documented in Lofgren, Harris
and Robinson (2002). Model implementation consists of applying the theoretical model
(Chapter IV) to an aggregated version of the Malian social accounting matrix from the
previous chapter using additional exogenous data on behavioral parameters. The Malian
CGE model was solved numerically with the General Algebraic Modeling Systems
(GAMS) software using the CONOPT 2 nonlinear programming solver.

The remainder of this chapter is organized in five sections. Section 6.2 describes
the aggregation of the Malian SAM into a smaller number of accounts used in the CGE
model. Section 6.3 discusses the calibration of the re-specified CGE model, with
emphasis on the behavioral parameters assumed for the analysis. Most of the materials
in this chapter are concentrated in Section 6.4, which presents and discusses the CGE
base year and simulation results. The stability of these results is tested in Section 6.5,
which presents the results of various sensitivity analyses involving model parameters
and macroeconomic closure rules. The last section (Section 6.6) presents further
discussions and conclusions.

6.2. Aggregation of the Malian Social Accounting Matrix

As indicated throughout this dissertation, this study is interested in the impacts

of trade reform scenarios on welfare and food security in Mali. Such an inquiry requires

the quantification of the direction and magnitude of price and policy shocks associated

94

with these reforms. Chapter III provided detailed information on scenario design. The
inquiry also requires the identification of the main components of welfare and food
security in Mali. Chapter V documents the structure of the Malian economy in terms of
the sources of incomes and the destination of expenditures. This study defines food
security broadly as changes in food consumption, without going into details regarding
the nutritional and utilization aspects of food security. Such changes could be computed
for food product groups, which are more aggregated than those discussed in Chapter V.
Changes and welfare can also be measured, irrespective to the degree of aggregation of
food commodities. The CGE model can be kept more tractable, and abnormally lower
entries in the SAM could be eliminated, if some commodities were aggregated in the
final specification of the model. Given these reasons, the Malian CGE model was run
using a more aggregated SAM. The aggregation was also justified on a more practical
ground. Indeed, an initial attempt to run the CGE model with more disaggregated data
did not result in a feasible solution, primary because of the extremely small entries in
the SAM.

Two types of aggregation were made. First, the two external sectors (West
Afiica and Rest of the World) were aggregated into a single rest of the world sector,
which is consistent with the CGE model outlined in Chapter IV. Second, the 37
commodities in the original matrix were grouped into twenty product categories shown
in Table 6.1. The first two columns in the table show the 37 initial commodities and
their corresponding SAM accounts, and the last two columns present the same

information for the 20 aggregated product groups.

95

Table 6.1: Mappingthe Original 37 Commodities Into 20 Aggr_egated Commodities

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SAM New SAM

Original Commodities Account Amted Commodities Account

Millet CMILL Coarse grains CCORS

Sorghum CSORG

Maize CMAIZ

Fonio CF ONI

Root and tubers CROOT Other food COFOO

Beans CBEAN

Other crops COTHE

Rice CRICE Rice CRICE

Groundnuts CGNUT Other industrial. ag products CIN AG

Tobacco CT OBA

Wheat CWHEA

Cotton CCOTT Cotton CCOTT

Cattle CCATT Cattle and sheep and goats CCASG

Sheep and Goats CSHGT

Pork CPORC Other livestock COLIV

Donkeys and Camels CDKCM

Poultry CPOUL

Eggs CEGGS

Milk CMILK

Hide & Skins CSKIN

Fresh fish CF FIS Fish CF ISH

Smoked fish CSFIS

Firewood CFWOO Forestry & gathering products CFOGA

Furniture wood CW OOD

Hunting/Gatheringproducts CGATH

Gold CPGLD Miningproducts CPGLD
Ago-industry and beverages CAIDT Agro-industry and beverages CAIDT

Textile products CTEXT Textile products CT EXT

Other manufacturing goods COMGG Other manufacturinggoods COMGG

Electricity, water and energy CEWEN Electricity water and emy CEWEN

Construction & public works CBBTP Construction & public works CBBTP

Trade services CT RAD Trade CT RAD

Transport & telecom Services CTT EL Transport & telecom C'I'I‘ EL

Other traded services COTSE Other traded services COTSE

Bank and insurance services CBISE Bank and insurance services CBISE

Nontraded services CNT SE Nontraded services CNTSE

 

Source: Based on the Malian Social Accounting Matrix (Chapter V)

Most of the aggregation refers to crops and livestock products. For example, four
cereals (millet, sorghum, maize and fonio) from the initial SAM were aggregated into
single product (coarse grains); Roots and tubers, beans and other crops from the initial
SAM were grouped into the “other food” product category; products from the industrial

agricultural activity (groundnuts, tobacco, and wheat) were combined into a single

96

“other industrial agriculture” product. Given the importance of products such as cotton
and rice in the Malian economy, they were kept unchanged in the aggregated SAM.
Also, there was a one-to-one mapping between commodities of the initial SAM and the
remaining 11 products in the aggregated SAM (fi'om mining products to nontraded
services in Table 6.1).

Additional mappings included two livestock product groups: (i) cattle, sheep and
goats, and (ii) other livestock, which includes commodities such as pork, donkeys and
camels, poultry, eggs, hide and skins, and milk; there was also a group for fishery
products and another group for forestry and gathering products.

Besides contributing to eliminate marginal values and provide more tractability
to the CGE results, there is another good rationale for aggregating the commodities in
the SAM. The construction of the detailed matrix SAM relied on simplifying
assumptions in order to facilitate the disaggregation of commodity groups into
individualized products. Since these assumptions may cause distorsions in the estimated
SAM, the aggregated SAM may be viewed as less distorted than the detailed SAM.

6.3. Model Calibration

The Malian CGE model was calibrated using IFPRI’s computer codes written in
General Algebraic Modeling Systems (GAMS) language. The model was solved
numerically using the CONOPT2 nonlinear programming solver. Inputs to the model
included the Malian social accounting matrix (SAM) and other behavioral parameters
on production technology, commodity trade, and consumer preferences. These

parameters were taken directly from the literature.

97

The model has two types of production elasticities, characterizing the two levels
of the nested production technology (described in Chapter IV). At the bottom of the
technology nest, production factors are CES aggregated into value added. Production
elasticities at this first level are characteristic of the rate of substitution between
production factors. These elasticities were obtained from Decaluwé, Dissou and
Robichaud (2000) who used them in a multi-country CGE study in seven West Afiican
countries, including Mali. They range fiom as low as 0.45 on cropping activities to as
high as 2 on services (see Appendix 6.1). At the top of the production nest, value added
and aggregate intermediate inputs are aggregated into outputs using a CBS production
technology. The second group of production elasticities corresponds to the substitution
between aggregated production factors (or value added) and aggregate intermediate

inputs. These elasticities were set to a generic 0.6, as suggested in Lofgreen, Harris and
Robinson (2002). The model also contains an output aggregation elasticity, which was
set to 6 for all activities, again following Lofgreen, Harris and Robinson (2002).
Similar to the case of production elasticities, there are two types of trade
elasticities. The first represents the demand side of the economy and corresponds to the
substitution between imports and local sales of domestic output. The second
characterizes the supply side of the economy and shows the rate of transformation of
total output into domestic sales and exports. These two sets of elasticities were taken
from Delacuwé, Dissou and Robichaud (2000). For each commodity, the CBS and CET
elasticities are equal to each other, taking the values of either 1.5 (e. g., agricultural and

mining products) or 2 (e. g., services). Details are available in the Appendix 6.1).

98

Finally, parameters for the linear expenditure demand system (LES) are
calibrated from the aggregated SAM entries, along with assumed values for expenditure
elasticities and the Frisch parameter. The Frisch parameter (Frisch, 1959) measures the
negative of the marginal utility of income with respect to income, also known as the
flexibility of the marginal utility of income. Following Lofgreen, Harris and Robinson
(2002), the Frisch parameter was set to -2 for all five households.28 Because quantities
consumed in the Malian CGE model were expressed in monetary units, changes in
consumption expenditures will be fully transmitted to changes in consumed quantities.
Thus, expenditure elasticities were set to unity for all commodities.

With all the calibrated values, the model can be solved for a base year solution
that replicates the aggregated SAM. A model that successfully replicates the base year
solution is said to be properly calibrated. With a few exceptions, most of the calibrated
figures were within +/-10% of base values (for details, see Appendix 6.2). These
deviations are high but remain nonetheless reasonable, given all the problems associated
with the Malian social accounting matrix (see discussions in Chapter V). A more
consistent database would probably lead to smaller deviations between the base and
calibrated figures.

Armed with the calibrated model, one can perform simulation analyses. Base
and simulation results are presented in Section 6.4. The calibrated model will also serve

as the basis for sensitivity analyses, discussed in Section 6.5.

 

2' The negative of the inverse of the Frisch parameter measures the average price elasticity of substitution
of a good i for a good j (Sato (1972). With a Frisch parameter of -2, the average elasticity of substitution
is 0.5.

99

6.4. Results

The interest of this study lies in the impacts of trade and government policy
reforms on welfare and food security in Mali. Most of the results will primarily
emphasize the welfare and food security dimensions of these policy reforms. However,
in order to elucidate the channels through which reforms are likely to affect welfare and
food security, the section will also present additional results on changes in other
economic variables, whenever this is relevant.

The results are organized in six parts, defined by the various impacts and
scenarios examined in this study. The first four parts focus on the measure of welfare
changes under various scenarios. Part 1 (Section 6.4.1) compares the results of two the
scenarios of incomplete reforms in the world commodity prices (FAPRI and OECD).
Part 2, (Section 6.4.2) compares the two IFPRI scenarios, which assumes no and full
implementation of WTO agricultural trade reforms. The two scenarios of change in
market access to Malian exports (DFA and EEP) are compared in Part 3 (Section 6.4.3),
whereas Part 4 (Section 6.4.4) discusses three government policy scenarios (BAN, CET
and GINV). Part 5 is exclusively dedicated to a discussion of the food security impacts
of the nine scenarios presented in Parts 1, 2, 3 and 4 of this Section. Finally, Part 6
(Section 6.4.6) examines the interaction between global trade reforms and productivity

gains in Mali and the welfare and food security impacts of this interaction.

100

6.4.1. Partial Reforms in World Commodity Markets: FAPRI and OECD
Scenarios
(i) Overview of the Scenarios

The F APRI and OECD scenarios are part of the four price change scenarios
discussed in Chapter III.29 Although the key differences between these two scenarios are
detailed in Chapter 111 (see Table 3.2), a quick reminder is useful for understanding what
assumptions underlie the results. The two scenarios are based on price projection
models that assume an average world GDP growth rate of approximately 3% and a
world population growth averaging 1%. The FAPRI and OECD models are built on the
expectation that major world currencies will appreciate against the US dollar. They also
assume that the climatic conditions will remain at their historical averages, though only
the FAPRI model states this explicitly. Also, the OECD model explicitly assumes
increased production efficiency over time, but this is not explicit in the FAPRI model.

The assumed agricultural and trade policies under the FAPRI and OECD are also
substantially similar. The two models assumed that current agricultural policies will
remain in place over the projection horizon, i.e. by 2010. Examples of such policies
include provisions of the 2002 US farm bill, the 2003 reform of the European Union
Common Agricultural Policy, and agricultural policy provisions for major agricultural
exporters and importers. The magnitude of policy provisions may differ between the
two models, hence the differences in the projected price changes (as shown in Table 6.2

later in the section).

 

29 Table 6.2 (below) presents the magnitude of change in commodity prices under the FAPRI and OECD
as well as the two IFPRI scenarios. These magnitudes are obtained by aggregating expected changes in
the prices of individual commodities (Table 3.2, Chapter III) into the corresponding product categories, as
described in Table 6.1. The value of each commodity is used as a weight in the aggregation.

101

Regarding trade policies, the FAPRI and OECD models are based on trade
reform commitments made by countries under the WTO Uruguay Round Agreements
on Agriculture (URAA). The commitments relate to limiting the use of export subsidies,
expanding tariff rate quotas, and reducing import duties and domestic support programs.
These commitments have been worked into national agricultural policies, which they
complement. Again, the main difference between the FAPRI and OECD models lies in
the difference in commodity coverage and the magnitude of committed policies.

The committed URAA reforms are assumed to be implemented through the
projection period under the FAPRI and OECD. While reducing worldwide protection of
agricultural markets, implementation of current URAA commitments will not result in
complete liberalization of these markets. Thus, the level of world trade liberalization
under the FAPRI and OECD scenarios may be viewed as an intermediate stage of trade
liberalization.

Table 6.2 shows the projected changes in world commodity prices for the FAPRI
and OECD scenarios and the two IFPRI scenarios (see Section 6.4.2 for more discussion
the IFPRI scenarios). The table is derived by aggregating price changes in Table 3.3
(Chapter 111), using the key provided in Table 6.1. The share of each commodity in the
total consumption value of the commodity group is used as a weight in the aggregation.

Table 6.2: Projected Changes in World Commodity Prices: FAPRI and OECD
Scenarios (% change from base)*

 

 

Commoditxgroups FAPRI OECD
Coarse grains -17.2 0.1
Rice -25.7 -18.7
Other industrial agricultural products -2.3 -0.4
Cotton ~20.3 -
Cattle, sheep and goats 6.4 9.6
Other livestock -6.1 l .6
_A_gro-industry and beveggs -25.9 ~23.8

 

Source: Based on Table 3.3, Chapter H1.

102

There are several important differences in the projected change in world
commodity prices between the F APRI and OECD scenarios. For example, one of the
main differences between FAPRI and OECD scenarios is a 20.3% projected decline in
world cotton prices under the FAPRI scenario compared to no change in the OECD
scenario. In addition, coarse grain prices are expected to decline sharply (-17.2%) under
FAPRI, while remaining substantially unchanged under OECD projections (+01%).
Similarly, FAPRI projects a 6.1% fall in the prices of other livestock products,
compared to a slight increase of 1.6% under the OECD scenario. Rice prices are
projected to fall under the two scenarios, with a sharper decline under FAPRI
projections (25.7% against 18.7%). The price of the “cattle, sheep and goats” would
increase under the two scenarios, with a slightly larger rise in the OECD scenario (9.6%

against 6.4%).

(ii) Change in Welfare in the FAPRI and OECD Scenarios

If the FAPRI and OECD prices were the prevalent world prices in Mali in 1997,
the country could have experienced a nearly 3% gain in welfare. This gain would be
concentrated in urban areas, but most rural households would lose only marginally from
the price changes. For example, for an average resident of Bamako and other urban
zones, welfare would rise by 11.7% and 9.9%, respectively, under the FAPRI price
change scenario. Under the same scenario, welfare would rise by 2.9% in Other Rural
zones, and fall slightly (03% to -0.2%) in the rural Kayes-Sikasso and River regions

(Table 6.3).

103

Table 6.3: Equivalent Variation: FAPRI and OECD Scenarios
(% change from base)*

 

 

 

Global (WTO)
Households BASE FAPRI OECD
Bamako 139.5 11.7 11.6
Other urban 93.1 9.9 9.5
Rural S-K 321.5 -0.2 1.7
Rural River 273 .9 -0.3 -1 .4
Other rural 228.7 2.9 1.4
Total 1056.8 2.9 2.8

 

There are three major factors driving these results: (i) the expected changes in
world commodity prices, (ii) the resulting changes in factor rental rates, and (iii) the
macroeconomic adjustment effects on the real exchange rates.

First, changes in commodity prices are one of the two channels through which
policy changes can affect household wealth, as discussed in Bourguignon, de Melo and
Suwa (1991).30 The Malian model is built on the assumption that the country will not
participate in the worldwide efficiency gain that would drive down world commodity
prices. In other words, the Malian input-output coefficients will remain constant. Thus,
Malian producers will face lower world prices that will not be offset by productivity-
induced lower unit costs of production. The direct consequence of lower agricultural
prices is reduced agricultural revenues, which have a negative impact on household
welfare in the heavily agricultural Kayes-Sikasso and River zones.3 I On the other hand,
urban households benefit from decreased real food prices, which have positive welfare
impact. Contrary to the Kayes-Sikasso and River households, the Other Rural

households consume a very low share of their own food production. They behave very

 

3° In fact, Bourguignon, de Melo and Suwa ( 1991) identify three channels. The other two channels are: (i)
changes in factor incomes, and (ii) capital gains and losses and portfolio decisions in models
incorporating assets markets. The latter channel is irrelevant for the Malian model, which does not
explicitly incorporate assets markets.

3 ' In Chapter V, households were classified based on the dominant farming system The Kayes-Sikasso
region was characterized by rainfed agriculture; in the River region uses a combination of rainfed and
irrigated agriculture; and the Other Rural regions are essentially pastoral or desert.

104

similarly to urban households, which are predominantly net food purchasers. Thus, they
would benefit fiom the fall in real food prices, but lose only marginally in agricultural
revenues. Consequently, the overall welfare effect is positive for the Other Rural
households under the FAPRI and OECD scenarios.

Second, changes in domestic commodity prices are also associated with changes
in factor incomes, also known as the functional distribution of income (Dixon and
Norman, 1980; Khan, 1998). Changes in factor incomes will, in turn, be translated into
welfare changes. This is the second channel, identified in Bourguignon, de Melo and
Suwa (1991), through which welfare can change following policy reforms. Changes in
world commodities prices have similar effects on factor incomes under the F APRI and
OECD scenarios (Table 6.4). Table 6.4 is divided into two parts: the first part shows the
changes in factor incomes resulting from changes in product prices, and the second part
presents the shares of domestic institutions (households and firms) in total factor
income. Consistent with expectations, changes in households’ welfare appear to be
highly correlated with changes in the incomes of factors in which these households are
most endowed. For example, the results show that incomes to the Rural River composite
farming factor would fall by nearly 5% under the OECD scenario. This change is
translated into 1.4% loss in welfare (Table 6.3). The Rural River composite farming
factor (F FCRV) is 100% owned by the Rural River households.

Other links between factor endowment, changes in factor incomes and welfare
changes can be established fi'om Table 6.4. Noteworthy is the strong effect changes in

wage incomes has on welfare changes in the urban (Bamako and Other urban) and the

105

other rural households. Nearly 90% of labor wages is distributed among these three
households (Table 6.4).

Table 6.4: Disaggregated Factor Income Distribution: FAPRI and OECD
Scenarios (% change from base)

 

 

 

 

 

Items F CAPI FWLAB FIENT FOUAC FFCKS FFCRV FFOTH FORAC
Initial factor return 119.3 184.1 46.6 48.1 310.6 232.2 173.7 80.0
Scenarios Percentage change from base
FAPRI -5.6 24.6 5.2 0.7 -1.5 -2.9 -3.3 -1.9
OECD -5.3 20.2 1.6 -1.8 0.0 -4.8 -5.8 -1.0
Institutions Share of domestic institutions in factor income
Bamako 0.0 39.3 18.4 0.0 0.0 0.0 0.0 0.0
Other urban 0.0 23.9 14.6 67.1 0.0 0.0 0.0 0.0
Rural K-S. 0.0 2.2 1.6 0.0 100.0 0.0 0.0 32.7
Rural River 0.0 9.5 16.6 0.0 0.0 100.0 0.0 14.2
Other rural 0.0 25.1 7.7 0.0 0.0 0.0 100.0 13.9
Firms 100.0 0.0 41.1 32.9 0.0 0.0 0.0 39.2
Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

 

Notes: Factors include: capital (FCAPI) and wage labor (FWLAB), as well as composite factors for
independent entrepreneurship (FIENT), other urban activities (F OUAC), Rural Kayes-Sikasso farming
(FFCKS), Rural River farming (FFCRV), Other Rural farming (FFOTH), and other rural activities
(FORAC).

While it is straightforward that changes in factor incomes would partly drive
welfare changes, it is less clear how the changes in world agricultural prices have
affected the returns to production factors. The general relationship between product
prices and factor rental rates is described under various versions of the Stolper-
Samuelson theorem. In a “two goods and two factors economy” changes in product
price are positively transmitted to changes in returns to the factor intensively used to
produce that good. This equivalence breaks down once one considers an economy with
several goods and factors, as is the case in this study. All attempts to generalize the
Stolper-Samuelson theorem to a multi-good, multi-factor environment have not
produced a useful grid through which the Malian results may be interpreted (a review is
available in Lloyd (1998)). Most discussions point to the necessity of having at least one

positive and one negative element in every row and every column in the Stolper-

106

 

Samuelson derivative matrix.32 No indication is made, however, on the exact location of
these mandatory positive and negative elements. In the Malian specific context, the use
of composite production factors adds another dimension of ambiguity regarding the
determination of the Stolper—Samuelson effects. Given the overall ambiguity, one may
expect any change in factor wages as a response to changes in relative good prices. This
was precisely what the Malian results reveal, with some factors gaining and others
losing from each of the six exogenous policy shocks. The patterns of these gains and
losses are similar to those described earlier regarding welfare changes. Changes in the
relative prices of final and intermediate demands appear once again to be the key driver
of the observed changes. Specifically, changes in product prices will be positively
correlated with changes in incomes of household-location composite factors in regions
that are net sellers of these products.

The third and final factor driving welfare changes is the adjustment in the real
exchange rate. Results suggest that the real exchange rate would appreciate under the
FAPRI and OECD scenarios (Appendix 6.4). The aggregate price index of nontradable
goods increased faster than the decline in the price index of tradable goods in these two
scenarios. It follows that the price ratio of tradables to nontradables declined, leading to
the appreciation of the real exchange rate. This appreciation would sharply raise the
demand for agro-industrial imports (Appendix 6.5). With at least 15% of the budget
share (Chapter V), agro-industrial products are important components of household

consumption. These products, which are based on value-added processing, tend to have

 

’2 The Stolper-Samuelson derivative matrix in a N goods and M factors economy is given by the M x N
matrix [(awn/apn)*(p.,/wm)], where the expression in the square bracket represents a proportionate change
in the rental rate of the m—th factor (wm) with respect to change in the price of the n-th good (pa).

107

higher income elasticity. Thus, changes in the consumption would closely reflect
changes in households’ incomes.
(iii) The Malian Welfare Results in the Context of the CGE Literature

The magnitude of changes in overall national welfare (as shown in Table 6.3)
lies within a range of one to three percent of the initial level. This range is consistent
with standard welfare effects in applied general equilibrium analysis. For example, the
2002 Global Economic Prospects (World Bank, 2001) indicates that world trade
liberalization would increase real income in developing countries by 1.6% (static gains).
These gains could reach 5% if productivity is allowed to grow in response to trade
reforms. Goldin, Knudsen and Van der Mensbrugghe (2003) revisit the 2002 Global
Economic Prospect estimates and project a 1.4% gain in world real income from firll
merchandise trade reform. According to these authors, the gains in low income countries
would average 3%. Also, Mattoo and Subramanian (2004) summarize findings that.
suggest that a 50% reduction in developed countries’ domestic support programs would
increase per capita GDP by 3% in the least developed countries (LDCs). A recent
review by Cline (2004) also suggest that trade liberalization in developed countries
would result in less-than-one percent static welfare gains in developing countries. Total
gains (including dynamic gains) would range fi'om one to three percent.

Specifically to Sub-Saharan Africa, Clines (2004) estimates that full trade
liberalization in the world would increase per capita GDP by 1.4%. In terms of
individual countries effect, the increases would be 3.2% in Mozambique, 1.3% in
Uganda, and 4.1% in Tanzania. The only CGE result available in Mali is the one

presented by Decaluwé, Dissou and Patry (1998) regarding the implementation of a

108

F

common external tariff (CET) regime in West Africa. The authors distinguish three
scenarios: (i) the creation of a free-trade area among WAEMU countries; (ii) the CET
regime itself (free trade area within WAEMU and common external tariff on non-
WAEMU imports); and (iii) a unilateral, indiscriminate liberalization. Their results
indicate that welfare would remain unchanged in the WAEMU free trade area scenario.
It would rise slightly (by about 0.2%) under the CET regime, and by nearly 1.3% under
overall liberalization.

It appears from these various welfare impacts that the Malian estimates are close
to those commonly found in the literature, even it is generally very difficult to compare
results from applied general equilibrium models. These models are usually different in
many ways, including size, regional and sectoral coverage, policy instruments, time
horizon, market structure, closure rules, database and functional form). These
differences will necessarily yield different results, but it is reassuring that the Malian
CGE results fall within the ranges generally found in the literature.

6.4.2. No and Full Market Reforms: IFPRI-l and IFPRI-2 Scenarios
(i) Overview of the Scenarios

The two IFPRI models have common features with the FAPRI and OECD
scenarios. In addition, IFPRI models explicitly assume an increased rate of urbanization,
which is expected to shift food consumption toward high-value commodities, such as
livestock products. The models firrther expect farm productivity to grow at a declining
rate over the projection period, i.e. 2020. This gain in productivity, combined with some
increase in production in low-cost regions of the world as trade barriers fall, is expected

to increase global supply and drive down the world price of most commodities.

109

Livestock products constitute an exception, as livestock prices will increase due to faster
growth in demand relative to supply.

The IFPRI models capture agricultural policies through the aggregate producer
and consumer subsidy equivalents. These subsidies, which create a price wedge between
domestic and world prices, are maintained under IFPRI-1 scenario and are completely
eliminated in the IFPRI-2 full liberalization scenario. IFPRI trade policies are less
explicit, but they are implicit in the implementation of IFPRI agricultural policy reforms
(i.e., maintaining or eliminating the producers and consumer subsidies). Table 6.5 shows
the projected magnitudes of changes in commodity prices under the two IFPRI
scenarios. Cereal prices are expected to fall, with higher rates of reduction in IFPRI-l
scenarios. Conversely, livestock are expected to rise, with higher increases in the IFPRI-
2 scenario.

Table 6.5: Projected Changes in World Commodity Prices: IFPRI-l and IFPRI-2
Scenarios (% changg from base)*

 

 

Commodity groups IFPRI-1 IFPRI-2
Coarse grains -29.7 -24.0
Rice -25.1 -14.7
Other industrial agricultural products -3.8 -3.1
Cotton - -
Cattle, sheep and goats 19.0 40.4
Other livestock -0.7 -5.5

_Agro-industry and beverages - -
Source: Based on Table 3.3, Chapter 111.

(ii) Change in Welfare in the Two IFPRI Scenarios

Overall welfare is expected to increase in Mali under the two IFPRI scenarios
(1.7% for IFPRI-1 and 1% for IFPRI-2). There are substantial differences between the
two scenarios regarding the cross-household distribution of this welfare gain. While

IFPRI-2 shows a welfare gain patterns similar to those of the FAPRI and OECD

110

scenarios (see Section 6.4.1), changes in household welfare are reversed in the IFPRI-1
scenario (Table 6.6).

Table 6.6: Equivalent Variation: IFPRI-1 and IFPRI-2 Scenarios
(“/0 change from base)

 

 

 

Global (WTO)
Households BASE IFPRI-1 IFPRI-2
Bamako 139.5 -5.0 3.6
Other urban 93.1 -2.0 5.9
Rural S-K 321.5 9.1 -0.3
Rural River 273.9 0.8 -0.4
Other rural 228.7 -1 .8 1.2
Total 1056.8 1.7 1.0

 

To a large extent, this result is puzzling because the magnitudes of change in the
cereal prices are similar under the two IFPRI scenarios. One would expect that
reductions in the prices of the main staples (cereals) would increase the real incomes of
net urban consumers and translate into higher welfare for these consumers. This was,
however, not the case, since other forces—such as changes in the real exchange rate—
have had a much greater influence on rural and urban welfare in the IFPRI-1 and IFPRI-
2 scenarios. For example, the depreciation negatively affected agro-industrial imports,
which fell by nearly 40% under the IFPRI-1 scenario. Conversely, these imports
increased by more than 26% in IFPRI-2 scenario where the real exchange rate
appreciated (Appendix 6.5). With at least 15% of the budget share (Chapter V), agro-
industrial products are important components of household consumption. These
products, which are based on value-added processing, tend to have higher income
elasticity. Thus, changes in the consumption would closely reflect changes in
households’ incomes.

In addition to changes in commodity prices and in the real exchange rate, the

resulting changes in factor incomes also contributed to observed movements in welfare

111

under the two scenarios (Table 6.7). For example, incomes to the Rural Kayes-Sikasso
composite farming factor would increase by 7.7% under IFPRI-1 scenario (Table 6.7).
This change is translated to 9.1% welfare gain (Table 6.6), offsetting the negative
effects of the fall in real food prices under the same scenario.33 Also, returns to wage
labor are expected to fall by more than 17%, contributing 5% and 2% falls in welfare in
Bamako and other urban zones, respectively.

Table 6.7: Disaggregated Factor Income Distribution: IFPRI-1 and IFPRI-2
Scenarios (% change from base)

 

 

 

 

Items FCAPI FWLAB FIENT FOUAC FFCKS FFCRV FFOTH FORAC
Initial factor return 119.3 184.1 46.6 48.1 310.6 232.2 173.7 80.0
Scenarios Percentage change from base
IFPRI-1 -7.5 -17.4 6.7 12.8 7.7 2.7 1.9 6.9
IFPRI-2 -9.3 14.7 9.1 8.0 -0.4 -1.2 -2.0 -0.1

 

Notes: Factors include: capital (FCAPI) and wage labor (FWLAB), as well as composite factors for
independent entrepreneurship (FIENT), other urban activities (F OUAC), Rural Kayes-Sikasso farming
(FFCKS), Rural River farming (FF CRV), Other Rural farming (FFOTH), and other rural activities
(FORAC).

6.4.3. Duty-Free Access and Preference Erosion: DFA and EEP Scenarios
(i) Overview of the Scenarios

As explained in Chapter III, the two market access regimes (DE A and EEP)
affect the economy mainly through their effects on the net price received by Malian
exporters. The EEP export regime will increase existing duties facing the Malian exports
in the importing countries, leading to a reduced net price of these exports. Conversely,
the duty-
free regime will eliminate these duties, thus increasing the net price of exports.

(ii) Change in Welfare in the DFA and EEP Scenarios

 

’3 These effects are also driven partly by the nearly 10% expected rise in the domestic price of “other
livestock” products under the IFPRI-1 scenario. In the Malian model, the Rural Kayes-Sikasso household
produces but does not home-consume their livestock output. They will, therefore, derive higher revenues
from selling this output in a market where the price is expected to rise by 10%.

112

Total welfare in Mali would increase by 1.3% if the Malian had effective duty-
free access to its export markets for cotton and other industrial agricultural products in
1997. The gain would be concentrated in urban households, with some rural household
in the Kayes-Sikasso and the River zones incurring some losses. Conversely, the
country would lose about 3.5% in welfare if it did not benefit from preferential access to
its major export markets. This loss would affect all households, with high impacts in
urban households (Table 6.8). The EEP regime is clearly less favorable than the base

year export regime, which was assumed to represent a preferential trade regime. Thus,

 

the 3.5% decline in welfare may be viewed as the welfare cost of potential erosion in
tariff preferences currently enjoyed by the Malian government.

Table 6.8: Equivalent Variation: EEP and DFA Market Access Scenarios

 

 

 

 

(”/o charge from base)

Preference Rgg’ms
Households BASE DFA EEP
Bamako 139.5 10.1 -8.7
Other urban 93.1 6.8 -6.4
Rural S-K 321.5 -0.8 -O.3
Rural River 273.9 -1.8 -2.5
Other rural 228.7 0.4 -5.0
Total 1056.8 1.3 -3.5

 

The EEP regime may be viewed as a negative shock to the country’s export
earnings. This shock would create an imbalance in the cmrent account balance, forcing
the real exchange rate to depreciate, reducing imports and restoring equilibrium in the
current account balance. This depreciation affects urban welfare by weakening the
purchasing power of imports (e. g., agro-industrial imports—see Appendix 6.5). The
depreciation will also weaken the purchasing power for intermediate inputs, which
would prevent rural producers from taking advantage of a more favorable export

environment brought about by the depreciation.

113

The strong correlation between factor returns and household welfare, discussed
earlier under the first four scenarios, has also been evidenced in the two market access
scenarios. For example, the returns to wage labor would increase by nearly 24% if
exports opportunities in Mali are expanded under the duty-free access scenario. These
returns would fall by nearly 16% if existing trade preference margins were to be
removed (Table 6.9).

Table 6.9: Disaggregated Factor Income Distribution: EEP and DFA Market
Access Scenarios (% change from base)

 

 

 

Items F CAPI FWLAB FIENT FOUAC FFCKS FFCRV FFOTH FORAC
Initial factor return 119.3 184.1 46.6 48.1 310.6 232.2 173.7 80.0
Scenarios Percentage change from base
DFP -7.9 23.9 -2.1 ~5.8 -1.0 -5.6 -7.0 -2.2
EEP -16.5 -15.9 3.2 6.8 2.5 0.0 -0.5 3.8

 

Notes: Factors include: capital (FCAPI) and wage labor (FWLAB), as well as composite factors for
independent entrepreneurship (F IENT), other urban activities (FOUAC), Rural Kayes-Sikasso farming
(FFCKS), Rural River farming (F FCRV), Other Rural farming (FFOTH), and other rural activities

(F ORAC).

6.4.4. Government Policy Reforms: BAN, CET and GINV Scenarios
(i) Overview of the Scenarios

This examines the welfare and food security impacts of three government
policies, which include: (i) applying the common external tariff (CET), (ii) imposing
cereal export ban (BAN), and (iii) increasing public investment by 50% (GINV). While
Chapter III discusses in detail the rationale underlying the choice of these three
scenarios, a brief justification of the scenarios is presented in this section.

First, the Common External Tariff regime (CET) represents a harmonized tariff
regime between eight West African countries. The regime eliminates tariffs on internal
trade between these countries, while imposing a common tariff on non-member imports.

The CET simulation uses tariffs rates fiom Chapter HI, and synthesized in Appendix

114

6.7.34 The simulation results are interpreted as the potential impacts of the CET regime,
were the regime in place in the base simulation year (i.e. 1997). Since the CET rates are
based on WAEMU averages, they may be viewed as higher than rates computed from
the average rates of all ECOWAS countries. The reason is that the scope of intra-
ECOWAS trade liberalization is likely to be higher than the intra-WAEMU
liberalization. In such a case, the CET simulation results reported in this study would
probably understate the impacts of a fully-implemented ECOWAS free trade area. On
the other hand, the ECOWAS and WAEMU free trade areas are not fully implemented
in reality. In such a case, the simulated results would probably overstate the real impact
of the WAEMU CET regime.

Second, banning cereal exports (BAN) increasingly appears implausible in Mali,
as the country moves toward greater market liberalization. Yet, there is evidence that the
country has resorted to this option in the past, and it may still ban cereal exports if faced
with sudden food shortages. This scenario would clearly be more relevant if combined
with conditions (such as adverse climatic shocks) that will force its use by the
government. The Malian CGE model, however, is not designed to address unexpected
shocks to the production. Simulation results from this scenario will therefore reflect
what changes a ban on cereal exports would induce in a normal (average) production

year.

 

3’ The CET regime was implemented by weighting the CET tariffs with the share of the rest of the world
in the total imports of each commodity. Under the CET regime, intra-West African tariffs are eliminated
whereas a harmonized tariff regime is applied to imports originating from outside the West African
region. Thus, for commodities which are 100% imported from outside the region (e.g., rice), the CET rate
would remained unchanged. Commodities that are entirely imported fi'om within West Africa (e.g., fish
and forestry and gathering products) will enter the region duty-free. The zero-duty on regional imports
will lower import duties on aggregate imports for commodities that are sourced from the two regions
(e.g., coarse grains and agro-industrial products). Appendix 6.7 shows more details.

115

Finally, the GINV represent as scenario of 50% increases in public investments.
Increased public investments (in “other manufacturing goods”, “building and public
works”, “transport and telecommunication”, and “nontraded services”) are expected to
affect the price of intermediate inputs, which will in turn affect their demand, and
thereby the level of activity. As explained Chapter III, the 50% increase is based on a
historic annual rate of 5%, cumulated over ten years.

(ii) Change in Welfare in the CE T, BAN and EEP Scenarios

Overall welfare is expected to increase under the three scenarios (Table 6.10).
The highest rise would occur under the CET regime (3%), followed by the scenario of
50% increased goverrnnent investments (1%). Total welfare would remain largely
unchanged if cereal exports were banned. As before, disaggregated effects would vary
with households. While all households would benefit from increased public investments
(with a very slight decline in the Other Rural households), the losses from cereal export
ban would be concentrated in rural (producing) zones (over 2% in Rural Kayes-Sikasso
and River), while gains will be reaped by urban dwellers (over 6%). Higher urban gains,
in order of nine percent or more, are expected under the CET regime. Rural River
households would lose about 1% of the welfare under the same regime.

Table 6.10: Equivalent Variation Under Government-Controlled Scenarios

 

 

 

(% cthg from base)
Household BASE CET BAN GINV
Bamako 139.5 13.9 7.1 0.5
Other urban 93.1 8.6 6.4 3.3
Rural S-K 321.5 1.3 -2.1 2.1
Rural River 273.9 -1.2 -2.5 0.2
Other rural 228.7 1.5 -0.2 -0.2
TOTAL 1056.8 3.0 0.2 1.0

 

A combination of changes in final and intermediate demand prices, as well as

adjustments in factor incomes, will again explain most the expected movements in

116

welfare under these three scenarios. Most of the gains to urban households, both under
the CET and BAN scenarios, would come from increased wage incomes (19% to 20%
rise) and decreased real commodity prices.35 Also, there would he falls in factor
incomes in the three rural households (BAN Scenario), and in the River and Other rural
households (CET Scenario). For example, the Other Rural households would lose more
than 6% of the composite farming factor income under the CET scenario and more than
7% with the BAN regime (Table 6.11). Additional results are shown in Appendices 6.8
and 6.9.

Table 6.11: Disaggregated Factor Income Distribution Under Government-
Controlled Scenarios (% change from base)

 

 

 

Items FCAPI FWLAB FIENT F OUAC F FCKS FFCRV FFOTH FORAC
Initial factor return 1 19.3 184.1 46.6 48.1 310.6 232.2 173.7 80.0
Scenarios Percentage change from base
CET -3 .4 20.2 -0.3 -6.2 0.8 -5.5 -6.4 0.0
BAN -4.2 18.7 4.0 —0.7 -2.3 -6.4 -7.5 -1.3
GINV 2.7 -2.4 10.2 7.1 1.7 -2.4 -2.4 4.5

 

Notes: Factors include: capital (FCAPI) and wage labor (FWLAB), as well as composite factors for
independent entrepreneurship (FIENT), other urban activities (F OUAC), Rural Kayes-Sikasso farming
(FFCKS), Rural River farming (FFCRV), Other Rural farming (FFOTH), and other rural activities
(FORAC).

These functional income distribution effects are exacerbated by the impact of
reduced commodity prices on household incomes. Reduced commodity prices would
decrease the value of households own production. Because own production is an
important component of real household incomes in rural areas, reduced valuation is
equivalent to lower the incomes received by households.

Banning cereal exports would lower their domestic prices (in order of 3%),

which in turn will negatively affect rural income while boosting real incomes in urban

 

35 The ban on cereal exports would generate domestic marketable surplus, which would increase the
volume of cereal trade. The trading activity pays 10% of its income to wage labor, so an increase in
trading activities would translate into higher wage incomes. Compared to cereal exports, domestic cereal
sales have additional linkages in the economy, and this explains the larger returns to trading activities in
this scenario.

117

areas. By closing the surplus-controlling circuit, which export markets represent, the ban
would generate marketable surplus domestically, thus driving down prices. In addition
to the price effect, returns to rural farming factors would decrease, while wage labor will
gain in incomes. The overall effects are welfare losses in rural areas and gains in urban
households. Banning exports would therefore penalize sellers (concentrated in rural
zones) while favoring purchasers (mostly in urban zones).

Finally, welfare in Mali would positively respond to increased government
investments (GINV Scenario), although the effects are likely to be small (1% welfare
gain). Public investments would result in lower prices of intermediate demands for key
activities, such the production of “other trading services” (-11.7%) and livestock
keeping (-9.5%). The investments would also lower the intermediate demand prices in
the production public utilities (as indicated by a 4.6% fall in these prices for the
“electricity, water and energy” activity), and public services (a nearly 1% fall the
“nontraded services” activity). A small cost reduction (-0.7%) is also expected in the
delivery cost of transportation and telecommunication services (Table 6.11). Cost
reductions at activity level would translate into higher profits (factor incomes) for all
factors, except wage labor and the composite farming factors in the River and Other
Rural zones. The decreased wage incomes would negatively affect wage-dependent
urban dwellers, but the effects would be tempered by gains to the “independent
entrepreneurship” and the “other urban activities” composite factors (Table 6.11). The
latter factor paid most of its income to other urban households, who would gain a 3.3%

increase in welfare. Government investments would also translate into lower consumer

118

prices for some commodities (see Appendix 6.10 for details), which would strengthen
effective household demands, and thereby increase real household incomes.

Although lower prices would induce lower factor incomes in producing regions
such as the River and Other Rural areas, they would also strengthen consumption. These
two opposing effects would leave welfare nearly unchanged in these two areas. Thus,
among all policy options examined in this study, increased public investments may lead
to the most Pareto-compatible distributional effects: the policy would unambiguously
increase welfare in the urban and the Rural K-S households, without worsening it in the
other two households.

6.4.5. Food Security

Changes in the state of food security are measured in two ways: changes in
households’ food consumption, and in the country’s capacity to pay for current food
imports out of export earnings. The presentation in this section closely follows the
discussion in the four previous sections. Specifically, the two partial trade reform
scenarios (FAPRI and OECD) are compared against each other, and so are the two
IFPRI scenarios, and the two market access or preference regime scenarios (DFA and
EEP). Food security results are also discussed with respect to the three government
policy scenarios (CET, BAN and GINV). The results of all nine scenarios are shown in
detail in Table 6.12 and 6.13, and this section makes frequent reference to these two
tables.

Household food consumption is divided into eight food categories: (i)-coarse
grains, (ii) other food (roots, tubers, fruits and vegetables), (iii) rice, (iv) other industrial

agricultural products (e. g., groundnuts), (v) beef, sheep and goats, (vi) other livestock,

119

Table 6.12: Changes in Households’ Food Consumption Under Global, Bilateral,
and Government Trade Policy Scenarios (% change from base)

 

Trade Reform Scenarios

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Partial No and Full More or Less Government Policy
Household food BASE Reforms Reforms Preferences Reforms
demand FAPRI OECD IFPRI] IFPRIZ DFA EEP CET BAN GINV
Coarse gains 178.1 6.3 0.5 5.7 7.6 0.2 -4.1 1.5 0.3 0.0
Bamako 10.6 20.1 11.8 -4.3 14.4 12.9 -1 1.0 14.8 1 1.8 0.3
Other urban 7.0 18.8 10.3 -0.8 17.3 10.0 -8.6 10.0 11.3 2.6
Rural S—K 67.6 4.0 0.7 13.1 6.6 -0.3 -O.7 2.0 -O.7 1.6
Rural River 59.3 3.8 -2.7 3.7 6.0 -2.6 -4.0 ~2.1 -2.3 -1.2
Other rural 33.7 8.2 0.3 -O.9 8.1 0.3 -7.9 0.6 0.6 -1.8
Other food 75.6 9.2 5.1 8.7 2.4 5.2 -4.5 5.1 3.0 6.4
Bamako 8.3 23.6 16.3 0.9 7.8 18.1 -10.8 18.3 13.9 7.9
Other urban 4.8 21.3 14.0 3.6 10.5 14.3 -8.3 12.8 12.8 9.2
Rural S-K 31.7 5.8 4.2 14.7 0.8 3.5 -1.7 4.2 1.1 7.5
Rural River 22.7 5.5 0.7 5.9 0.8 1.1 -4.2 0.4 -0.4 4.2
Other rural 8.1 11.1 4.4 3.6 2.8 4.9 -7.6 4.0 2.8 4.8
Rice 100.5 8.2 5.7 3.9 4.2 1.0 -5.0 8.0 0.9 0.2
Bamako 10.2 26.3 19.9 2.0 12.2 9.9 -10.5 27.5 8.9 0.6
Other urban 7.7 21.7 15.9 3.5 13.3 7.3 -7.8 19.0 8.3 2.4
Rural S-K 26.3 3.3 4.0 11.0 1.8 -0.7 -1.2 6.0 -1.1 1.8
Rural River 19.2 3.3 0.9 3.4 1.7 -2.5 -3.7 2.3 -2.3 -0.6
Other rural 37.0 6.4 3.4 -O.3 3.2 0.2 -6.3 4.8 0.4 -1.1
Other 1nd. ag 19.5 4.7 3.5 4.7 0.0 3.3 -7.0 2.0 1.1 -3.0
Bamako 2.0 15.1 12.0 -3.6 4.9 13.1 -13.0 12.0 9.5 -2.7
Other urban 0.9 13.4 10.1 -0.3 6.9 9.9 -10.0 7.5 8.6 -0.8
Rural S-K 8.6 2.1 2.8 9.8 -1.4 1.9 -4.6 1.2 -0.3 -2.2
Rural River 3.9 2.0 0.1 3.1 -1.2 0.1 -6.1 -1.4 -1.4 -3.7
Other rural 4.1 5.7 2.8 0.9 0.0 2.8 -9.1 0.9 0.9' -4.8
Beef, sheep, goats 75.1 7.3 23.7 8.2 2.1 4.5 -5.2 8.3 4.4 7.5
Bamako 8.1 22.5 61.2 10.4 7.2 18.8 -9.7 28.9 17.5 14.6
Other urban 6.1 19.3 48.7 10.7 9.5 14.1 -7.3 20.3 15.2 14.0
Rural S-K 16.1 5.8 36.7 24.6 0.6 4.4 -1.1 12.0 4.0 13.4
Rural River 22.1 1.6 6.7 3.0 -0.2 -0.8 -4.1 0.2 -0.6 3.2
Other rural 22.7 5.3 10.8 0.2 1.4 1.9 -6.9 3.1 2.1 3.2
Other livestock 24.5 13.8 0.7 -3.1 6.1 1.2 4.8 16.5 11.4 10.1
Bamako 3.1 26.4 8.1 -9.6 l 1.4 9.9 0.5 31.4 22.2 12.0
Other urban 1.8 22.5 6.5 -6.0 12.6 7.0 1.7 23.1 19.2 12.1
Rural S-K 9.5 9.4 -0.9 0.9 4.0 -0.8 7.2 13.8 7.7 10.4
Rural River 6.7 12.0 -1.5 -3.7 4.9 -0.8 5.7 10.6 7.4 8.4
Other rural 3.3 13.6 -O.6 -6.0 6.0 0.0 2.0 16.2 12.0 10.0
Fish 33.7 -8.3 9.7 1.1 4.0 23.0 -0.1 9.0 5.1 6.5
Bamako 2.5 -3.7 23.4 -4.4 -1.8 46.6 -3.8 24.8 16.3 9.5
Other urban 2.5 -2.6 19.2 -1 .5 1.3 36.7 -2.2 17.7 14.2 9.9
Rural S-K 6.3 -10.8 9.8 8.2 -5.5 23.4 3.6 9.4 3.6 8.7
Rural River 1 1.9 -9.1 5.2 1.2 -4.7 15.8 0.5 4.3 1.7 4.9
Other rural 10.4 -8.2 9.1 -1.5 -4.1 22.0 -1.6 8.1 4.8 5.4
Agro—lndustry 82.8 0.0 -0.1 ~0.4 0.0 ~0.3 -1.1 -3.8 -5.1 -2.6
Bamako 0.6 11.2 10.8 -7.6 5.2 11.5 -7.0 1.1 -1.6 -3.5
Other urban 0.3 10.3 9.2 -4.0 7.7 8.8 -4.8 -0.8 -0.9 -1.5
Rural S-K 12.3 -1.8 1.0 6.8 -0.9 -0.1 2.1 -4.8 -6.6 -1.7
Rural River 34.8 -1 .5 -1.8 0.0 -0.8 -1.9 -0.4 -6.0 -6.5 -2.9
Other rural 34.9 1.9 0.9 -3.2 0.8 1.0 —2.8 -5.6 -6.2 -3.9

 

120

(vii) fish, and (viii) agro-industrial products, such as sugar and cooking oil (Table 6.12).
The table shows the percentage deviation in food consumption between the base year
and the nine scenarios examined in this study.
(i) Changes in Food Consumption in the FAPRI and OE CD Scenarios

For the country as a whole, and with minor exceptions, food consumption is
expected to increase in both the F APRI and OECD scenarios. Exceptions include a more
than 8% fall in fish demand in the FAPRI scenario, and a very marginal decline in the
consumption of agro-industrial goods OECD scenario (Table 6.12). This aggregate
national impact hides, however, important cross-household differences. Following a
pattern identified earlier with welfare impacts, changes in food consumption would be
larger in urban zones (Bamako and other urban households). Under the F APRI scenario
for example, all positive changes in food consumption are expected to be greater than
10% for the two urban households. The magnitude of changes would average more than
20% for five out of the seven food categories for which positive changes are expected.
Under the same price change scenario, average positive changes in food consumption in
rural‘households would be as low as 1.6% for beef, sheep and goats in the Rural River
region, and as high as 13.6% for “other livestock” products in Other Rural households
(Table 6.12). Earlier discussions emphasized the fact that expected increases in welfare
in urban and “Other Rural” households could be attributed to combination of lower real
agricultural prices, real exchange rate appreciation, and higher factor incomes. This
positive welfare impact would translate into increased real consumption (Appendix 6.6),

both for food and nonfood products.36

 

’6 Changes in households’ consumption of nonfood products are not shown, as the interest of the study
lies primarily in food products.

121

With the exception of the Rural River households, the consumption of coarse
grains, other food, rice, other industrial agricultural products, as well as this of “beef,
sheep and goats” are expected to increase in all households under both the FAPRI and
OECD price change scenarios. With sharper price falls for crop-based products under
the FAPRI scenario, the corresponding increased consumption would generally be
larger than the increases with the OECD scenario. Similarly, the consumption of other
livestock products would increase in response to falling prices under the F APRI
scenario, while declining in ma] areas in response to a slight rise in prices under the
OECD scenario. Contrary to intuition, however, urban consumption of other livestock
products is expected to increase, despite the slight price rise. As explained earlier,
domestic conditions ultimately determined movements of consumer prices, which in
turn determined the changes in the final demand. Consumer prices of other livestock
products fell by more than 25% (Appendix 6.3), resulting in nearly 14% rise in overall
demand (Table 6.5) under the FAPRI scenario. Both final prices and demands remained
largely unchanged in the OECD scenario. Also counterintuitive are the substantial
increases in the consumption of beef, sheep and goat products under the two scenarios,
despite the 6.4% and 9.6% rises in world prices under F APR] and OECD scenarios,
respectively. Again the price rises—whether modeled as affecting both imports and
exports or only imports—would result in lower domestic prices, which in turn triggered
higher demands. Here also, the downward pressure on livestock prices may result from

the overall reduction in cereal prices. 37

 

’7 The expected decline in the consumer prices of most livestock products with its associated increased
demand, even as the world prices these products are expected to increase, raises question about the
equilibrium process at work in the model. Initial reaction to higher prices would normally be a short-run
cut in supply, as producers hold back breeding stocks from the market. By doing so, they might

122

In conclusion, despite expected decline in fish consumption in all households
and in the consumption of agro-industrial products in the Rural Kayes-Sikasso and
River households, the FAPRI price change scenario would unambiguously increase food
consumption in Mali. Similarly, despite a fall in the consumption of a few products
under the OECD price change scenario, overall food consumption would improve in all
households (except perhaps in the Rural River households, where consumption fell by
2.7% for coarse grains, 1.5% for other livestock, and 1.8% for agro-industrial products).
(ii) Changes in Food Consumption in the Two IFPRI Scenarios

As was the case with the F APRI and OECD scenarios, food consumption in Mali
is also expected to increase under the two IFPRI scenarios, although exceptions exist
(the demand for “other livestock” would fall 3.1% in IFPRI-1 scenario and the demand
for fish would decline by 4% in the IFPRI-2 scenario.) Again, there are important cross-
household differences regarding the magnitude of changes in food consumption.

Also important are the expected increases in agro-industrial food consumption in
urban areas in the IF PRI-2 scenario, and a fall in urban agro-industrial food demand in
the IFPRI-1 scenario. With its relatively higher income elasticity (e. g., as compared to
coarse grains which are mainly sourced from own production), the demand for agro-
industrial products is highly dependent on fluctuations in households’ incomes. The
reduced urban consumption of agro-industrial products in the IFPRI-1 scenario may be

viewed as a direct consequence of the negative shocks to factor incomes (and thereby

 

“overshoot” in their supply response to higher prices, thus ultimately leading to lower prices. But the
overshooting is not in itself an equilibrium result, and cannot be a solution of this CGE model. The most
plausible explanation remains, therefore, the induced effects of lower cereal prices. In any case, it would
be a useful extension to investigate the time-dynamic path of breeding stock adjustment, an issue not
addressed in the present static study.

123

total income) in urban households. These shocks appear to be reversed as one switches
from the IFPRI-1 (status quo) to IFPRI-2 (full trade liberalization) scenarios.

One of the main difference between IFPRI-1 and IFPRI-2 scenarios lies in a
sharper increase in the world prices of livestock products in IFPRI-2, particularly for the
“beef, sheep and goats” (19% against 40%, as shown in Table 6.2). The increased world
prices appeared, however, not to be transmitted to the domestic economy, where
consumer prices for “beef, sheep and goats” fell by about 20% and 3% under IFPRI-1
and IFPRI-2 scenarios, respectively (Appendix 6.3). This lack of transmission reflects
the assumption that price shocks affect the domestic economy primarily through
imports. But since there was no importing activity involving “beef, sheep and goats”,
the expected world price changes for this product were not directly recorded in the
domestic economy. The observed prices resulted fiom the relative demand of all
products (both agricultural and non agricultural) in the general equilibrium. There
appeared to be a downward pressure on the domestic price of livestock products, as
other products (particularly cereals) became less expensive in the domestic market. This
is because lower cereal prices would reduce the real incomes of rural producers, thus
shifting downward the market demand of livestock products. This shift will particularly
be important, because livestock products are generally luxury goods in Malian
households, with income elasticity greater than unity (Carnara, 2004). Rural households
consume about 70 percent of the total livestock and fish demand in Mali (see Appendix
5.4), so that overall a reduction of their income could translate into overall decline in

livestock prices.

124

The assumption that international prices are transmitted to the domestic
economy through imports is restrictive. An alternative (and a more realistic)
specification would be to apply the expected changes in world prices to both imports
and exports. This experiment was performed, and results suggest that the domestic price
of livestock products would still decline, inducing increases in consumer demand of
these products. Some of the implications of this result are discussed later in this section,
but one should already note that this result is puzzling. Mali exports its livestock
products to the West African regional market, with Cote d’Ivoire absorbing a substantial
amount of these exports. It is expected that increases in the world price of livestock
products will be transmitted to import demand prices in importing countries in West
Africa. Beef import prices in care d’Ivoire, for example, are directly linked to the
Malian cattle export prices, so increases in the first should be reflected in increases in
the second. A rise in export prices would in turn be reflected in increases in domestic
livestock prices. The Malian livestock prices (particularly for cattle, sheep and goats)
were instead reduced in the equilibrium. A possible explanation is that while rising
international prices would bid up domestic livestock prices, the equilibrium prices result
from the imperfect substitutability between the domestic and foreign demand. The
results seem to indicate that the rising international prices has tempered, but not offset,
the downward pressure that falling cereals prices put on livestock prices.

To a large extent, this result may be an artifact of the distortions in the social
accounting matrix. For example, the Malian SAM did not show home-consumption of
livestock in the rural Kayes-Sikasso households, shifting all their livestock consumption

to the market. The market demand of livestock is therefore probably overestimated

125

(cattle, sheep and goats share about 8% of the consumer budget in the Kayes-Sikasso
Rural area, compared to 2% or less in other rural areas; see Table 5.11). It follows that
the magnitude of fall in livestock prices is also probably overestimated to a point that it
offsets rises in international livestock prices. The results are likely to different with a
more accurate data on home-consumption of livestock in the Kayes-Sikasso region.
(iii) Changes in Food Consumption in the DFA and EEP Scenarios

For the country as a whole, food consumption of the eight food categories
increased under the DFA scenario. The alternative export regime, that is, the EEP export
regime, would unambiguously decrease food consumption in Mali, mainly due to higher
domestic prices of food (Appendix 6.3). The average country effects, as described
above, hide important differences across households.

First, a shift to the EEP export regime would unambiguously reduce food
consumption in all five households, with larger impacts in urban areas. This result is
consistent with observed evidence in West Afiica, where Diagana et al. (1999) found
evidence of declined cereal consumption as a result of the 1994 devaluation of the local
currency, the CFA franc. Singare et al. (1999) specifically reported a similar evidence
for Mali. The EEP regime had a devaluation-like effect, exerting a 12% depreciation of
the real exchange rate. This led to increased intermediate and final demand prices,
which resulted in fall in consumption. With no productivity growth, moving fi'om the
current preferential export regime to the less favorable EEP regime would therefore
unambiguously decrease food consumption in Mali.

Second, the alternative duty-free export regime (DFA) would lead to increased

food consumption in urban and “other rural” households, while having mixed effect in

126

rural Kayes-Sikasso and River households. The consumption of coarse grains, rice,
other livestock, and agro-industrial products would fall slightly in these latter two
households, mainly due to negative factor income shocks.
(iv) Changes in Food Consumption in the CE T, BAN and GINV Scenarios

Changes in national food consumption would very closely follow the expected
welfare changes discussed in Section 6.4.4. Except for 5% or less falls in the
consumption of agro-industrial products in the three scenarios as well as a 3% reduction
in the consumption of “other industrial products” in the GINV scenario, national food
consumption would increase in Mali under the three scenarios (Table 6.12). The
decreased consumption of the agro-industrial products would be marginal relative to
increases in the quantity consumed of other commodities. From the Malian national
perspective, food consumption would therefore increase with the implementation of the
CET regime, a ban on cereal exports, or a 50% increases in government investments.

From the perspective of household groups, however, the picture is less clear-cut.
With minor decreases in for Rural River households in the BAN scenario (-0.4% for
other food and -0.6% for beef, sheep and goats), the consumption of “other food”, “fish”
and livestock products (beef, sheep and goats; and other livestock), would increase for
all remaining households under the three scenarios. The consumption of the remaining
products in Table 6.12 (except agro-industrial products) would also increase in urban
households under the three scenarios. Thus, food consumption would improve on
average for urban dwellers when CET rates are applied, cereal exports are banned or
government investments are increased. Higher increases are expected with the CET

scenario, followed by the BAN, and finally the GINV scenarios, which are consistent

127

with the expected welfare changes in urban areas (Section 6.4.4). As before, the
increased urban food consumption would be driven by falls in consumer prices
(Appendix 6.10) and increased factor incomes (Table 6.11 above).

In rural areas, Kayes-Sikasso households would benefit from increased food
consumption under the CET regime and the GINV scenario. Incomes to the composite
Rural Kayes-Sikasso factor would increase under these scenarios, which would more
than offset the decreased value of their own production following the induced price
falls. The net food purchasers within this household group would, however, benefit fi'om
commodity price falls, as these falls would strengthen their effective demand for these
commodities. As one can see from Table 6.12, Rural Kayes-Sikasso households faced
with cereal export bans would on average decrease their consumption for coarse grains
(-0.7%), rice (-1.1%), other industrial products (-0.3) and agro-industrial goods (-6.5%).
While the decreased demand of agro-industrial products can be clearly linked to nearly
20% price increases (Appendix 6.10), fall in other consumptions (even as their
associated prices decline) may be traced to reduced factor incomes. For net food
purchasers in Rural Kayes-Sikasso, the negative effects of falling factor incomes on
food demand could outweigh the positive effects of reduced prices that trigger larger
effective demand. Net food sellers, faced with the need to maintain their sales income,
would sell larger quantities of their production, while cutting the own consumption (if
production remained unchanged). Additional results (not reported here) suggest that
overall coarse grain sales would rise by 1.5%, while home consumption would fall by

nearly one percent in Rural Kayes-Sikasso households. The decline in home

128

consumption could even be higher in households that rely heavily on domestic cereal
sales.

The cereal export ban would have more pronounced effects in Rural River
household, where incomes to the main endowment would fall by more than 6%. This
household group would also face a decreased consumption of coarse grains and other
industrial products (in addition to agro-industrial goods) under the CET regime. Under
the GINV scenario, the consumption of agro- and other industrial products would fall in
all households. In addition, the River and Other Rural households would consume less
coarse grains and rice. The driving forces of the changes remain those discussed
throughout this chapter: all scenarios are associated with changes in the prices of final
and intermediate demands (Appendix 6.10), which directly affect the effective food
demand, and indirectly the households’ factor incomes. Similar to welfare effects, the
food consumption effects of increased government investments will most likely favor
consumers in the urban centers (through lower food prices) and farmers in the Kayes-
Sikasso agricultural zone (through increased returns to their main production factor, the
Rural Kayes-Sikasso composite farming factor).

(v) Ratio of Food Imports to Total Exports

The share of food imports in total exports measures, as explained in Chapter II,
the capacity of the country to pay for current imports using export receipts. A decrease
in this ratio suggests a strengthened capacity to pay for current food imports out of
export revenues, as the country would devote a smaller share of these revenues to food
imports. But, as will be shown below, the capacity to pay for current imports may be

strengthened without improvements in food security.

129

In computing the food imports/total exports ratio, food was defined to include
the following commodities: coarse grains (CCORS), other food products (COFOO,
which includes, fruits, vegetables, roots and tubers), rice (CRICE), beef, sheep and goats
(CCASG), other livestock products (COLIV), fish (CFISH) and agro-industrial products
(CAIDT). The commodity “other industrial agricultural products” (CINAG) was
excluded in the definition of food because it contains tobacco (in addition to some food
products), which is not a food product. This omission will have little effect on the
results because the CINAG commodity group constitutes a very small share of
household demand (see base year figures in Table 6.13).

Table 6.13: Ratio Food Imports/Total Exports Under Global, Bilateral, and

Government Trade Policy Scenarios (% change from base)
Trade Reform Scenarios

 

 

Partial No and Full More or Less Government Policy
Household food BASE Reforms Reforms Preferences Reforms
demand FAPRI OECD IFPRII IFPRIZ DFA EEP CET BAN GINV
Real food imports 72.3 19.3 6.8 -22.1 12.4 4.4 -15.8 12.4 -0.3 —3.6
Total real exmrts 479.9 -0.2 2.7 0.3 -2.9 2.4 -0.1 4.3 0.3 -0.2

 

Ratio (% A base) 15.1 +19.5 +4.0 .22.3 +15.7 +1.9 -15.8 +7.9 -0.6 -3.4
Food imports represented about 15% of export receipts in the base year. With the
exception of the IFPRI-1 price change and the duty-free (DF A) export regime, the
capacity to pay for current food imports from export revenues diminished under all the
nine scenarios. In other words, the share of food imports in total exports increased in all
but the IFPRI-1 and DFA scenarios (Table 6.13).
These changes are explained either by changes in food imports, changes it total
exports, or both. But changes in imports appeared to be highly correlated with changes
in the food imports/total exports ratio. For example, the ratio increased by 19.5% under
the F APRI price change scenario. Real food imports rose by about a similar magnitude

(19.3%) under the same scenario. Similarly, the ratio fell by 22.3% under IF PRI-l and

130

by nearly 16% under EEP export regime; real food imports fell by similar magnitude
under the two scenarios, respectively. To a minor extent, changes in real exports
attenuated or exacerbated the magnitude of change in real food exports. This may be
seen under the IFPRI—2, OECD and DFA scenarios. Under IFPRI-2, a 12% increase in
real food imports combined with nearly 3% fall in exports resulted in nearly 16% rise in
the share of food imports in total exports. Under the OECD and DF A scenarios, both
food imports and total exports increased, so that the resulting changes in the food
imports/total exports ratio were smaller than changes to food imports (Table 6.13).

Compared to the first food security indicator, that is, actual levels of household
food consumption, the food imports/total exports ratio showed that a strengthened
capacity to pay for current food imports (that is, a decreased ratio—as under IFPRI-l,
EEP, BAN and GINV scenarios) did not necessarily translate into increased food
consumption. Instead, an opposite effect seemed more likely. Despite the strengthened
capacity under IFPRI-1 and EEP scenarios, real food imports simultaneously declined,
with an ultimate negative effect of food consumption (food consumption unambiguously
worsened under the EEP scenario). Conversely, the capacity to pay for current imports
worsened under the remaining scenarios (FAPRI, IFPRI-2, OECD, DFA and CET). Yet,
in at least one of these scenarios (FAPRI), food consumption unambiguously increased,
mainly due to a combination lower real food prices and higher factor incomes.

These results confirm observations in the literature (e. g., FAO, 2003) about how
a govemment’s goal to reduce its food import dependency (by improving the food
imports/total exports ratio) may be incompatible with the more crucial goal of ensuring

food security (broadly defined as increased food consumption). For example, real food

131

consumption and the capacity to pay for current food imports moved in opposite
directions under the FAPRI and EEP scenarios, implying that the share of food imports
in total exports revenue is not a perfect measure of food security. Therefore, one needs
to be very cautious about interpreting falling import dependency as an unambiguous
sign of greater food security. This food imports/total exports ratio can only be
meaningfully interpreted in light of more fundamental indicators of national and

household food security.

6.4.5. Global Trade Reforms With Productivity Gain in Mali

Up to now, the CGE simulation results are obtained under the assumption Mali
will be insulated from the worldwide gains in productivity, which is the main driver of
the expected decline in global commodity prices in the FAPRI, OECD, IFPRI-1 and
IF PRI-2 scenarios. In this section, we relax this assumption by allowing a 10% gain in
productivity in Mali. The productivity gain scenario is simulated as a 10% increase in
output using the same levels of inputs (i.e. aggregate intermediate input and aggregate
production factors; see Chapter IV for details). We analyze the productivity shock
scenario within the IFPRI-2 price shock scenario. Table 6.14 shows the welfare results
and Table 6.15 show the food security impacts under the productivity shock scenario.

Table 6.14: Welfare Impacts of a 10% Productivity Gain Under the IFPRI-2
Global Commodig' Price Change Scenario

 

 

 

 

Productivity Scenarios

IFPRI-2 IFPRI-2
Households BASE With No Productivity Gain With 10% Productivity Gain
Bamako 139.5 3.6 4.3
Other urban 93.1 5.9 6.0
Rural S-K 321.5 -0.3 -0.1
Rural River 273.9 -0.4 -0.5
Other rural 228.7 1.2 1.1
Total 1056.8 1.0 1.4

 

132

As can be seen from Table 14, a 10% increase in productivity in Mali would add
an extra 0.4 percentage point to the country’s welfare, had the IFPRI-2 prices prevailed
in the base year (i.e., 1997). The price reduction, whether this is driven by a domestic
productivity gain or a fall in the exogenous world market prices, would benefit urban
consumers. Under the IFPRI-2 scenario, a 10% productivity gain in Mali would increase
the welfare gain in Bamako and other urban households by 0.7 and 0.1 percentage
points, respectively.

With respect to the base year, welfare in the two main agricultural regions
(Kayes-Sikasso and River) would fall when agricultural productivity increases in Mali
under the IFPRI-2 price. The gain in productivity would mitigate the fall in welfare loss
for Kayes-Sikasso households (the loss decreases from -0.3% to -0.1% of initial welfare
level). However, under the same IFPRI-2 price change scenario, River households
would lose marginally (the loss increases from -0.4% to -0.5% of the initial welfare
level) from the shock. As before, the combined effects of productivity gain and world
price fall would result in lower domestic prices, which in term determine factor returns
and institutional incomes. In the Rural Kayes-Sikasso, output gain effects would
dominate lower price effects, hence the smaller loss in welfare. In the River households,
the increased output (income) induced by productivity gain would fall short of
compensating the lower-price-induced income losses. Higher productivity could lead to
larger increases in output, which could more than compensate income losses resulting
fi'om lower prices. But the exact effects can only be established through empirical

analyses.

133

Table: 6.15: Food Security Impacts of a 10% Productivity Gain Under the IFPRI-2
Global Commodity Price Change Scenario

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Productivity Scenarios

IFPRI-2 IFPRI-2

Household food demand BASE Wltb No ProductivitLGaln With Productivity Gain
Coarse grains 178.1 7.6 7 .4
Bamako 10.6 14.4 13.9
Other urban 7.0 17.3 16.8
Rural S-K 67.6 6.6 6.5
Rural River 59.3 6.0 5.9
Other rural 33.7 8.1 8.0
Other food 75.6 2.4 2.4
Bamako 8.3 7.8 7.8
Other urban 4.8 10.5 10.5
Rural S-K 31.7 0.8 0.8
Rural River 22.7 0.8 0.8
Other rural 8.1 2.8 2.8
Rice 100.5 4.2 4.2
Bamako 10.2 12.2 12.5
Other urban 7.7 13.3 13.4
Rural S-K 26.3 1.8 1.8
Rural River 19.2 1.7 1.7
Other rural 37.0 3.2 3.1
Other 1nd. ag 19.5 0.0 0.0
Bamako 2.0 4.9 4.9
Other urban 0.9 6.9 6.9
Rural S-K 8.6 -1 .4 -1.4
Rural River 3.9 -1.2 -1.2
Other rural 4.1 0.0 0.0
Beef, sheepfioats 75.1 2.1 2.1
Bamako 8.1 7.2 7.1
Other urban 6.1 9.5 9.3
Rural S-K 16.1 0.6 0.6
Rural River 22.1 -0.2 -0.2
Other rural 22.7 1.4 1.4
Other livestock 24.5 6.1 6.2
Bamako 3.1 11.4 11.5
Other urban 1.8 12.6 12.7
Rural S-K 9.5 4.0 4.1
Rural River 6.7 4.9 5.0
Other rural 3.3 6.0 6.1
Fish 33.7 -4.0 -4.1
Bamako 2.5 -l .8 -l.8
Other urban 2.5 1.3 1.3
Rural S-K 6.3 -5.5 -5.7
Rural River 1 1.9 -4.7 -4.8
Other rural 10.4 -4.1 -4.2
_Ag:mindustry 82.8 2.7 2.9
Bamako 0.6 6.3 6.5
Other urban 0.3 7.5 7.8
Rural S-K 12.3 0.9 1.0
Rural River 34.8 0.8 0.8
Other rural 34.9 2.3 2.4

 

134

1.—

As was the case with welfare effects, the level of food consumption would
change only marginally when the productivity increases by 10% (Table 6.15).
Compared to the no productivity gain scenario, food consumption in the increase farm
productivity scenario would deviate by a maximum of -/+2% percentage points. For
examples, overall rise irrthe consumption of coarse grains would decline by 0.2
percentage points (fi'om 7.6% to 7.4%) in the agricultural productivity scenario. Also,
the increase in agro-industrial demand would increase from 2.7% to 2.9%. The intra-
household distribution patterns of gains and losses would remain substantially
unchanged (Table 6.15).

6.5. Sensitivity Analysis

The sensitivity analysis is performed with respect to model parameters and
macroeconomic closure mics. Model parameters of interest include factor
substitutability (elasticity of substitution between production factors), output
aggregation elasticity and the CES (Armington) and the CET (Powell-Gruen)
elasticities. The sensitivity analysis consists of a 50% decrease and a 50% increase in
the initial value of each of these parameters. Macroeconomic closures refer to the
equilibrium conditions atwork in the factor markets, the saving-investment balance, and
the foreign exchange market. The stability of the initial results is tested under each of
these alternative closurerules. In all cases, the welfare indicator (EV) is used to
illustrate how variations in model parameters or in the choice of macroeconomic
closures would affect deviations from the base year solutions. All the changes apply to

the OECD price change scenario.

135

 

6.5.1. Sensitivity to Model Parameters

As discussed in Chapter IV, the Malian CGE model is based on a nested
production technology in which production factors are combined into value added, and
intermediate inputs into aggregate intermediate demands. Outputs are then generated
from an additional combination of value added and aggregate intermediate inputs. Both
factor aggregation and output production follow a CBS technology, whereas aggregate
intermediate inputs are generated fiom a Leontief technology. Also characteristic of the
production technology is the aggregation of a total output of a commodity produced by
various activities. Regarding production, the sensitivity analysis of interest relates to
factor aggregation and output production. The sensitivity analysis does not cover
parameters involved in the formation of aggregate intermediate demand and the
aggregation of outputs. The formation of aggregate intermediate demand follows a
Leontief production technology with fixed aggregation coefficients derived fi'om the
Malian social accounting matrix. A sensitivity involving these coefficients would be
equivalent to changing the structure of the matrix, a task not undertaken in this study.
Regarding output aggregation, it applies to commodities that are produced from more
than one activity. This does not apply for the modified model presented in this chapter.

As can be seen in Figure 6.1, increased factor substitutability exacerbates
welfare gain in urban areas and losses in rural areas under the OECD price change
scenario. Decreased substitutability would lead to the opposite effect, thus reducing the
variability in welfare effects between urban and rural zones. This is because rural
production factors, such as the composite farming factors, are location-specific are

relatively fixed compared to urban production factors (e.g., wage labor) which are more

136

mobile and more substitutable. Increased factor mobility will therefore work against less
mobile factors, whereas decreased mobility will work against relatively more mobile
factors. The initial results appear highly stable to changes in the second production
elasticity of interest, that is, the elasticity of substitution between aggregate factors
(value added) and aggregate intermediate inputs. Changes in overall EV in the OECD
and other scenarios (not shown here) remained largely identical to their initial levels.
This suggests that, at a national level, the relative mix of aggregate production factors or
intermediate demands may have a little bearing on how trade reform scenarios would

affect welfare (and consequently food security) in Mali.

 

 

 

 

 

 

 

 

 

 

 

16 *
Aal‘nako
A; 12
I”)
a
m
E
E s
Q / . ' .
ea .
i: . '
a , v
5 . ' ’ Total
‘5 4 L /
o
g “—1— + /
e, ‘ ‘ ‘ . __ Rural K-S
> i T fl
T ‘ ~ - ,. ~ Other
‘ “""~—<':_ Rural River
~o
.4 i l l
50% Decrease Initial simulation 50% Increase
Change in Factor Substituability

 

 

 

Figure 6.1: Change in EV in Response to Change in Factor Substitutability

137

(e.g. of OECD Price Change Scenario)

Changes in trade elasticities are indicated in Figures 6.2 and 6.3. The CET
elasticity characterizes the degree to which domestic sales can be converted into
exports, and the CES elasticity shows the degree to which imports can replace sales
from domestic output. The higher these elasticities, the faster domestic sales will be
converted into exports, or be displaced by imports. Regarding CET elasticities, the
Rural Kayes-Sikasso appeared to be the biggest loser from a potential reduction in the
tradability of exports with respect to domestic sales (a more than 10% loss with respect
to initial parameters.) Kayes-Sikasso is the most agricultural region of the country, so
the decreased export opportunities may increase domestic surplus, lower producer
prices, and consequently reduce farmers’ incomes. The effects are comparable to those
observed under the cereal export ban scenario under which both the Kayes-Sikasso and
the River rural households are expected to lose. Surplus-induced lower prices would
benefit urban households. These effects are reversed when the CET elasticity is

increased (Figure 6.2).

138

 

201

 

 

 

 

 

 

 

 

 

 

 

 

§~I6
a
12
E
E
in 8
0
E”
a 4
a
U
30
$4
>
H -8
-12

50% Decrease ln'tial s'mihtion 50% hicrease
Change in Export Tradability (CET Elasticity)

 

 

 

Figure 6.2: Change in EV in Response to Changes in Export Tradability (e.g. of
OECD Price Change Scenario)

Regarding Armington elasticities, changes would also result in differential _
effects in urban and nrral areas (Figure 6.3). Decreased substitutability between
domestic sales and imports would reduce consumption flexibility, thus penalizing urban
dwellers in an international environment where commodity prices are expected to
decline (the OECD scenario). Conversely, increased tradability would work against
rural residents who will experience real income fall following the lower domestic prices
associated with relatively less expensive imports. In practice, both export and import

tradability can be influenced by policy levers, such as the development of grades and

standards.

139

 

16-

 

12

 

 

 

 

-4

 

EV (Percent Change From Base)

 

 

-8 I fi

50% Decrease In'ml sirulation 50% Increase
Change in Import Tradability (Armington Elasticity)

 

 

 

Figure 6.3: Change in EV in Response to Changes in Import Tradability (e.g. of
OECD Price Change Scenario)

Overall, the results are less stable than often thought in CGE analyses (e.g.,
Wobst (2001) and Mbabazi (2002)). Quantitative changes in welfare are important
under alternative factor substitution, export transformation and import substitution
elasticities, and so are the qualitative changes. In many instances, changes in elasticities
would determine whether a household lose or gain fi'om the OECD price change
scenario. It is worth remembering that the Malian CGE model is based on a peculiarly
low quality data, and this may partly explain this relative instability that characterizes
some of the results.
6.5.2. Sensitivity to Macroeconomic Closures

In the base model, all factors are fully employed. In addition, wage labor and
capital were allowed to be mobile across sectors, whereas the composite factors were

constrained to be sector-specific. This constraint was justified on the grounds that

140

composite production factors are not conventional in the neoclassical sense, because
they represent a mixture of household labor, land and capital endowments. Thus, they
are location-based and very specific to activities. They cannot, therefore, be mobile
across sectors. The restriction on factor mobility proved justified when performing
sensitivity analyses in which composite production factors were allowed to move across
sectors. Under such a specification, the Malian CGE model did not have a feasible
solution. Factor mobility would be more meaningful when the various components of
the composite factors were isolated, so that factor returns are truly determined by inter-
sectoral arbitrage opportunities. This would require a redesign of the Malian social
accounting matrix, provided information on classical factor allocation is available.

The foreign market closure in the base model allowed for fixed foreign savings
and flexible exchange rate, which adjusts to balance the trade account. The alternative
closure is to allow the foreign savings to vary, while maintaining a fixed exchange rate.
This alternative closure makes sense in the context of Mali where the nominal exchange
rate is fixed, as the local currency (the CFA flanc—CF AF ) is pegged to Euro as part of
a monetary cooperation between the West Afiican Central Bank and the French
Treasury. But since most international transactions are executed in US dollars (U SD),
the fluctuations in the USD-Euro exchange rate are automatically reflected in the
nominal exchange rate between CFAF and USD. Thus, specifying a flexible exchange
rate seems to be the most realistic in the Malian situation. Notwithstanding the fact, we
examine how the initial results would change were the exchange rate to be fixed. The

welfare results are depicted in Figure 6.4.38 Fixing the exchange rate would flee foreign

 

’8 Other evidence supporting the interpretation of the results are not presented here. The main effect is,
nonetheless, consistent with previous observations, as explained in the remaining part of the section.

141

savings, which would nearly double (a 179% increase) under the OECD scenario. The
inflow of foreign capital would lead to higher appreciation of the real exchange rate
(12% against 3% in the initial simulation), which would result in export contraction. In
addition, lower import prices (flom the OECD scenario) would translate into lower
producer prices. Real urban income would increase in response to lower domestic
prices, which, together with decreased export opportunities, would reduce real rural
incomes. The effects would be most severe in the Rural Kayes-Sikasso, where welfare
effects could be reversed as one moves flom flexible to fixed exchange rate (i.e. for

fixed to flexible foreign savings).

 

 

£3 Initial Sirmlation El Fired Exchange Rate

 

 

 

20

 

 

 

 

 

 

 

 

 

 

 

Change in EV (Percent Change from Initial Simulation)

 

‘5 . H \ \
$4- &,5? 4“” «or
a t» o"
20
Households

 

 

 

Figure 6.4: Comparison of the Changes in EV with Flexible (Initial Simulation)
and Fixed Exchange Rates (e.g. of the OECD Price Change Scenario)

142

The last series of sensitivity analysis involves the savings-investment closure. In
the standard model, the closure is said to be “investment-driven”, that is, savings adjust
endogenously to equilibrate the demand for fixed investments. We consider two
alternative saving investment closure rules, one is saving-driven (i.e., investments are
endogenously determined) and the other is a balanced saving-investment closure (all
components of the total absorption adjust to balance the savings- investment account).
Welfare results under these two alternative closures (along with the initial results for the
OECD price change scenario) are shown in Figure 6.5.

It follows flom the figure that, under the saving-driven closure (savings are
fixed), investments must adjust and clear the fixed savings of the saving-investments
account. The increased investments would mainly come, in rural areas, flom increased
stocks of livestock.39 Rural consumption would then be reduced, resulting in rural
welfare loss. Urban consumption would remain largely unchanged (slight increases
were observed), and the overall national welfare would still improve under the OECD
scenario, albeit by a lower margin. This result, because it is confined to a single period
is misleading, as current investments would have future returns that are clearly not
accounted for when evaluating welfare losses in rural areas. Lofgren, Harris and
Robinson (2002) warn about this possibility and suggest that welfare results are more
reliable when investments are fixed (this is the initial simulation).

The balanced saving-investment closure considered here is investment-driven

(investments are fixed) and it involves uniform adjustment in saving rates and changes

 

’9 Domestic livestock prices would increase under this specification, and this would trigger a short-run
withholding of breeding stocks flom the markets (or vice versa, i.e. increased supply would reduce
domestic supply leading to the price increase—the direction of causality in not clear from the single-
period model used in this study). Detailed results are not shown here.

143

in direct taxes across domestic institutions (based on Equation 4.58, Chapter IV). Using

this closure rule would reverse initial welfare effects in urban areas. These effects would

 

 

 

 

 

 

 

  

 

 

 
    
  
 
  
 
    

 

 

 

 

 

 

 

 

Initial Simulation CI Fixed Savings 13 Bahnced Closure
20
E
E
E
3
gig 10 ‘
5 A
6 .5.
a 2.5
i ..E.
a m
5 0 d :3:
.5 {-
o If
on .-.
t: ..-.
3 I
0 Rural Rural Other Total
K-S River rural
-10
Households

 

 

 

Figure 6.5: Changes in EV with Alternative Closures of the Saving-Investment
Account (e.g. of the OECD Price Change Scenario)

mainly stern flom higher taxes on wage labor, which would reduce wage incomes and
decrease urban welfare. Change in national welfare, however, would remain marginally
positive. The various macroeconomic closures appeared to preserve the direction of
change in overall welfare, though there are differences at household levels. This seems
to confirm observations made in Devarajan, Lewis and Robinson (1990) and Lofgren,
Harris and Robinson (2002) that under full factor employment, various macroeconomic
closures would have no or little effect on aggregate income. Despite this observation,

the Malian CGE model results can be said to be less stable than standard results in CGE

144

analyses, partly due to the understandably lower data quality in the Malian analysis. The
results remain useful in understanding the linkages between global trade policies,
welfare and food security in Mali, even if the magnitudes of expected effects would
vary depending on model parameters and macroeconomic closure rules.

6.6. Summary and Conclusion

This chapter discusses the implementation and the results of the Malian CGE
model. The interest lies in understanding how changes in world commodity prices and
domestic policy reforms would affect welfare and food security in the country. Welfare
changes are measured by the equivalent variation, and changes in the state of food
security were assessed through direct changes in food quantities consumed. The analysis
is done for the country as a whole and for each the five household groups defined in the
Malian social accounting matrix.

Nine global, bilateral and domestic trade policy reform scenarios are analyzed in
the chapter. These nine scenarios are organized into four major groups based on the
nature and source of the policy reforms. First, we compare two scenarios of incomplete
reforms in the world commodity prices (F APRI and OECD). Second, we examine two
comparable scenarios of change in world commodity prices under the conditions of no
and full implementation of WTO agricultural trade reforms (IFPRI-1 and IFPRI-2.).
Third, we compare two scenarios of bilateral trade reforms (DFA and EEP), based on
whether Malian exports would have preferential or non-preferential access to OECD
markets. Finally, we examine three government policy scenarios, including banning
cereal exports (BAN), effectively implementing a common external tariff regime (CET),

and increasing public investments in key sectors.

145

The simulations are performed using IFPRI’s standard CGE model, as discussed
in Chapter IV. The analysis was based on an aggregated version of the Malian social
accounting matrix presented in Chapter V. The Malian SAM was aggregated in order
keep the analysis tractable, and to facilitate the generation of a feasible solution for the
CGE model.

The results identify three interdependent channels through which price and
policy scenarios would affect welfare and food consumption in Mali. These include
changes in domestic prices, in factor incomes, and in the real exchange rates.

Price changes for both final and intermediate demand affect the real incomes of
producers and consumers, often with opposite effects in urban and rural areas. For
example, it was found that reduced world cereal prices in the two partial trade
liberalization scenarios (FAPRI and OECD) would reduce domestic equilibrium prices,
which in turn will mostly benefit urban households in Bamako and other urban zones.
The same results are expected under full trade liberalization in the IFPRI-2 scenario.
Regarding the IFPRI-1 status quo scenario (i.e. no further trade liberalization), however,
commodity price effects appear to be dwarfed by the effects of exchange depreciation.
The depreciation of the real exchange rate (RER) would weaken the purchasing power
of imports, which in turn will lower consumption for net consumers (often urban
households) while increasing incomes for net producers (often rural households). Except
the EEP scenario, the RER would appreciate in all the other scenarios examined. In
general, rural households would lose flom RER appreciation, which would usually force
a contraction in exports and depress the real producer revenues. A depreciation of the

RER would have an opposite effect (c. g. of the IF PRI-l scenario). In the EEP scenario,

146

however, the depreciation would stern flom worsened conditions in the Malian export
markets. This would slow the entire economy and reduce welfare across all households.

Changes in commodity prices also drive the welfare results in the three
government policy scenarios. Banning cereals exports (BAN) would generate domestic
surplus and lower cereals prices. Effectively implementing the CET regime would
reduce average tariffs and reduce domestic prices. Investing in inflastructure would
reduce production cost, which will also reduce domestic prices. In the BAN and CET
scenarios, urban households would again benefit florn reduced domestic prices, whereas
nrral households would lose. The cereal export ban could also have dynamic effects (not
captured in this analysis) of undermining Mali’s reputation in export markets as a
reliable supplier, thereby hindering trade in firture years. In the increased investments
scenario, however, the gains a more evenly distributed between urban and rural
households, because a reduction in production costs could increase both the producer
and consumer surpluses.

Factor incomes would respond to changes in commodity prices. This response is
commonly known as the Stolpler-Samuelson effects. In many cases, wage labor would
gain in income, because it was assumed mobile across sectors (contrary to the composite
production factors that were sector-specific). Factor mobility would permit a more
efficient use of inter-sectoral arbitrage opportunities, hence an increase in its returns
relative to the returns to fixed factors.40 The Other Rural households share about one-
quarter of wage labor incomes, and this proved very useful in compensating or even

offsetting adverse shocks to producer prices observed under the price change scenarios.

 

‘0 One should note that perfect mobility of labor is an extreme assumption as it assumes no private or
social costs to intersectoral or rural-urban migration.

147

In many instances, welfare and food consumption would unambiguously improve in
these households. At the same time, welfare loss and negative or ambiguous food
consumption effects would be observed in the Rural Kayes-Sikasso and River
households.

All these results vary with changes in model parameters and in macroeconomic
closures, and are less stable than often reported in standard CGE analysis. The core
qualitative results remain valid in terms of the channel through which global and
domestic policy scenarios would affect welfare and food security in Mali. Because of
the relative low quality of the Malian data, the magnitudes of the simulated effects need
to be interpreted with caution. Except the EEP export regime, Mali is likely to gain flom
future reforms and productivity gains in world commodity markets. However, the gains
would primarily be concentrated in Bamako and other urban zones (plus the Other Rural
areas in many instances), whereas Rural Kayes-Sikasso and River households would
shoulder most of the losses. Increased public investments would deliver the most

Pareto-compatible welfare and food security results.

148

CHAPTER VII: SUMMARY, CONCLUSION AND RECOMMENDATIONS

“. . .commodity policies are called a game in which politics
overpowers economics and economists. In this characterization,
anything that economists do, such as modeling consequence of
alternative policies, is often thought to be for their own
professional entertainment”.

John E. Lee, Choices, Fall 2002, p. 11.

7.1. The Research Question

This study has attempted to evaluate the potential impacts of WTO and West
African agricultural trade reforms on welfare and food security in Mali. The WTO trade
reforms involve a concerted effort among 145 nations toward increased liberalization of
the world commodity markets through reductions in the level of trade-distorting
supports to agriculture. West Aflican trade reforms underscore the harmonization of
agricultural trade policies among West Aflican nations, with the aim of deepening the
commercial integration in the region. These ongoing reforms are expected to change the
level of welfare and food security in Mali and in other developing countries. But it is not
clear a priori whether Mali would gain or lose flom the anticipated increased
liberalization of the World and West Aflican agricultural markets. In addition, there are
uncertainties regarding the extent to which WTO member countries would implement
their trade reforms. These uncertainties have fueled speculations about the possible
directions and magnitudes of changes in world commodity prices, and in the tariff
protection regimes under which Mali will trade with its partners in West Afiica and the
rest of the world. This study has offered an approach to understand better the nature of
the uncertainties by evaluating changes in welfare and food security in Mali under

plausible scenarios of trade reforms.

149

The approach used in this study is based on a computable general equilibrium
(CGE) analysis, and it consists of three key objectives. First, the study emphasizes the
design of trade reform scenarios based on expected changes in world agricultural prices,
in preferential trade opportunities for the Malian exports, and in government trade and
fiscal policies. A second focus of the study is the construction of a social accounting
matrix (SAM), which is the main database used in the Malian the CGE analysis. Finally,
the study uses the SAM and the CGE model to simulate national and household-level
welfare and food security impacts of trade and domestic policy reforms. Each of these
three objectives is evaluated in sections 7.2, 7.3, and 7.4, respectively. Section 7.5
discusses the achievements and limitations of this study.

7.2. Trade Reform Scenarios

The first objective was to design trade reform scenarios, based on changes in
world commodity prices, in the level of border protection applied to Malian exports, and
in government trade and fiscal policies. There is an infinity of possible trade reform
scenarios, which prompts the need to select a limited number of scenarios for tractable
policy analysis. This study considers nine scenarios, each of which covering a specific
aspect of global, bilateral or domestic policy reforms of interest.

Specifically, the study first compares two scenarios of change in world
commodities prices that would result flom a partial reform in world agricultural
commodity markets (FAPRI and OECD).The magnitude and direction of change in the
F APRI and OECD scenarios depend on of change on a multitude of projection

assumptions which differ in the two scenarios.

150

Second, the study also considers two comparable price change scenarios, based
on the assumptions of no and full liberalization of the world commodity markets
(IFPRI-l and IFPRI-2). Contrary to the first price change scenarios where the
underlying projection assumptions are different, only the assumed level of trade
liberalization differs between the IFPRI-1 (status quo) and IFPRI-2 scenarios (full trade
liberalization).

A third group of scenarios refers the two possible reforms in the bilateral trade
opportunities in Mali. Such reforms would change the level of border protection (tariffs)
facing key Malian exports to the rest of the world (ROW). Mali could face erosion in
the preferential tariff rates that the country enjoys on its ROW exports, and this
possibility is captured in the Effective Erosion of Preferences (EEP) scenario.
Alternatively, Mali could obtain additional preferences in the form of Duty-Free Access
(DFA) to exporting markets in the OECD countries. The DF A scenario is also
considered in the analysis.

Finally, among the infinite policy actions the Government of Mali can take to
manage its economy in rapidly changing world trade enviromnent, this study
emphasizes three scenarios. These include (i) a switch to a common external tariff
regime (CET), (ii) an imposition of cereal export ban (BAN), and (iii) increased public
investments for key sectors (GINV).

The rationale underlying each of these scenarios is discussed in detail in Chapter
HI, which deals exclusively with the design of trade reforms scenarios. The CGE

simulation results in each of these scenarios are interpreted as the effects that the

151

policies and market conditions underlying these scenarios would have on the Malian

economy, were these policies in place in 1997 (the base year used in the simulations).

7.3. Construction of a Detailed Social Accounting Matrix for Mali

Constructing a disaggregated social accounting matrix (SAM) for Mali is the
second main objective of this study. To our best knowledge, this was the first matrix
with such a level of detail in Mali. The matrix included 17 production sectors that
produced 37 commodities, 25 of which were agricultural. In addition, there were nine
production factors, five household types, and six different government policy
instruments (taxes). The matrix also differentiated imports and exports based on their
West Aflican and ROW origins and destinations. The Malian SAM was built flom very
diverse and highly inconsistent data sources, and this created important imbalances
between the demand (expenditures) and supply (incomes) accounts in the initial matrix.
These initial imbalances were corrected using a procedure known as the cross-entropy
(CE) SAM balancing method. The procedure is based on information theory and
consists of finding a balanced matrix with the smallest deviation flom the initial
(unbalanced) matrix. The result of the CE procedure is a balanced social accounting
matrix for Mali.

Although balanced, the matrix presented important structural differences with
the initial matrix. This raises the question about whether the initial or the balanced SAM
represents the “true”, but unobserved, structure of the Malian economy. It is impossible
to answer this question in light of the lirrrited data that were available when the current

SAM was built. Since then, however, the results of new household surveys and input-

152

output tables have been prepared, and their public release is imminent. These new (and
certainly more consistent) data could reduce the observed disparities between the supply
and demand sides of the economy. Improvements could also involve a definition of
more homogenous production sectors, production factors, and household groups. The
current matrix defines household groups based on their location. It also defined
location-based production factors as a composite of classical production factors, such
capital (including land) and labor. Thus, the returns to a composite factor represent
aggregate returns to all of these factors. Besides, the localized nature of these factors
makes them quasi-immobile in the CGE simulations. Such an assumption is reasonable,
given structural rigidities in the Malian economy. Nonetheless, a definition of more
conventional production factors can help sharpen the understanding of inter-sector shifis
in factor allocations in response to reforms in world commodity markets.

7.4. A CGE Analysis of Welfare and Food Security Impacts of Trade Reformsin
Mali

The central focus in this dissertation is the use of a computable general
equilibrium model to simulate changes in welfare and food security under various trade
reform scenarios. Welfare is measured by equivalent variation (EV), and food security is
captured through the changes in food consumption at national and household levels, and
in the capacity of the country to pay for its current food imports. While these indicators
are not perfect measures of changes in welfare and food security, they are useful for
approximating the effects of interest in this study.

The Malian computable general equilibrium model belongs to the class of the so-
called neoclassical structuralist models. These models are based on the neoclassical

general equilibrium of markets. They also include additional structural features, such as

153

home consumption of own productions. The Malian CGE model is essentially adapted
flom a standard neoclassical structuralist model developed by researchers at the
International Food Policy Research Institute (IFPRI). Applying the standard CGE model
the Malian data disaggregated social accounting generates an infeasible solution due to
the presence of extremely small values in the matrix. Small entries pose serious
convergence issues in nonlinear optimization models. In order to eliminate small entries,
it was necessary to aggregate the 37 commodities into 20 commodity groups. The
changes in the re-specified model have inevitably reduced the level of details initially
contemplated. Nonetheless, the resulting analysis is still rich in details, allowmg sectoral
and household-level assessments of the impacts of global, bilateral and domestic trade
policy reforms on welfare and food security in Mali.

The results indicate that under most of the scenarios, Mali will benefit flom
welfare gains and increased food consumption. As a country, Mali is likely to gain flom
expanded global (or North-South) trade, whether the expansion comes flom partial
(F APRI and OECD) or full (IFPRI-2) liberalization of world commodity markets.
Welfare would also increase if there were no further liberalization of world agricultural
trade but productivity in the rest of the world continues to grow (IFPRI-1), or if the
country secures an effective duty-flee access for its exports to the world markets (DFA).
The magnitude of welfare increases would vary flom 1% (IFPRI-2) to 2.9% (OECD),
and these differences are due to variations in the projected prices. Overall welfare would
also increase under the three government policy scenarios (common external tariff,

cereal export ban, and increased public investments in key sectors). The country’s

154

 

welfare would decrease, however, if current trade preferences were eroded under the
EEP trade regime.

The overall country effects hide important cross-household differences.
Household effects are primarily determined by changes in domestic prices, which in
turn affected factor incomes through the Stolper-Samuelson linkages. In general,
reduced domestic prices would increase real urban incomes and decrease real incomes
in rural areas. Reduced domestic prices would occur under partial (F APRI, OECD) and
firll (IF PRI-2) liberalization of the world commodity markets. As net commodity
purchasers, the two urban households (Bamako and Other Urban) as well as the Other
Rural households (which has urban consumption pattern) would benefit flom welfare
increase whenever prices fell. Being net sellers of agricultural commodities, the
remaining two rural households (Rural Kayes-Sikasso and River) would suffer more
flom the depressed real incomes associated with falls in domestic prices.

The status quo price change scenario (IFPRI-1) would have a mixed effect on
domestic prices. Cereal prices would fall while livestock prices (only for “other
livestock products”) would increase, and the resulting effects would favor rural
households, particularly those in the Kayes-Sikasso Region. Thus, the continuation of
current world commercial policies would most likely benefit rural dwellers, with
possibly net welfare losses in urban centers and in the Other Rural households.

In addition to changes in domestic prices, changes in factor incomes would also
contribute to the observed changes in household welfare. The declines in domestic
prices would generally be associated with higher returns to wage labor, which is has

been specified as a mobile factor in the Malian CGE model. Reduced agricultural

155

productivity stemming flom reduced commodity prices would force a reallocation of
wage labor flom less productive to more productive sectors, resulting in higher returns
to this production factor. Since urban household own the quasi-totality of the wage labor
factor, they would gain in welfare when the returns to wage labor increase in all but the
IFPRI-1 and EEP scenarios. In the latter two scenarios, wage income would fall
resulting in welfare losses in urban areas.

Changes in the real exchange rate also appeared to be associated with the
observed welfare changes. In general, urban welfare would increase when the exchange
rate appreciates, and fall under a depreciation of the real exchange rates. Conversely,
rural household would generally benefit flom the depreciation (IFPRI-1). But an
exception occurred under the EEP scenario in which real exchange rate depreciation
could lead to reduced imports of intermediate inputs, and ultimately to generalized
losses in welfare across all households. Rural households would lose flom a currency
appreciation, which could reduce exports and thereby producers’ incomes.

Welfare results under the three government policies (implementing CET regime,
banning cereal exports, and increasing public investments in key sectors) have the same
characteristics as the results under the other scenarios. As before, these policies affect
welfare through three interdependent effects on domestic prices, factor incomes, and
real exchange rates. Overall welfare would increase under the CET and increased public
investments scenarios, while remaining substantially unchanged when cereal exports are
banned. Cereal export bans would raise domestic supply and reduce prices, which would
have negative welfare impact in rural households and a positive impact in urban zones.

The CET regime would reduce average tariffs, and thereby the domestic prices. The

156

 

increased public investments would also result in overall welfare gains in the country,
with most of the gains concentrated in the Other Urban and Kayes-Sikasso households.
In terms of welfare, this scenario will deliver a Pareto-compatible outcome in which no
household would substantially lose flom the positive investment shocks. These
investments would help both producers and consumers through lower intermediate and
final demand prices. Underlying the investment result is the assumption that the
distribution of aggregate public investments mirrors the pattern in the base year, and is
relatively well spread out across households. In practice, however, public investments
are made at the margin, and they tend to be sector and location-specific. Thus, the
expected Pareto-compatible effects of increased public investments will most likely
weaken, if the increases were to be allocated at the margin, instead of the current across-
the-board allocation.

In most cases, food security effects will follow closely the welfare impacts. Food
consumption would unambiguously increase under expanded liberalization of world
commodity markets (particularly the FAPRI scenario). Conversely, erosion in tariff
preferences would unambiguously reduce food consumption across all households in
Mali. The ratio of current food imports to total exports revenues did not present any
systematic relationship with the level of food consumption in Mali.

All these results are obtained under the assumption Mali would not benefit flom
the worldwide productivity gain, which would drive reduction in world commodity
prices. An alternative scenario is to investigate the case when also benefit flom
productivity gains, concurrently with the rest of the world. This scenario was

investigated and the results indicate that increased productivity in Mali would reduce, or

157

even reverse, losses incurred in rural areas. This is mainly because a gain in productivity
in Mali would increase the competitiveness of domestic production, limiting the
negative effects on increased imports on domestic production.

An important conclusion flom the results is that the South-South liberalization
(the CET regime) would result in 3% overall welfare increase, and a similar magnitude
of gains is expected under the North-South partial liberalization scenarios (FAPRI and
OECD). So, flom the perspective of the country as a whole, there is as much gain to be
expected flom effective implementation of the CET regime in West Aflica, as there are
gains flom increased liberalization in the rest of the world. This confirms the common
belief that poor countries would benefit flom own trade liberalization. However, most of
the expected gains would go to urban households, who would benefit flom important
falls in final demand prices, real exchange rate appreciation and increased returns to
their endowment.

This unequal distribution of the potential gains (which would favor urban
dwellers) could encourage lobbying for and against the maintenance and an effective
implementation of trade reforms. As indicated in the quote at the beginning of this
chapter (Lee, 2002), it is these sorts of conflictual interests that empower politics to
overpower economics and economists, when it comes to the making of agricultural
commodity policies.

7.5. Achievements and Limitations of the Dissertation

One of the limitations of this study is that some of the quantitative results appear

to be less stable than standard results in computable general equilibrium analysis. The

Malian CGE model was moderately sensitive to parameter values and to some

158

macroeconomic closures regarding the foreign market and the savings-investment
accounts. This moderate sensitivity of the CGE results may be due to the relatively
limited quality of the Malian social accounting matrix. Despite this sensitivity of
quantitative estimates, qualitative results can be viewed as robust, particularly in terms
of how the fundamentals (movements in commodity prices, in factor incomes, and in the
real exchange rates) would affect welfare and food security in Mali. Because the
magnitude of the simulated effects is less stable, caution is needed when using the
quantitative results for policy applications.

Despite the various limitations outlined throughout the study, we can say that the
key objectives of the study have been met in their essence. Yet, this study may be
improved in many respects. First, we indicated earlier that there is a tremendous room
for improving the quality of data used in this study, once recent micro and
macroeconomic survey results are made available. The quality of the social accounting
matrix is crucial to the quality of the CGE results, and it important that a representative
matrix for the country be put together for Mali. This is best done by a team of sectoral
experts using the best available information flom national and multilateral institutions.
We have also identified serious imbalances in the supply and demand sides of the
economy. Creating a balanced social accounting matrix by a team of experts would be
one of the best ways to reconcile production and consumption data in Mali. It could also
permit more specific distinction of household groups and production factors in the
economy. The SAM could finally be used for a wide range of policy analyses, and

would be great addition to the economic analysis toolkit for the country.

159

 

Second, there is also a room to modify the standard analytical model to reflect
more closely a more empirically-based specification of the production technology and
consumer preferences in Mali. The current generic LES demand and CES/Leontief
technologies could eventually be replaced with evidence-based specification of
consumer preferences and production technologies that are more representative of the
Malian economy. An example of such evidence is Camara’s (2004) characterization of
consumer demand in Bamako within the Almost Ideal Demand System (AIDS)
flarnework.

Third, the Malian model is static, but results suggest that there could be several
adjustment and readj ustrnent processes at work between the base year and the
counterfactual general equilibria of the economy. For example, changes in investment
demand has both contemporaneous and future effects, but the lack of dynamics in a
standard CGE model would fail to capture future effects if investments are
endogenously determined. A dynamic model would, therefore, be useful for a richer
understanding of the adjustment path the economy will follow in response to policy
shocks.

Fourth, it is clear that CGE models represent a useful flarnework to understand
the impacts of alternative trade and government policies. Yet, smaller models that trace
the impacts of specific policy changes at sector or even at commodity (filie‘re) levels are
necessary to complement CGE-based results. Diagana et al. (1999) is an example of
such an alternative analysis, applied to the impacts of the devaluation of the CFA flanc

on the agricultural sector in West Aflica.

160

F ifih, the analysis is limited to the inter-household effects of policy changes.
There are ample evidence, however, that policy changes has both inter- and intra-
household effects. There is an emerging literature on micro-simulations that permit
analyses of both effects within general equilibrium models. Such analyses are based on
actual households, as opposed to standard analyses that rely on representative
households. The expenditures and incomes of real households are directly obtained flom
household surveys, and the social accounting matrix in micro-simulation analyses
contains as many households as there are in the surveys. Micro-simulation analyses are
possible if there are reliable income and expenditure data on each real household.
Currently in Mali, only the 1994 survey meets this criterion. Unfortunately, the survey is
of limited value for current policy analyses because several structural changes took
place in Mali since the survey was conducted immediately after the 1994 devaluation of
the CF A flanc. A combined income-expenditure survey would be required for reliable
micro-simulation analyses in Mali.

Finally, a general message flom this study is that in a world where agricultural
productivity is growing faster than in Mali, urban consumers benefit flom fleer trade
regimes that allow them to access that increased productivity. Rural farmers with
immobile resources, producing import substitutes, generally lose. The losses are
mitigated, or even reversed, if the Malian agricultural sector becomes equally
productive as in the rest of the world, because domestic productions are able to compete

more effectively against imports.

161

APPENDICES

Appendix 4.1: Mathematical Model: Sets, Parameters, Variables and Equations"l

SETS
a e A
a e ACES(C A)

ceC
ceCD(c:C)

c e CDN(C C)

c e CE(C C)

c e CEN(C C)

c e CM(C C)

c e CMN(c: C)

c e CT(c C)

c e CX(c C)

f e F

is INS

ie INSD(<: INS)
is INSDNG(C INSD)
h e H(c INSDNG)

PARAMETERS
Latin letters
cwtsC

dwtsc

icacfl

icdaco
icec’co

icmaco
Intaa

tvaa

 

mpsi
mpsOlc
Pwec
Pwmc
qutc

qgc

Activities

Activities with CES function at the top of the technology nest
Cormnodities

Commodities with domestic sales of domestic output
Commodities not in CD

Exported commodities

Connnodities not in CB

Imported commodities

Commodities not in CM

Transactions service commodities

Connnodities with domestic production

Factors

Institution (domestic and rest of the world)

Domestic institutions

Domestic non-govemment institutions

Households

Weight of commodity c in the CPI
Weight of commodity c in the production price index
Quantity of c as intermediary input per unit of activity 0

Quantity of corrmiodity c as trade input per unit of c’ produced and sold
domestically
Quantity of conunodity c as trade input per exported unit of c’

Quantity of commodity c as trade input per inrported unit of c’

Quantity of aggregate intermediate input per activity unit

Quantity of value-added per activity unit

Base saving rate for domestic institution i

0-1 parameter with 1 for institutions with potentially flexed direct tax rates
Export price (foreign currency)

Irrrport price (foreign cmrency)

Quantity of stock change

Base-year quantity of government demand

 

“ This Appendix substantially uses materials flom Lofgreen, Harris and Robinson (2002).

162

 

qinvc
Skiff, f
Shifty

Greek letters

Base-year quantity of private investment demand

Share of the domestic institution i in income of factor f

Share of net income of i’ to i (i'e INSDNG';i e HVSDNG)
Tax rate for activity a

Export tax rate

Direct tax rate for factor f

Exogenous direct tax rate for domestic institution i

0-1 parameter with 1 for institutions with potentially flexed direct tax rates
West Aflican import tariff rate 5
Rate of sales tax

Transfer flom factor f to institution i

Rate of value-added tax for activity a

Efficiency parameter in the CES activity firnction

Efiiciency parameter in the CES value-added function

Shift parameter for the domestic commodity aggregation function
CES function shift parameter (imports and domestic sales)

CET shifi function parameter (exports and domestic sales)

Marginal share of consumption spending on home commodity c flom
activity a for household h

Marginal share of consumption spending on rmrketed commodity c for
household h

CES activity share parameter

Share parameter of domestic commodity aggregation function
CES function share parameter (imports and domestic sales)
CET function share parameter (exports and domestic sales)

CES value-added function share parameter for factor f in activity a
Subsistence consumption of marketed commodity c for household 11

Subsistence consumption of home commodity c flom activity a for
household 1:

Yield of output c per unit of activity a
CBS production function exponent

CES value-added function exponent

163

EXOGENOUS VARIABLES

 

Consumer price index

CPI
DTINS Change in domestic institution tax share (= 0 for base; exogenous variable)
FS A V Foreign saving (foreign currency)
G A DJ Government consumption adjustment factor
IA DJ lnvestrnent adjustment factor
MP SA DJ Saving rate scaling factor (= 0 for base)
Q FS Quantity supplied of factor
f
TINS A D J Direct tax scaling factor (= 0 for base; exogenous variable)
— Wage distortion factor for factor f in activity a
WDFDISTM
ENDOGENOUS VARIABLES
DWS Change in domestic institution saving rates (= 0 for base; exogenous
variable)
DP] Producer price index for domestically marketed output
EG Government expenditures
EH h Consumption spending for household
EYR Exchange rate (local currency per unit of foreign currency)
GO VSHR Government consumption share in nominal absorption
GSA V Government savings
[NVSHR Investment share in nominal absorption
WS. Marginal propensity to save for domestic non-govemment institution
' (exogenous variable)
PAa Activity price (unit gross revenue)
PDDC Demand price for commodity produced and sold domestically
P DSC Supply price for commodity produced and sold domestically
PEc Export price (domestic currency)
P [NT/10 Aggregate intermediary input price for activity a
P M c Import price (domestic currency)
PQC Composite commodity price
PVAa Value-added price (factor income per unit of activity)
P X c Aggregate producer price for a commodity
PMC Producer price for commodity c for activity a

a,c

(2Aa
QD.
Q5.
Q1“)...
QG.
QHA

a,c,h

Quantity (level) of activity

Quantity sold domestically of output

Quantity of exports

Quantity demanded of factor f flom activity a
Government consumption demand for commodity

Quantity of household home consumption of commodity c flom activity a
for household h

164

QEc
QF f ,a
QGc
QHAa,c,h
QINTAa
QINTM
QINVc
QM c
QQc

QT c
QVAa
QX c
QXA Ca c

TABS
TINS,

my“.

YFf
YG
Y1,-
1’15”;

Quantity of exports
Quantity demanded of factor f flom activity a
Government consumption demand for commodity

Quantity of household home consurrrption of commodity c flom activity a
for household h

Quantity of aggregate intermediary input

Quantity of cormnodity c as intermediary input to activity 0
Quantity of investment demand for commodity

Quantity of imports

Quantity of goods supplied to domestic market (composite supply)
Quantity of commodity demanded as trade input

Quantity of (aggregate) value-added

Aggregate marketed quantity of domestic output of cormnodity
Quantity of market output of commodity c flom activity a

Total nominal absorption
Direct tax rate for institution i (i E INSDN G)
Transfer flom institution i’ to i (both in the set INSDG)

Average factor price f
Income of factor f

Government revenue
Income of domestic non-govemment institution

Income to domestic institution 1' flom factor f

165

EQUATIONS
1. Price Block
Import price:

Export price:

Demand price of
nontraded goods:

Absorption

Market output
value
Activity price

Aggregate
intermediate input
price

Activity revenue
and cost

Consumer price
index

Producer price
index for
nontraded market
output

PM, = pwm,.(1+tm,).m + ZPQ,.jcm,.,

c'eCT
PE, = pwe,.(l - te, )FH - Z PQcticecv,
c'eCT
PDD, = PDS, + Z PQ,.rcd,.,

c'eCT
PQ,.(1- tq, ).QQ, = PDD,.QD, + PM, .QM,

PX, .QX, = PDS, .QD, + PE, .QE,

PA, = ZPXAC,, 19,,

ceC
PINTA, = Z PQ.ica ,,

ceC

PA,.(1-za,).QA, = PVA,.QVA,
+ PINTA, .QINTA,
-C_P_I = Z PQ,.cwts,

ceC

DPI = 2 PBS, .dwts,
ceC

2. Production and Trade Blocks

CES technology:
Activity prod.
Function

CES tech.: value-
added
intermediate
input quantity
ratio
Value-added and
factor demands

Factor demand

0 a
QA, = ag{5g.QVA;/’a + (1 —5g).Q1NTA;Pa ) pg

1

 

 

QVAa = QINTAa 5: H63
QINTAa PVAa 1-5:
I
_ va pva
QVAa =aza{ z 5;:QF up“ ] a
feF

Wffl’FDIST fa = PVAa.(l - tvaa).

l

—1
_ va _ va_
2 egg, Pa ] 53913., p. 1
f 61" '

166

1

ceCM
CECE

ceCD

C E
(CDOCM)

ceCX

06A

06A

(1 e ACES

a e ACES

aeA

aeA
feF

(4.1)

(4.2)

(4.3)

(4.4)

(4.5)

(4.6)

(4.7)

(4.8)

(4.9)

(4.10)

(4.11)

(4.12)

(4.13)

(4.14)

Leontief tech.:
demand for
aggregate value-
added
Leontief tech.:
demand for
aggregate
intermediate
input
Disaggregated
intermediate
input demand
Commodity
production and
allocation
Output
aggregation
function

First-order
condition for
output
aggregation
function

Output
transformation
(CET) function

Export-domestic
supply ratio

Output
transformation
for non exported
commodities

Composite supply
(Armington)
function

Import-domestic
demand ratio

Demand for
transaction
services

QVAa = ivaa .QAa

QINTAa = intaaQAa

QINTca = icaca .QINTAa

QXACa, + ZQHAach .—_ gaC°QAa
hell

1
ac_1

_ ac '-
QXC =ag0[25g§.QXAC,,pC ] pc

 

06A

-1
_pac
PXA C,, = PX, .QX,[ 263,0 .QXAC,, “ J .

as A’

-ch -1

6:: .QXACa,
l

t t —
QXc = a; {65.QE5’C + (l — 5:).QDg’c )Pé
1

QE,_ PE, 1-5; 55-1
QDc PDSc'l-ag

QXC = QDc + QEc

 

1

_ q _ q ‘—
QQc =ag {53°QMcpc +(1_ch)°QDC Pc ] pg

1

 

 

 

QM, _ P01), 5,1 1+5:
QDc P M c l - 63
QT c = Z icmcccQMco + icecchEcv + icdcchDcv

c'eC'

167

a e ALEO (MS)
a e ALEO (4.16)
a e A (4.17)
c e C
a e A (4.18)
c e CX
c e CX (4.19)
a e A (420)
c e CX
c e (4.21)
(CE n CD)
0 6 (4.22)
(CE 0 CD)
6 e (4.23)
(CDnCEN)
u
(CE u CDN)
C 6 (4.24)
(CM nCD)
c 6 (4.25)
(CM n CD)
c 6 CT (427)

3. Institution Block

Factor income

Institutional
factor income

Income of
domestic non-
government
institutions

Inna-institutional
transfers
Household
consumption
expenditure

Household
consumption
demand and
marketed
commodity

I-IH consumption
demand for home
consumption

Investment
demand
Government
consumption
demand
Government
revenue

Government
expenditure

 

YFf + wa.WDISTfa.QFf, f e F
aeA
i e INSD
mg, -_- shif,7((1— tff).YFf — tmsf,,wf.m) f e F
Y1,- = 21717,, 1'6
f 6F INSDNG
+ 2 m1,» + tmsfigov.a’7 + trnsfl-mwm
i'eINSDNG'
TRIIii' = Shiftina — MPSi').(1— TWISi').YIi' i6 INSDNG
i'e INSDNG'
h e H
EH), = 1- 2mm), .(1— MPSh).(1- YNISh).YI,,
ieINSDNG
PQc-QHch =PQc-72h'l' Z11" CEC
h e H
h
[EHh - ZPQc'WZ'ih - Z ZPXACa,C'°7a,c',h]
c'eC aeA c'eC
PXACa,,.QI-1Aa,,,h = PXACa,,.7Z,C,h + j’c’h. “ E A
c e C
[E11, - 2&5..th - z ZPMCa,,e.)/Z,c.’h] h 6 H
c'eC aeAc'eC
QINV, = IADJ.qinv, C 6 CIN V
QG, = GADJEIE,’ c e C

Y0 = Q ZTINSI'YII
ieINSDNG
+ Z tff.YFf + XmaPt/A,.QVA,
feF USA
+ Z taa.PAa.QAa + ztmcpwchMc EXR
aeA ceCM
+ Zte,.pwec.QEc.EXR+ thcPQc-QQC
ceCE 06C
+ Z YIFgovJ + trnsfrgov’mwm
f 6F
E6 = ZPQc-QGC + Z‘msf’ieowcpl
ceC ielNSDNG

168

(4.28)

(4.29)

(4.30)

(4.31)

(4.32)

(4.33)

(4.34)

(4.35)

(4.36)

(4.37)

(4.38)

4. System Constraint Block

Factor market

Composite
cormnodity
market

Current account

balance for rest of

the world (in

ZQFfm = 9178]. f e F

aeA

QQc = ZQ’NTM + ZQHCA +QGc +QINVc +94% +QYc
aeA he”

2 pwchMc + Ztrnsfrmw’f
ce CM f EF

ceC

fOMS“ WW) = Z pwec.QEc + Ztmsjfiy’mw + FSA V

Government
balance

Direct
institutional tax
rate
Institutional
saving rate

Saving-
investment
balance

Total absorption

Ratio of
investment to
absorption
Ratio of
government
consumption to
absorption

ceCE
Y0 = BC + GSAV

i eINSD

TINS, = Tia-(1 + T1NSADJ1insOl,-) + DTINSJinsOli) 1' e INDSNG
MPSi = mpsi.(1 + MPSADJ .mps01i) + DMPSmpsOli) i E UVDSNG

ZMPS,.(1— HNS,).Y1,- + GSA V + mPSA V =
i eINSDNG

ZtPQ..QINV. + 2199.442.)
ceC ceC

TABS = Z ZPQ,.QH,,, +

heHceC

Z Z ZPMCam-QHAamh

aeAheHceC

+ Z PQ,.QG, + ZPQ,.QINV, + ZPQ,.qut,
ceC ceC ceC

INVSHRJABS = Z PQ, .QINV, + Z PQ, .quz,
ceC ceC

GOVSHRJ’ABS = ZPQ,.QG,
ceC

169

(4.39)

(4.40)

(4.41)

(4.42)

(4.43)

(4.44)

(4.45)

(4.46)

(4.47)

(4.48)

Appendix 5.1: Justification of Entries in the Malian Macrosam

There are 34 nonzero entries in the Malian macrosam, as presented in Table 5.2. This
Appendix documents the sources and the rationale of each of these entries. Throughout
this section, a cell is referred as [Row, Column] where the first element refers to the
row, and the second to the coliunn. All figures are expressed in million CFAF flancs
(CFAF).

Both the 729,045 million CFAF attributed to intermediate inputs [Commodities,
Activities] and the 1,310,343 million of value added at factor cost [Factors, Activities]
were taken directly flom the national accounts published by DNSI. The sum of these
two values is the gross output, evaluated to be 2,039,388 million.

Along the row, activities received income flom marketed output [Activity,
Commodities] and flom home-consumed output [Activities, Households]. The input-
output (I-O) table indicated that households’ home consumption [Activities,
Households] in 1997 was 320,311 million, whereas the private final consumption
[Commodities, Households] was 663,356 million, and the final consumption of the
government [Commodities, Government] was 221,563 million. The value of the
marketed output [Activities, Commodities] was 1,719,077 million, the difference
between the gross output and the home-consumed output.

Regarding the marketing margins [Commodities, Commodities], the auxiliary
DNSI accounts suggest that they represented about 12% of the total private
consumption, which would be 118,040 million. For the macrosam, we have retained a
more precise estimate of the aggregate marketing margins, evaluated to 126,500 million
based on detailed estimates discussed later in Appendix 5.2.

The Malian total exports [Commodities, ROW] in 1997 were valued to 361,610
million CFAF, while imports [ROW, Commodities] run at 486,200 in the same year.
These figures were taken directly flom DNSI national accounts, which also valued
the1997 changes in stocks [Commodities, Stock Change] to be -22,789. It is
straightforward to verify that this figure is the sum of the variation in the stocks of
cotton (-28,514) and livestock (5,755). A careful analysis of the I-0 table revealed,
however, changes in the stocks of other commodities in 1997. These include gold (-30),
food crops (8,426), other manufacturing goods (5,533), water, electricity and energy
(3,200) and other trading services (16,800). Changes in stocks for traded services refer
to equipments specific to the delivery of these services. It is unclear why these latter
variations in stocks were not included in the DNSI official figures on stock variation,
which was set to -22,789 million. Given that the 1-0 table was one of our main data
source used in building the SAM, we have treated the official figures as inaccurate. We
have, therefore, used all the details provided in the 1-0 table, and this changed the total
variations in stocks to 11,170 million CFAF.

Commodities also received payments flom private and public investments.
These payments were entered in cells [Commodities, Capital Investment] and

170

[Commodities, Government Investment], respectively. The main DNSI accounts only
provided the value of the total investment, known as the Formation Brute Capital Fixe
(FBCF). This value was 348,093 million CFAF. This figure is very close to the 347,300
million allocated to FBCF in the BCEAO database. BCEAO figures also indicated that
the public sector accounts for 31.9% of the total investment, while private investment
makes up the difference. Thus, we used the ratio 31.9:68.1 to distribute the total
investment to the public and private sectors, which receive 1 10,893 and 237,200,
respectively.

The Tableau Economique Integré (TEI), which is an unpublished source flom
DNSI, indicates that factor return to households made about 80% of value added. This
was equivalent to 1,048,874 million CF AF , entered in cell [Households, Factors]. The
total factor income and tax to government amounted to 88,555 million according to TEI
figures (factor taxes alone were 10,305 million). This amount was adjusted by
government service of foreign debt, evaluated to 13,000 million according the IMF
sources. Thus, the entry to [Government recurrent, factor] was 75,555 million CFAF.
Total factor payments to the rest of the world in cell [Rest of the world, Factors] were
43,687 million, based on the TEI figures. The last entry in the factor column, that is the
factor return to enterprises, was obtained as a residual so that the total payments to
factors (domestic and foreign) balanced with factor income to domestic and foreign
institutions. The residual was entered in cell [Enterprises, Factors] and amounted to
153,924 million CFAF. The 1997 TEI also indicates that factor payments flom abroad
were 11,097 million [Factors, Rest of the world].

The entry for indirect fiscal revenue to the government recurrent account in cell
[Government Recurrent, Commodities] was 149,700. Direct fiscal receipts, or
institutional payments to the government recurrent account, came flom households and
enterprises. These payments amounted to 46,600 in 1997, according to BCEAO records.
Households paid about 9% of this amount (4,208), whereas enterprises paid the rest
(42,392). The ratio 9:91 came flom the preliminary TEI, which indicated that
enterprises paid 35,000 (91%) out of the total 38,474 that governments received as
direct fiscal income. In the macrosam, one should note that the entry of households’
payments to the government included not only taxes, but also social security payments.
The TEI evaluated these latter payments to 24,642 in 1997. Thus, the entry to
[Government Recurrent, Household] was 28,850 (4,208 + 24,642) and this for
[Government Recurrent, Enterprises] was, as indicated earlier, 42,392.

Regarding savings, the preliminary TEI showed a total saving of 299,263 million
CF AF distributed between households (19%), enterprises (42%) and government (39%).
According to the DNSI national accounts, total savings in 1997 were 314,552. We
distributed this latter figure to households, enterprises and government, using the ratio
0.19:0.42:0.39, respectively. Thus, saving entries were 130,619 for [Capital Investment,
Enterprises], 60,467 for [Capital Investment, Households] and 123,466 for [Capital
Investment, Government Recurrent]. Foreign savings in cell [Capital Investment, ROW]
were 24,634, taken directly flom the DNSI national accounts.

171

With regard to transfers between Mali and the rest of the world, the TEI sources
indicated that the country received total current transfers of 158,091 million flom
abroad. This was shared, mainly between the government recurrent account which
received 99,306 in [Government recurrent, ROW], and households, which received
58,785 in [Households, ROW]. Transfers flom ROW to enterprises were 5,509 million,
also taken flom the 1997 TEI. On the other hand, the 1997 national accounts revealed
that current transfers to the ROW amounted to 51,351 million, of which 12,024 may be
attributed to households [ROW, Households], according the 1997 TEI. The difference,
39,327 million, was paid by the government recurrent account [ROW, Government
recurrent]. Also, the national accounts indicated that enterprises transferred 13,800
million to ROW, as indicated in cell [ROW, Enterprises]. One may notice that all ROW
transfers to governments are paid to the recurrent account, though some of these
transfers may be for investment purposes. We have adopted the standard treatment of
these transfers, which are first cashed in the government current account, which transfer
them to the capital investment account, which in turn allocates them to the government
investment account.

There are two remaining entries that are unexplained entries to the macrosam.
First, the 58 million CFAF to [Households, Enterprises] represent enterprises’ dividend
payment to households. This figure was taken directly flom the preliminary TEI.
Second, households also received a transfer of 28,489 million flom government
recurrent account in cell [Households, Government Recurrent]. The 1997 preliminary
T EI provided these two figures.

All the 34 nonzero entries in the macrosam will be disaggregated in order to fill
in the relevant cells of the microsam , which is discussed below in Appendix 5.2.

172

Appendix 5.2: Justification of Entries to the Malian Microsam

The microsam has 83 accounts, as mentioned earlier. Entries to the nricrosam were
obtained by disaggragating the 34 nonzero entries of the macrosam. This section
discusses the disaggreation procedure and the main sources of the data needed to carry
out this procedure.

Intermediate inputs

Intermediate inputs were differentiated by their domestic and import origins in the base
I-O table. The structure of the microsam does not allow for such differentiation, so these
intermediate consumptions were aggregated.

One of the main characteristics (and perhaps limitations) to the entries of
intermediate input uses relate to the distribution of a single intermediate consumption
value to many commodities, in situations where one activity produces several
commodities. In most cases, it was not possible to access information on the share of
each commodity in the total consumption of intermediate inputs by activities. We used
activity volume or value to prorate the use of these inputs, guided by common sense or
additional insights into the production processes. For example, the activity “food
production, excluding rice” (AF OOD) consumes 4,525 million CF AF worth of food
products as intermediate inputs. This intermediate consumption is primarily made of
seeds, which were evaluated to be between 2% (maize and roots) and 3% (millet,
sorghum, fonio, and beans) of the value of the outputs. “Other crops” intermediate
consumption is determined as residual. These rates come flom unpublished DNSI
sources.

Other activities, such as “agro-industry and beverages” (AAIDT), “trade”
(ATRAD) and “nontraded services” (ANT SE) also used food products as intermediary
inputs. Most of the agro-industry activities are informal food processing at home, such
as home brewery. The 1997 agricultural survey evaluated the rate of home food
processing to be 1.37% for millet, 1.13% for sorghum, 0.36% for maize, 0.13% for
fonio and 1.25% for beans. Each rate was applied to the output values of the
corresponding commodity. The resulting total (1,349) was adjusted to reflect the initial
entry of 352, using the proportionality rule.

The activity “trade” (ATRAD) also used food products (flom activity AF OOD)
as intermediate inputs. We used the commercialization rate of these products to
distribute the 122 million CFAF of intermediate input consumption (flom the I-0 table)
to the seven food commodities. Similarly, the production of nontraded services, mainly
by public administration and nonprofit organizations, used food crops as intermediate
inputs. Again, we assumed that uses are proportional to the value of the activities, thus
distributing the 1,532 million of that consumption to the seven food products.

Products flom the industrial cropping activity (groundnuts, tobacco and wheat), and the
use of these products as intermediate inputs are disaggregated in the microsam
following the same assumptions used for the food cropping activity. For example, seeds
represented 2% of groundnuts, 3% of tobacco and 3% of wheat. Home processing rate

173

were 3.3% on groundnuts and 27.1% for tobacco, while commercialization rates were
20.2% and 25.6%, respectively for groundnuts and tobacco. Most processing activities,
however, are based on oil making (groundnut, so we allocate most to groundnut).
Regarding the processing of animal products by the agro—industrial activity (AAIDT),
payments to commodities were limited three commodities (hides and skins, cattle, and
sheep and goats), again using the proportionality rule. The activity “production of other
manufacturing goods” (AOMGG) used four commodities: firewood (CFWOO),
furniture wood (CWOOD), gathering products (CGATH) and gold (CPGLD),
proportional to their values, whereas the activity “buildings and public works” (ABBTP)
used only furniture wood (CWOOD).

Production factors

The Malian rrricrosam has nine production factors. The factors were identified mainly
based on the results of the EMCES 94, which described households in terms of activity
intensity in each factor category. The EMCES 94 was carried out between March and
July 1994, and the published results were based on data flom 9,800 households, sampled
across the entire country. The survey, though comprehensive in many regards, had
nevertheless a limited focus in terms of cropping activities, if compared to the yearly
permanent agricultural survey, the Enquéte Agricole Permanente (EAP). The EMCES
1994 covered specifically the main traditional cereals (sorghum and millet) and more
input-intensive cereals (maize and rice), as well as export and agro-industrial crops
(cotton and groundnuts). The survey put less emphasis on horticultural crops, such as
fluits and vegetables, though these crops represent strategic activities of women and
urban farmers. Despite these limitations, experts at DNSI argue that the main crops
surveyed are representative of the agricultural activities undertaken by a “large majority
of rural households” (Toure, 1995; p. 7).

We have distinguished three location-based farming factors (Kayes-Sikasso,
River, and Other rural), and Table E.l (Table 5.9 in the main text) provides further
details on the composition of these regions. This table is based on Charmes (1994).
Table E.1 indicates, for example, that the Kayes-Sikasso composite farming factor
includes farming factors flom Sikasso Region (excluding the Dioila Circle) and flom
Kayes Region (excluding Noiro Circle), on so on. Touré (1995; p. 17) shows the shares
of households in each socioeconomic group that are involved in the main agricultural
activities. The sum of these should normally sum up to 100, but in many cases
(sorghum, maize and rice), it was slightly less than 100, and we corrected the
inconsistency through proportional readjustment of these shares.

Regarding nonagricultural activities, their importance was determined based on
their relative contributions to household incomes. A nonagricultural enterprise was
defined as a “production unit having as main function the production of goods and
services in sectors other than agriculture, livestock keeping, fishery and forestry”
(Toure, 1995; p. 8).

Payments of the “other manufacturing goods” activity (AOMGG) to all factors,
except capital, were obtained flom the cost share of these factors in the total value added

174

of the activities. Similarly, the return to food and industrial cropping activities was
shared between five “composite” factors: Kayes-Sikasso Farming, River Farming, Other
Rural Farming, Rural Composite and Urban Composite. The returns were computed as
weighted averages, using cost shares as weights. The EMCES 94 supplied these cost
shares, which we discuss later in this Appendix.

Table E.l: Description of locations” used to define composite production factors
Locations Composition

 

 

Kayes-Sikasso Sikasso Region (plus Dioila Circle)
Kayes Region (minus Nioro Circle)
River Koulikoro Region (minus Dioila and Nara Circles)

Segou Region (minus Niono Circle)
Mopti Region minus Bandiaggra, Koro, Bankass, Douentza Circles)
Other Tombouctou Region
Gao Region
(plus Nioro, Niono, Bandiagara, Koro, Bankass, Douentza Circles)

Source: Based on Charmes (1994; p. 19)

Payments to capital were assumed to correspond to activities for which the 1-0
table indicated that investments have been made. Three activities received almost the
totality of investment expenditures in 1997. These include “building and public work”
(ABBTP) with 61.7%, “other manufacturing goods” (AOMGG) with 25.7%, and
“livestock and fishery” (ALIFI) with 8.6%. The activities “trade” (ATRAD) with 2.3%,
“transport and telecommunication” (ATTEL) with 0.2% and “non-traded services”
(ANT SE) with 1.6%, made up the rest of investments. Thus, return to capital is
restricted to these six activities. Assuming factor price equalization across sectors, the
return to capital was considered to be proportional to the value of investments. Returns
to two activities, “building and public work” (ABBTP) and “other manufactiuing
goods” (AOMGG) which also take place in the informal sector, are adjusted by the
share of modern sector in these activities. Making the adjustment resulted in a more
even distribution of the returns to formal capital across the six activities, as shown in
Table E.2 below.

Table E.2: Initial and Adjusted Shares of Activities in Total Investment
Share of total investment Adjusted shares based on

 

 

Activities in the 1997 1-0 table the proportion of modern
(in percent) activities (in percent)

ALIFI 8.6 18.7

AOMGG 25.7 27.4

ABBTP 61.7 45.1

ATRAD 2.3 5.0

ATTEL 0.2 0.4

ABTSE 1.6 3.4

Total 100.0 100.0

 

Source: Based on the 1997 1-0 table and other data flom DNSI.

The value added of the “livestock and fishery” (ALIF I) activity is distributed to
capital and the location-based composite factors. Payments to composite factors are

175

proportional to the intensity of livestock activities in the five agricultural composite
“factor” categories.

The value-added flom activity “forestry and gathering” was distributed between
the three main farming factors (“Kayes-Sikasso”, “River” and “Other Farming”),
proportional to the intensity of the forestry and gathering activity in these localities. The
intensity is based on the survey of rural labor presented in Doumbia and Kamate (1997;
p.8), in which the figures were presented by administrative regions. We used Charmes’
(1994; p. 19) mapping of administrative regions into the location-based composite
factors (Table E. 1). It was not possible to obtain figures that were disaggregated to the
level of administrative circles, implying that our distribution was solely based on
aggregate regional figures. As a consequence, it is possible that payments of the activity
“forestry and gathering location” to the composite factor “River” (include three
administrative regions minus seven circles) was overestimated, whereas the payments to
“Other Farming” composite factor (which include two regions plus six circles) was
underestimated.

In order to distribute the value-added of the agro-industrial activity (AAIDT), we
used the fact that the modern sector represented about 94% of value added of that
activity, according to DNSI auxiliary figures. We assumed that this figure can be
considered as the payment for wage labor. Thus, the remaining 6% was distributed to
the composite factors, proportional to the manufacturing intensities in each of the five
locations. A similar procedure was use to distribute the value added of the “textile”
activity to the wage labor factor (67%, corresponding to the share of the modern sector
according to DNSI figures), and to other location-based composite factors which shared
the remaining 33% proportionally to their cotton production intensity.

All factor payments of the “banking and insurance” activity (ABISE) went to
wage labor.

Home consumption

Home consmnption was observed with six activities. Except the activity “Other
nontraded services”, all the activities where home consumption was observed were
related to agriculture, again defined broadly to include, in addition to cropping
activities, fishery, livestock keeping, forestry and gathering. Home consumption was
proportionally distributed to different household groups, based the intensity of these
activities. Given home consumption, marketed outputs in the sub-matrix [Activities,
Commodities] are obtained for each commodity as the difference between total output
and home consmned output. The DNSI sources provided the rates of home consumption
of each commodity considered in the microsam.

Final household consumption
Households’ final consumptions were obtained by combining the information in the 1-0
table, the BBC 89, EMCES 94 and WAEMU 99.

176

EMCES 94 evaluated total household expenditures to 648,529 million CFAF,
which is very close to the 663,356 million CFAF indicated in the I-0 table as final
household consumption. However, the structure of the expenditures by main commodity
groups was drastically different in these two sources (Table E.3). One of the sources of
the differences is that, while consumption of the commodity “trade” was directly
incorporated to the final consumption values in the EMCES 94, it was separated in the I-
0 table and amounted to more that 19% of the total final consumption. However,
differences persisted, even when the consumption of the “trade” is distributed
proportionally across the remaining commodities groups. The most striking differences
include: (i) underestimation of the consumption of food items, particularly of the
products flom food and rice cropping activities; and (ii) overestimation of final
consumption of commodities such as agro-industrial products (CAIDT) and other
manufacturing goods (COMGG). There is no easy way to reconcile these drastic
differences, so we will use to the EMCES 1994 structure of household’s expenditures,
instead of the one in the I-0 table, as the latter is based on the much older BBC 89
survey. It is plausible to assume that the structure of final household consumption has
changed since the devaluation of the currency in 1994.

The EMCES 94 survey itself was carried out in the second quarter (March to
July 1994) following the devaluation, making the results very susceptible of “picking
up” extra short-term effects that are likely to inflate the budget shares for food
consumption. Thus, while EMCES 94 evaluated the budget share of food at 62% in
Bamako, WAEMU consum tion surveys indicated this share to be 50% in 1999
(AFRISTAT, 2001; p. 271) 2. The 1999 figures are likely to reflect more accurately the
situation prevailed in 1997, the year to which the Malian SAM was calibrated. It was
not possible to do similar comparisons for the remaining four location-based household
types. For these latter households, the most recent budget shares were the ones
documented in the EMCES 1994 survey. The final structure of households’ expenditure
retained in this study was a compromise between the EMCES 94 structure and the most
recent budget share information for Bamako.

For Bamako, we used a simple average of the two sources (1994 and 1999), in
order to minimize the short-term devaluation effects that inflated the 1994 figures,
particularly regarding the budget share of food consumption. For the other four
household groups, we have taken the percentage change in the average figures with
respect to the 1994 figures to reflect aggregate relative price effects that affected not
only Bamako dwellers, but also residents in the remaining part of the country. Thus,
these percentage changes were applied to each broad consumption categories to obtain
the budget share of that category. Further disaggregation of the category budget shares
into commodity shares was done while retaining the infra-category structure of the 1994
data.

In some limited cases, expenditure shares were taken flom the BBC 89, which
provided very detailed information on commodity budget shares. Thus, detailed budget

 

‘2 The figures, published in AFRISTAT (2001; p. 271), represent the weight of food production in the
price WAEMU index. This weight was 5,001 on a scale of 10,000.

177

shares for beans, groundnuts and forestry products were taken flom BBC 89. These
budget shares were applied to total consumptions in each location-based household
group to obtain the corresponding final consumptions. Among the three

forestry/ gathering products (firewood, furniture wood and gathering products) the BBC
89 figures left the furniture wood and gathering products aggregated. We therefore used
the shares of these two products in the total output (of the forestry/ gathering activities)
and adjusted them to be consistent with the budget shares of firewood for the five
household groups.

F inal consumption of the commodity “hides and skins” (CSKIN) was
determined in the same manner with regards to the livestock production activity,
particularly outputs of cattle, and sheep and goats. Budget share of “other meat”
(besides beef and meat flom sheep and goats) flom BBC 89 was distributed to two
commodities: “por ” (CPORC) and “camels and donkeys” (CDKCM), proportional to
output value, again using the relative importance of these commodities in the total
output of livestock activity.

The budget share of agro-industrial products (CAIDT) was determined as
residual, so that the constraint on the share of food in total expenditure is satisfied. The
expenditure share of “other food items” (COTHE) was primarily based on the shares
allocated to fluits and vegetables.

The I-O table showed that the final private consumption of “building and public
works” (CBBTP) was relatively small, if compared to the investment expenditures on
that commodity. EMCES 94 indicated budget share of lodging rent in total expenditures
in the five household groups, and these shares are used as proxy for final consumption
of BTP.

Budget shares for final consumption of “other traded services” (COTSE) were
mainly made of education and health services, excluding medicines and school
materials, which were treated as other manufactured goods.

Table E.3: Comparison of budget shares of main expenditure items in Bamako in
1994 and 1999

EMCES WAEMU Average of Percent

 

 

Expenditure Items 1994 1999 the two change

sources w.r.t. 1994
Education 0.80 1.65 1.23 53.1
Health 2.40 2.30 2.35 -2.1
Clothing 4.70 5.29 5.00 6.3
Lodging 10.10 18.71 14.41 ' 42.6
Transports and Communication 16.80 11.88 14.34 -l4.6
Food 62.00 50.01 56.01 -9.7
Others Expenses 3.20 10.16 6.68 108.8
Total 100.00 100.00 100.00 0.0

 

Source: Based on EMCES (1994) and AFRISTAT (2001)

178

Other traded services (mainly health and education) made up about 3.8% of
household expenditures, based on the 1999 figures. We assumed that this figure is
slightly lower in rural areas, about 3%. Similarly “bank and insurance services”
(CBISE), besides regular insurance services on transport equipment, were assumed to be
confined to urban areas, and to represent about 0.3% of total expenditures, as suggested
by the Bamako consumption survey. Final consumption of “trade” (CTRAD) was
directly incorporated to each commodity, so that “trade” can be better viewed as
intermediate input. Budget shares for “other manufacturing goods” ( COMGG) are
obtained as residuals over total budget shares, including an average of 3% for intra-
household transfers.

Using the information above, it is possible to obtain the structure of consumption
for the five household types. The total final consumption obtained amounted to 637,498
million CFAF, which is slightly less than the LC and the unbalanced macrosarn’s final
consumption entry of 663,356 million CFAF.

Trade

Data on cereal trade is highly inconsistent. For example, while the I-0 table showed no
trade in cereals, other DNSI tables evaluated cereal exports at over four billion CFAF.
Parallel studies, such as Diakite and Samaké (2002), evaluated the same exports at
nearly 1.4 billions. The high disparity in the data may be attributed to the importance of
informal cereals trade between Mali and its neighbors, most of which is occurring
through very porous borders. For the purpose of this SAM, we will consider that Diakité
and Samaké’s figures, which are based on the survey of informal cereal trade, are
closest to what is really happening in Mali. Furthermore, they constitute a middle
ground between DNSI’s zero and four billion extreme figures. Diakité and Samaké’s
figures have the additional advantage that they show the structure (in volume) of these
cereal exports, dominated at more than 80% by millet exports. We converted the
volume-based structure into a value-based one, using domestic consumer valuation of
each of the cereals studied by Diakité and Samaké. This reduced slightly the share of
millet in total cereal exports to 78.6%.

DNSI external trade data showed milk and eggs imports of about 9,233 million
CFAF, 98.4% of which can be attributed to milk and milk products, according to the
FAO statistical database. A similar procedure (i.e., total value flom DNSI external trade
source and distribution flom FAO database) was applied to cereals imports, which were
evaluated to nearly 8,821 million CFAF shared between rice (49.9%), wheat (48.9) and
maize (1.2%). Also nearly 2,083 million CFAF worth of exports of oilseeds and beans
were shared between groundnuts (91.4%) and beans (8.6%), a ratio suggested by the
1997 FAO trade data for Mali. About 45 million CFAF of beans were imported, a figure
taken directly flom DNSI trade database.

Imports and exports of the aggregate commodities were taken directly flom the
I-0 table. The West Aflican regional and non-regional distribution is based on the 2001
detailed analysis of the origins of imports and the destinations of exports. Also rice
imports, evaluated at 8,683 million CFAF were taken directly flom the 1-0 table. DNSI

179

estimated tobacco imports at 4,283 million CFAF, and this was used directly in the
microsam. Tobacco exports were negligible.

The DNSI trade database indicates imports of fruits and vegetable at 1363
million CFAF. FAO trade data suggests that roots and tubers made about 14% of that
figure in 1997. Furthermore, 94% of these imports were tropical roots and tubers such as
yams, and cassava, as well as other starchy staples such as plantains. These imports
were attributed to the West African region.

Imports of wheat were obtained as the residual over the 10,542 million CFAF
imports of the three commodities produced by the nonfood industrial cropping activities,
as indicated in the I-0 table. Thus, wheat imports were 6,259 million CFAF.

About 52% of exports of “other food products” (COTHE), mainly composed of
fluits and vegetables (mangoes, green beans, potatoes and onions), went to three
regional countries (cote d’Ivoire, Burkina Faso and Senegal). The remaining 48% went
to Europe (Diakite and Samake, 2002). Besides, DNSI trade data suggested that three-
quarters of imports of fluits and vegetables (COTHE) came flom within the region.

Fish exports are dominated by dried fish, which were exported mainly to
regional markets. Fish imports are assumed to be equally shared between flesh and
dried/smoked fish, mainly of regional origin.

Imports of furniture wood and firewood were assumed to be essentially regional
and made of fiuniture woods. Data were taken flom the DNSI trade database. Exports
were also assumed to be essentially made of furniture woods, distributed roughly in
equal proportion to the West Aflican and non-West Aflica regions.

Exports of cotton, gold and groundnuts were essentially non-regional, whereas
exports of most commodities were dominantly oriented towards regional markets. A
small share of the hides and skins produced may have been traded internationally,
particularly in the Middle East (Metzel et al., 1997). We assumed this share to be
roughly 10% of the value of exports.

Though we have used the best available information to allocate exports and
imports of most commodities by the destination and origin, we were not able to do for
all commodities. For all the remaining commodities not discussed above, we have used
the overall share of regional exports and imports as a proxy.

Imports of combustibles, main component of the “electricity, water and energy”
commodity (CEWEN), were essentially of a regional origin (about 98% according to
DNSI trade statistics).

Regional imports of textiles represents about a third (30%) of total textile and

apparel imports, and these of agro-industrial products (CAIDT) represented about 57%
(DNSI trade statistics). Also, about 3.9% of “transport and telecommunication”

180

(CTTEL) imports were flom regional origin, so were nearly 9.4% of “other
manufacturing goods” (COMGG) imports (DNSI trade statistics).

Regional exports represented about 12.9% of the total Malian exports in 1997.
Imports of regional origin were much higher, about 41.4% of the total imports in 1997
(BCEAO, 2001). The export figure is certainly an underestimation, because it did not
appear to have accounted for unrecorded cereal exports between Mali and its neighbors.
These exports are obtained flom parallel sources (as discussed above) and are included
in the SAM.

The 1-0 table estimated “trade” exports tol 1,590 million CF AF . Following the
format adopted in this SAM, these exports will be treated as marketing margins on
exported commodities, and these margins are paid by exports to the marketing margin
accounts.

Based on the entries discussed above, all exports added up to 341,177 million
CFAF, which is slightly less than entry in the unbalanced SAM, 361 ,610 million. The
balanced macrosam show a total export of 371,967 million CFAF, and the initial entries
were adjusted upwards to reflect the macro aggregate. A similar adjustment was
required for imports, which totaled 490,5 89 million CFAF whereas the entry in the
unbalanced macrosam was 486,200. Total imports in the balanced macrosam amounted
to 472,226 million CFAF.

Marketing margins

Most of the information on marketing margins on domestic sales and imports (as
percentage of marketed output) was provided in the DNSI’s I-O document. Exports are
dominated by livestock, cotton and gold, and the marketing margins on these
commodities were taken flom other sources. For example, Diakite and Samake (2002)
evaluated the marketing margin of livestock exports to be about 12% of the export
values.

Marketing margins on exports were assumed to be at least equal to domestic
marketing margins, which were used as lower bound. Where there is no information, we
used a generic margin 12%, which is the average marketing margins used in the DNSI
auxiliary accounts.

Applying the respective margin coefficients to each commodity resulted in a
total marketing margin that was slightly greater than the 126,500 million CFAF that
were entered in the unbalanced macrosam (Table 5.2). The difference was corrected
through a proportional readjustment of the calculated margin for domestically sold,
imported and exported commodities. In addition total marketing margins on exports
were constrained to equal 11,590 million, which is the I-0 table’s valuation of these
margins in 1997.

Variation in Stocks

181

The variations in stocks were evaluated in the 1-0 table to -22,789 million. This figure
corresponded to the sum of the variation in the stock of cotton (-28,514) and livestock
(5,755). The same I-O table showed a negative change in stocks of gold and minerals (-
30), as well positive changes in stock of food crops (8,426), other manufacturing goods
(5,533), water, electricity and energy (3,200) and other trading services (16,800). As
mentioned earlier, changes in stocks for traded services refer to equipments specific to
the delivery of these services. It is unclear why these latter variations in stocks were on
included in the DNSI “official” data on stock variation, which was the -22,789. Given
that the I-0 table was one of our main data source for building the SAM, we’ve treated
the official figure as inaccurate, and we used the provided details, which changed the
total variations to a positive 11,170 million CFAF in 1997. The change in livestock
stocks was assumed to be divided between sheep and goats and cattle, in proportion of
their output. The same proportionality condition was used to distribute the change in
stocks of food commodities to the five home-grown cereals included in the microsam
(i.e., millet, sorghum, rice, maize and fonio).

FBCF: Public and Private Investment

As explained earlier, the I-0 table retained six investment items: (i) livestock, (ii) other
manufacturing goods, (iii) building and public works, (iv) trade, (v) transport and
telecommrmication and (vi) nontraded services. Total investment was evaluated to
348,093 million, of which 31.8% was public (110,893). The public investment
expenditures flom the I-0 tables were drastically different flom the investment
expenditures published by the IMF (2002), which estimated them to 196,200 million
CFAF in 1997. The difference may be explained by the fact that IMF (2002) treated
expenditures on agricultural research and extension, and other public services as
investments, whereas the 1-0 table treated them as current expenditures that
corresponded to final consumption of nontrated services. We retained the I-0 table’s
treatment, and used the IMF (2002) figures to allocate the 110,893 million FCFA to the
six commodities that have received public investment.

Total investments on “other manufacturing goods” (COMMG) were 89,449
million CFAF in 1997. We allocated 42,100 million CFAF to the public sector, which
roughly corresponded to govemment’s investments in the secondary sector, as indicated
in IMF (2002). The residual was allocated to the saving-investment account.

Government investments in “buildings and public wor ” (CBBTP) were
obtained so that the total government investments equal the 110,893 million CFAF. The
difference between the total and govemment’s investments in CBBTP went to the
saving-investment account. Investments in livestock went to cattle, sheep-goats and the
group of camels and donkeys, proportional to their values.

Savings

Household savings were assumed to be proportional to their income. All foreign savings
were attributed to rest of the world (ROROW).

182

Distributed profits
Transfers by enterprises to households were marginal (58 million CF AF ) and was
assumed to go to Bamako households.

T tvces

Tax receipts on international trade were 112,500 million CFAF in 1997. This figure is
consistent across sources (DNSI and IMF). About 34,200 were attributed to value-added
tax (VAT) on imports, another 29,200 were allocated to special duties on petroleum
products, and 2,100 were allocated to duties on gold exports (IMF, 2002). The
remaining 47,000 million CFAF were attributed to tariff receipts on imported
commodities. The total indirect tax revenue was 149,700 million. The difference
between this total and tariff receipts on international trade (that is, 37,200) represented
essentially the VAT on domestic production. IMF (2002) estimates of VAT collected on
domestic productions was lower, about 26,500 million CFAF. We maintained the
37,200 million CFAF, which is consistent with the overall indirect tariff receipts of
149,700, as indicated in the BCEAO publication (BCEAO, 2001).

The normal value-added tax (VAT) rate in 1997 was 15% and the reduced rate
was 10%. The reduced rates applied to agricultural equipment and inputs, medical and
school supplies and chemicals. Exports are exempted, but a special export tax of 3%
was applied to gold exports (IMF 1998; 1999; 2002). We used these rates for each the
37 commodities in the microsam, but the resulting value-added tax receipts on domestic
productions were much higher than the 26,500 million CFAF suggested by IMF (2002),
let alone the 37 ,200 million FCFA retained in this study. One of the reasons for this
overestimate was that only 41% of the production of goods and services was classified
“modem”, that is, the taxable base of the economy. We therefore used the share of
“modem” production in total production to adjust for the VAT receipts, constraining to
total to be consistent with the 37,200 million CFAF mentioned above. A similar
procedure was used on the value added tax receipts flom imports, evaluated to 34,200
million CFAF in 1997. The difference here was that there was no need to adjust the
receipts by the taxable shares of production branches, as all qualified imports were
automatically subject to VAT on the ports of entries. The total value added tax receipts
on products and services were the sum of the receipts on domestic sales and imports.

Petroleum products (an essential component of commodity CEWEN) paid
29,200 million CFAF in special duties that included both import tariffs and the value-
added tax. We treated the special tax as part of the “Other Taxes on Products”. The
2,100 million CFAF export duties on gold exports were also treated under other taxes.

Factor taxes amounted to 10,305 million CFAF. Firms (including public firms)
paid 8,211 million CFAF and households paid the difference (2,094 million CFAF).
Households’ factor tax payments were allocated to the factor called “independent
entrepreneurship” and enterprise payments were assumed to be taxes on capital. The
TEI also showed a production subsidy of 13,097 million CFAF in 1997, but this does
not appear in any the other sources, such as BCEAO (2001) and IMF (2002). We,
therefore, did not consider this subsidy in the rrricrosarn. Considering such a subsidy

183

would have resulted in a negative net factor tax, which would imply a net effective
subsidization of production factors. These subsidies are likely to come in the form of tax
breaks of production inputs and equipments, and this was accounted for to a large extent
in the relatively lower tariffs and taxes collected on these products.

Households paid a direct tax of 28,850 million CFAF (income tax and social
security payments), distributed to the five households using the shares of wage income.
They also received a government transfer of 28,489 million CFAF. This income was
distributed to household using their shares in total wage income as weights.

T arm’s

According to the 1997 1-0 table, average tariffs on “other manufacturing goods”
(COMGG) were about 6.9%. The government perceived a total import tax (VAT +
duties) of 5,533 million CFAF on a subset of COMGG imports that was valued at
80,149 million CFAF. Assuming the average rate applied to all COMGG imports, the
total duties collected on this commodity would be about 28,735 million CFAF, 30% of
which are tariff revenues.

Nearly 66% of agro-industrial (CAIDT) imports are made of sugar. The tariff
rate on sugar was about 70% in 1997 (55% for variable tax on imports and about 15%
normal custom and fiscal duties), while average duties on the remaining components of

the CAIDT composite commodities were 15% on average. Thus a weighted average
tariff rate on CAIDT was 51.3%.

Diakité and Samake (2002) evaluated the import tariff on rice in 1997 at 36%.
The same rate was applied to other imported products that were also produced
domestically. An average rate of 15% was applied on all remaining commodities and the
total tariff receipt was constrained to sum up to 47,000 million, as discussed earlier.

Transfers

There was no basis for distributing income transfers flom or to abroad by the five
household groups. We arbitrarily assign initial distribution of these transfers, which
were confined to urban households. (Though some rural households may have transfer
with the rest of the world, it is assumed that the transfers are made or received in via
urban service centers). Bamako dwellers receive and send three-quarter of these
transfers whereas other urban dwellers transact in the remaining quarter. Also unknown
is the share of West Aflica is these transfers. We again arbitrarily assign initial values,
sharing the both the incoming and outgoing transfers equally between the two foreign
accounts: West Afiica and the rest of the world without West Aflica. Necessary
readjustrnents were left to the cross-entropy SAM balancing process.

184

Appendix 5.3: Ratio of Expenditures to Incomes in the Unbalanced Microsam

 

 

 

Account Account Description Ratio expenditures
Name to incomes
AFOOD Food production, excluding rice production 1.00
ARICE Rice production 1.00
AINAG Industrial agriculture, excluding cotton production 1.00
ACOTT Cotton production 1.00
ALIF I Livestock keeping and fishery 1.00
AF OGA Forestry and Gathering 1.00
AMDII Mining 1.00
AAIDT Agra-industry, beverage & tobacco 1.00
ATEXT Textile production 1.00
AOMGG Production of other manufacturing goods 1.00
AEWEN Production of electricity, water and energy 1.00
ABBTP Construction and public works 1.00
ATRAD Trade 1.00
ATTEL Transport and telecommunication 1.00
AOTSE Other traded services 1.00
ABISE Bank and insmance services 1.01
ANT SE Nontraded services 1.00
CMILL Millet 0.14
CSORG Sorghum 0.21
CMAIZ Maize 0.28
CFONI Fonio 0.17
CROOT Root crops (potatoes, yanr, cassava) 0.23
CBEAN Beans 2.1 l
COTHE Other crops 1.03
CRICE Rice 0.50
CGNUT Groundnuts 2.74
CT OBA Tobacco 0.57
CWHEA Cotton 1.20
CCOTT Wheat 0.90
CCATT Cattle 0.86
CSHGT Sheep and Goats 0.60
CPORC Pork 0.60
CDKCM Donkeys and Camels 0.35
CPOUL Poultry 0.33
CEGGS Eggs 6.65
CMILK Milk 1.14
CSKIN Hide & Skins 0.39
CFFIS Firewood 0.70
CSFIS Furniture wood 2.83
CFWOO Hunting/Gathering products 8.37
CWOOD Fresh fish 0.58
CGATH Smoked fish 7.68
CPGLD Gold 0.93
CAIDT Agro-industry and beverages 2.39
CT EXT Textile products 1.34
COMGG Other manufacturing goods 1.08
fiWEN Electricity, water and energy 1.25

 

185

Appendix 5.3: Ratio of Expenditures to Incomes in the Unbalanced Microsam

 

 

 

 

 

 

 

 

 

 

 

 

 

(continued)

Account Account Description Ratio expenditures
Name to incomes
CBBTP Construction and public works 1.06
CTRAD Trade services 1.02
CTTEL Transport and telecommunication services 1.14
COTSE Other traded services 0.80
CBISE Bank and insurance services 1.23
CNT SE Nontraded services 1.09
MMDOM Marketing margins on domestic sales 1.00
MMIMP Marketing margins on imports 1.00
MMEXP Marketing margins on exports 1.00
FCAPI Capital 1.27
FWLAB Wage labor 0.95
FIENT Independent entrepreneurship 1.01
FTRAD Composite factor for flading services 0.96
FOUAC Composite factor for other urban activities 0.98
FFCKS Composite farming factor in Rural Kayes-Sikasso 0.96
FFCRV Composite farming factor in Rural River Region 0.96
F FOTH Composite farming factor in other rural 0.96
FORAC Composite factor for other rural activities 0.98
HBMKO Bamako residents 1.44
HOURB Residents in other urban zones 1.24
HRUKS Residents in Rural Kayes-Sikasso Region 0.84
HRURV Residents in Rural River Region 0.87
HORUR Residents in other rural zones 1.17
ENTRE Formal enterprises 1.00
GRVAT Value-added taxes for government recurrent account 1.00
GRT AR Tariff incomes for government recmrent account 1.00
GROTP Other taxes on products 1.00
GRFAC Factor taxes for government recurrent account 1.00
GRDTA Direct tax on institutions 1.00
GROTH Other taxes for government recurrent account 1.00
GRECU Government recurrent account 1.00
GINVE Government investment account 1.00
SINVE Saving investment account 1.00
WESTA West Aflica 0.53
ROROW Rest of the world, excludingWest Africa 1.14
STCHG Change in stock 1.00
Source: SAM

186

 

5m 040.5

NwN— _ SN 0830
Na— 3” — . COBHU

n . A. mEmU

m 5— min—U

eBN u—q—EU
$2 mOOmU
ova .520
o — N EUv—OU

N2 N3. ammo

chm mm 3 51:50

3: new. 5 36— :000

E. tin—:30

0v m (508

5.— wmon oav 5200
no. A. NUS—U

N3 mEOU

vn— ~20mU

an new OdOmU
mm 0mm _ 13:20

mma< mth< APP—k Dg< mhmm< Zm3m< OOEO< EXP—k hQ-<< ~Z=z< <00m< _..=.._< EOU< O<7=< mUS—e. GOOL<
eons: 5.25 2:56 an: ..e 552 “52.884. .525 32.2.5 83 < "...m 52.2.3.

187

 

 

 

3ch we»: 3an. @3th mmNcmN .vam 83a «mNaN gnaw vaN: can; v33. cemmN— an—N 830 £29 .._<.—.O.—.
010.5
30MO¢
firmer»
m>z._m
m>EO
DUMBO
EOMO
(4.—.20
0(4me
PHOy—O
asp—O
h<>¢0
Elm
«5&0:
>135;
45de:
EDD:
032m:
man. :8 95mm mNew e _ ON new N _ 2 nDVN— «EN New” 82H 0:0..—
553 Saw” ammo v8— 3: on _ RNM mason new. moNN Now: 3&Nm IHOHE
mo:— ahnb glen :2. com— won En; ogmo 03.0— wwNv 2v: whmmm >mUHE
852 SAN men _ n mm 5 EN meN NMONN on: no — N 535 we?” won: Ne _ co much—m
NmNS new: mom... ovoN 8N N am amm — _ _ _ Now U<DOm
—Nc_ mmaNN 2E
2.0m— N . mm. 83. em _ 2 23 chN film:
wamm :Nv NS 2.2: on: 2.5—N 5— .9K 2.va m<43m
mONm Nw Em mm»: N. _ .N an. MN 730.»
53:22
$3522
2822
mmeU
ago Nv ”NVN— moan anv m _ v m2. SN 3&— mwN_ mmEU
33 mm Nsma new. _ mNVN— $1 33 OS. ween h _ 2 3A: awNm mthU
we: «can ocoN veg NNww So ONNN men no: — S «v 52 N3 VNn BNN om: — APP—U
m3 whoN wemN an: 3ch :NM 2.:— ENM he: men. :0 3mm mace Nmn 35m wnN_ GED
Nwm 3N— Nmon 5n 3: b9 2% 2: N 31 Nw Ahmmu
m9. 3.; SN: omen :2. 30m mvcm cnvN 3.2 thn 30m 03: .m— «:1— oom ZmBm—U
o2. . Etna— occm annex mmoN eNNmn mgN meow ocQMN N _ mm 3mm 83..— mn _ N «a _ h mag 00200
82 men cm— 2. EXP—b
:0 exam nno PAH—<0
mma< mmho< APP—k §< anm< Zm3m< 00—204. .Cmmfik hn=<< _Z_E< <OOm< 7:111 EOU< D<Z~< MUS—e, QOOm<

 

€35.88 .3: ..e 552 Eeaaoofis 325 32.2.5 8.: < "am 522:?

188

 

mm
mm—

$—
0:.

cm
.3
:2

cm. NGSN Noe—m
ccme—
:2 .3. 33
men:
ammNm QVN 54m— 3: New w _ mmN Nommm
UMOLU hommo .E.<UU ...—.000 <m§u (£08 #3200 mus—U th00 7738 HOOD—U ~ZOh—U N_<ZU OxOmU 41:20 mmHZ<

€85.88 :4: E 552 $5282., 325 885.5 32 < "am 522.3.

0420
E<OU
00030
COED
m-mmo
mrEU
2.9—mu
M520
wOOmU
1520
ZOO—DU
“Va—0&0
hOImU
.E.<UU
:OUU
<m=30
<m08
SZOU
MUD—U
MEHOU
Z<mmU
hooxu
~ZOn—U
N220
OKOmU
13:10
mmhz<
mm_m<
mtht.
APP—k
DAD—.54
Lhmm<
Zm3m<
DOZO<
EXP—k
.—.D_<<
_Z=2<
<8m<
—..=4<
:OU<
O<Z~<
mUE<
0005‘

189

 

 

883 :N can: 35 5 082— 2 z: 3.3 v3: 2.34m «23 cow coaN mMNN am;— mnmmm R: E 00030

mi mnmn 530— n _ n N on mmN @750
0: mm: m 33 NN min—U

cww vow» we .E.<UU
v me: See own an SN Nw —N am— a ma aN SN 32 M an :000

A. _ on FOOMU
wN — v ~ZOmU
No3” N~<EU

3cm Ads—U
mmcv mwhz<
oomv mum—mix
3mm mmh0<
9.2.0 AMP—Li
m3”— mme 32 9 m8— mh _VN— NN onNN 05.5,

83 E: 8% NS _ _ n3 ca Ahmm<

S :2 SN? 83. noun 0N3 co n5 32 we 2‘ new» ONNN— Zm>2m<

3:: #9.:
303 _2=2<
NhoaN (00.74
was 7:12
gemN :OU<
3N3 0<Z~<
one mug
go 000:
UMOAU HOZmU .E.<UU ...—.000 <m230 (mp—.0 SZOU mun—U mEOU Z<mmU ROOM—U .ZOmU 5520 OMOmU 4:20 Mmhz<

83.528 :42 .8 8543:2883. 38m .85.:— 8.: < "am 55%?

190

 

 

MMNcMN
.vaM
eohwa
vMNoN
canon
Nv¢n__
heme mmon NMOm
nmom some. mmV. momm New. cvom ewwm

mhmmu 295.90 00200 .585 5.0—<0 GAGA—U Ih<OU D830 OOBHC wEmU min—U 2350 v3.20 mOOmU 1.30.5 SUV—GU

€955.8an :32 .8. xtaazdeteseuoe. Enem vegan—am 53 < 38 5.52:?

DAOLU
*Pfi<mXu
Aunxu>>nv
nXu>9mnv
wEmU
mpmunv
~Anxwo
“flu—<9u
mmxumHv
Afizuanv
20x00
Huxnxhu
EOIwU
.Fh<nxu
hhhfiuu
Atmvtshv
¢J=Uhhv
.rfiszxu
mHuanv
mthU
Z<mmU
hhxuxnv
.ZOmU
Nfi<§éu
Omcmu
Aw=eznv
Mmhz<
mma<
mmHO<
.Hmhhgx
AH¢JPE<
.wrmmax
.Lm>?M(‘
mvm=2n2<
hymwfi<
hh=l$<
_7=<4<
rcfixum4‘
ES<
thHu<.
mv<7=¢‘
mn=g5<
.unXum¢‘

191

 

NemNMN _NQMN— N~oa_v
owN chcmN

«noun ceca—

mocmN
«N mmwn _mNN_
oGMN sav~ hNMQ-

vow cmcmN

wow—o me_N_ mmomN_
o_NN_ N_mN

w_v> Nobo—

__mo
no.0
woov _me_

mm
owns chN
Nwo_ NNv

cosh

oNN

v~o_—

Nmov

woe
new

n.a—

MNoc NcNm

no
em. o_N

mm—
mm—
hmw ova

_NwoN

—a_—

no.
v_w

0.0
mmov

emom mNOON

enow
moN

enn—
—w swa

wc_N
oNNN
omN NNNN

_No—

mm

m—
we

c—
moN

vam

no.

va

mom

£§¢_ A<HOF

wm_ h<>x0

mom mynwegv‘
8mm 20022

mmEU
mthU
.mwrru
nzfimhhv
arrmfinu
~Lm>>mnv
maveenxu
gunmuhu
urn—<0

...—Emu Emu 00200 .588 hQ—<U DADA—U E<OU 90030 834.0 mEmU mEmU 2350 21:20 mOOmu 150.5 ZOO—DU

€35.58 :4: he he“: $5.583. Boom 82.2.5 82 < "am 559:,

192

 

 

83mm
envom
mmmmw
33m.
$35

Ibo: >6"; BmUmm U<DOm Cg HZmE m<E 720..— me22 5.522 2822 mmHZU Mum—mu meOU 15E fig

€35.88 is ...... 5553:5293 .325 382:: 32 < "1m 5593

GAGA—U
IH<OU
90030
COED
mEmU
9&0
2550
31:20
MOOmU
4320
20x8
UMOmU
hOimU
.E.<UU
:000
«5:30
(mp—.0
PDZOU
mUD—U
n.a.—FOO
Z<mm0
FOOMU
Eons
NED
OKOmU
AdZU
umhz<
mma<
umh0<
APP—k
§<
Ahmm<
295%
0020<
EXP—k
he:
=1=2<
(005‘
53<
:00<
O<7=<
MUD—4‘
900%

193

 

 

whoa: 351K name; 32% N59 9.2.9 n33. NMNNE mZS 52m 3:0 omommm :5.” can: @932 uneasm 4499—.
. DIP—M

8N.“ 334 w: n 33 BOKOK

comm m2 (Emma

m>7=m

m>Z_0

Doug—0

mane EOy—O

(hm—do

cm: Sm”. 0.520

.5010

N _ 2 N92. amvm «SS—O

vam 32 no.2 $2 h<>xo

3.8— m_mnm Nmmfi. g2”..—
msoo: Nmnm Nowv enema MDKO:
nonvom ace 8N2 om; >135;
meme—m an: :2: Nam >63“:
83m :2: 3.8 fnvm mud—DO:
nmmg 34: 88v Ov—sz

0&0.“

Eon—..—

>¢U~E

wxut

U<DO..—

GE

Him—..—

@133..—

3(9—

flaws—2

Ava—22

283:2

mmPZU

mmEU

umbOQ

4mg

2;— 82m 32b 0%

F550

ZMBmU

00—200

EXP—.0

PAH—(U

 

Eot >6: 38k 9.58 92: SE 3.5; :5: ..me2 .35: 28.22 mmhzu $8 $98 48.6 928
€253.39 =a2 ...... 55:2 atuscuuaV Eoem 63:23— 33 < 3am n.a—3%:

194

 

 

0.9.0

 

New 8a. .moN no. a. ......<OU

mN «N D0030

Nmm eNa. mch .... 3. 003.0

Nun ovo— omo ..wm mmn m....mU

Pmm mom» 3:. ..me vow. mimo

..N. mm. ngu

39 . .mm we: a... «NmN $4.20

we. won oo. . .N mOOmU

...v n... vaN . .v .3 ADOLU

3... mm... mom. 3: .w. . 3N. 3.0300

3.. .n on UdOmU

2km mmm va mmcm 3m. as. HOImU

no.2 0mm. .mwN Nb; ooNv 25¢ .....<UU

.....OUU

No.N «av. Nmmm 8.. N.NN €5.30

3v. 2%. mm: 0N» cNh «59.0

N2. .3 NN cm ...DZOU

vwow wMNo N53. no: 003 muzu

.34. 83. BNm . on .v 30¢ thOU

..n v Z<mmU

mmm 00.. c . m wen. ...OOxU

80 3.. Now Na Na .ZOmU

VNem whom an? 8.. .oh N.<S.U

nwww I...» «08. NVNN acmN OMOmU

.mom. 2.8. 8.... NwwN 03m 44.3.0

mwhz<

mw_m.<

mwov .cmN m .3 5mm ow. mm...O<

amt<

2.5.

.....mm<

Zm3m<

OOZO<

...Xm...<

...D.<<

.Z..2<

hm . ..N :8.” .93 <00m<

mmowN «mama omo N. . .n..\.<

...—.004.

memN mMNN «NS 2. mm O<Z.<

oowMN ON . N. 8%. wmw E mU§<

m8. .Nccv as; .3 an 900......
m.>Z.m m>7=0 DUmmO ......OMO <...D~.O U<m¢0 ...HOMO d<...~.0 ...<>MO mEZm MDmO... >¢D¢I away... EDD... 0235.. 0’20“.

 

82.5.58 a“: ...... Eszwégsé .325 .3328 5: < new 522:?

195

 

mmm.mm

.mNe..

ww.vov

nn.nc nmcme mono.

wee.m omvnv NQONN aovmn. hwnoNN

McNon

mwwmmn

cmvmo.

moooN. Novh..

A<h0h

 

nncm

.mNo..

mmNN
owvmv.

ommmv

meow.

Nnm

wheNo

«moon

mmoav
mc.hN.
mm.mo

mmcwo momc.

cone

mmoN

mwmo.N

VNNN.

oov.N. MNom.

voo.n omvhv NmCNh

moNom w.mm

VN

meow
on.m

mwN
thn
mme.
vaec.
mo.w.

vM.mN

v.5m

Noon
NemoN

.Nw
wvmw
mooNN
.com.
mnwVN

vomVN

va

awn”
mmcoN

«we
”n.a
vanm
c.ow.
n.a.v

noon

cam..

ammo

on.
mc.m
wvoo.

cch
omNo
mo.w
oon
n.h..

own».

comb

QnNo

m.cw¢
acme.

h¢ec.

va
ommm
v..w.

w.ov
mva
cm.”
Noac
Mva.

OIUhm
BOMOM
arrmm>>
mw>7=m
m>z.0
NKuMMzu
IEkuxmv
A<hthmv
ne<m¢mv
mhhvxmv
u<h¢0
h<>¢0
wahggm
xDMOI
>x3¢1
253.5.
my:.n:.
AUV=2M=.
“U¢J=vm
wrrmymm
>xUmL
mofiumm
U<DOm
..(Jwrm
hzm...
m§<4>>m
.L<Um
.hxme4s.
m3=e4=.
(anxuvgvg
mthU
mmau
mthU
.mhLLu
ndfimhhu
mhhxmu
..m>>mnv
mxuegnxu
.hxmwru
...n..<U

 

m>5=w

m>5=mv

DUMKO IHOMO (FQMO U<mm0

abomo ¢<hMO h<>mo M¢HZm maxO:

>x3m1

>53”...

md=.n:.

OXEmI U<MOL

 

.3225. is 3. gazduuaasé 38m .5223 5.: < 3% 55...?

196

 

 

 

 

3% «was 0 . h... . 4<...0...
mmVn 0:0...m
mwNavv 30x0y.
0.55.. .firmms?
own . mm m.>Z.w
.nNe. . m>Z.0
mm .vov .Nmoo. 3030
mm . no ...—.010
$80 «$.30
men... 0<h.~.0
V8 . n .....0M0
can». .....th
NQKN h<>¢0
003K. . «new Em.
SmoNN J”...—v.0...
noNaNm >~.D~....
mwmnmm 39...»...
3.35. show . £30...
macaN. a . ..3 oNho . 02m...
Nev... . 044.0“.
38... ...—0......
hcnvoN >.~.0......
chv . n 3.0....
«9.3 0<30...
N . cmm 0%
2.2.0 ...Zm...
3‘03. m<43...
NmNNo. 3%. :20“.
n. . N . 00.22
.c. .m 2.22
no . 5 20022
oNamMN mmhvhu
52.3. mm.m.0
cmwN. . come. «3. N8. mm...00
ocmoe. «.Nom ..MELLV
mmoabN 0%
SmNMN ......mm0
.thN. .3 m 7538
N .3 . v mmmm 00200
wow .0 ..st
me . N. ...n..<0
4<...0... 0:0...m 30x0”. (....me

 

cash
¢.o..
MNoo
NcNo
.NmoN
vmmm
”NoeN
.Nm.
vam
evvo.
.vN
oom.n
voa.o
oownN.
m..o.
chm”
vmae.
mhvvm
no.0v
woe
oooN
mnNN
am.v.
mmnmm
oc..m

Show
wNm . o.

.28.. 0:05 39.9. Emma
€035.50. ..a2 ...... 5.552 nut—583. .525 wanna—am E...— < "1m 56:2...<

omvN
Ema

GNNVM-

oooN

CNN
. .e
98.
83

.thm.

wmwm

man.

an...

.Nw

.N.

8....
.nww.

th
.wm

NMN

av:

......<00
00030
003.0

me0
9......0
2050
0.4.20
m00m0

430.0
200.00

0M0..0

...0...m0

......<00

......000
(H.520

450.0

......200

m0.~.0

m......00

Z<mm0

...00M0

.ZO...0
N.<20
0x0m0
44.20
mmhz<

mm.m<
mm...0<
4m.....<

2.5.

.....mm<

Zm3m<
0020<

..Xm...<

...n..<<
.Z.2<
<00...<
.m.4<

......00<

0<Z.<

m0.y.<

000....

197

Appendix 6.1: Base Production and Trade Elasticities

 

 

SAM Account SAM Account
and production & CES and CET

Activities elasticities Commodities elasticities
Food production excluding rice AF OOD 0.45 Coarse grains CCORS 1.50
Other food COFOO 1.50
Rice production ARICE 0.45 Rice CRICE 1.50
Industrial ag excluding cotton AINAG 0.45 Other industrial. ag products CINAG 1.50
Cotton production ACOTT 0.45 Cotton CCOTT 1.50
Livestock keeping and fishery ALIF I 0.45 Cattle and sheep and goats CCASG 1.50
Other livestock COLIV 1.50
Fish CFISH 1.50
Forestry and Gathering AF OGA 0.45 Forestry & gathering products CF OGA 1.50
Mining AMINI 0.60 Mining products CPGLD 1.50
Agra-industry AAIDT 1.50 Agra-industry and beverages CAIDT 2.00
Textile production ATEXT 1.50 Textile products CT EXT 2.00
Production, other manufactures AOMGG 1.50 Other manufacturing goods COMGG 2.00
Production, elec. water & energy AEWEN 1.50 Electricity water and energy CEWEN 2.00
Construction and public works ABBTP 0.95 Construction & public works CBBTP 1.50
Trade ATRAD 2.00 Trade CTRAD 1 .50
Transport & telecom ATTEL 2.00 Transport & telecom. CTTEL 1.50
Other traded services AOTSE 2.00 Other traded services COTSE 2.00
Bank and insurance services ABISE 2.00 Bank and insurance services CBISE 2.00
Nontraded services ANTSE 2.00 Nontraded services CNTSE 2.00

 

Source: Adapted from Decaluwé, Dissou and Robichaud (2000), p. 9.

198

Appendix 6.2: Base and Calibrated Data for Selected Indicators

 

Calibrated Percent change

 

Items Base Year Data Data from base
Total absorption 1421.0 1466.4 3 .2
Total investment 209.7 209.7 0.0
Public expenditure 219.9 248.6 13.1
Total exports 467.5 467.4 0.0
Total imports 534.6 554.9 3.8
GDP, market price 1461.7 1401.4 -4.1
GDP, factor price 1343.5 1272.4 -5.3
Net indirect tax 118.2 129.0 9.1
Private consumption, including: 991.4 1056.8 6.6
Bamako 119.8 139.5 16.4
Other urban 84.5 93.1 10.2
Rural S-K 300.2 321.5 7.1
Rural River 287.0 273.9 -4.6
Other rural 199.9 228.7 14.4
Total factor income, including: 991.4 1056.7 6.6
Capital 187.2 175.7 -6.1
Wage labor 175.6 188.3 7.2
Composite independent entrepreneurship 54.6 50.7 -7.1
Composite trading services 15.0 13.2 -12.0
Composite other urban activities 52.5 48.1 85
Composite farming , Rural S-K 320.2 310.6 -3.0
Composite farming, Rural River 264.7 232.2 -12.3
Composite farming , Other Rural 182.1 173.7 -4.6
Composite other mral activities 91.5 80.0 -12.6

 

Source: Author

199

Appendix 6.3: Final and Intermediate Demand Prices Under Price and Market
Access Scenarios (% change from base)

 

 

 

Items BASE FAPRI IFPRI] IFPRIZ OECD DFA EEP

Commodities Consumer Prices (Percent change from base)
Coarse grains 0.7 -6.7 -2.0 -8.9 -0.1 -l.7 4.3
Other food 0.9 -12.6 -10.2 -3.1 -6.8 -9.1 5.1
Rice 0.9 -3.0 0.1 -2.0 -3.0 -1.6 3.9
Other industrial. ag products 2.1 -10.4 -l3.2 3.9 -8.7 -12.5 18.3
Cattle and sheep and goats 1.0 -1 1.5 -19.6 -2.6 -38.9 -9.8 4.2
Other livestock 2.4 -25.5 9.8 -12.0 0.9 -2.5 -21.9
Fish 1.4 26.3 -1 .7 1 1.5 -17.0 -36.0 -5.8
Forestry & gathering products 0.8 0.9 2.2 0.0 -1.2 -2.8 -2.5
Agra-industry and beverages 1.9 20.5 -40.8 3.0 12.4 16.1 -15.6
Textile products 1.9 8.5 75.9 13.6 20.6 -2.8 -35.7
Other manufacnu'ing goods 1.2 -4.0 11.3 -13.9 -1.8 -4.2 6.2
Electricity water and energy 1.0 35.3 13.8 52.3 37.1 67.2 33.2
Construction & public works 0.6 1.4 11.2 0.7 1.2 0.8 9.8
Trade 0.4 2.5 11.0 2.3 1.8 1.9 11.7
Transport & telecom. 0.9 3.4 23.6 16.2 0.6 -8.7 7.7
Other traded services 0.3 29.5 -41 .6 42.8 9.7 31.7 71.5
Bank and insurance services 0.5 4.9 15.5 8.2 2.6 3.8 16.2
Nontraded services 1.0 0.9 -0.6 3.7 1.2 0.7 -4.9

Activities Price of Aggregate Intermediate Inputs
Food production, excl. rice 0.9 -1.2 8.2 0.9 0.7 -l .3 7.3
Rice production 1.1 -3.6 2.9 -1.6 -2.0 0.9 7.4
Industrial ag, excluding cotton 1.1 0.1 11.4 0.8 1.0 -1.2 7.7
Cotton production 1.0 2.1 11.9 1.5 1.8 2.2 11.7
Livestock keeping and fishery 1.1 -1.6 3.9 1.4 -6.2 «1.3 4.7
Forestry and gathering 1.1 -0.4 10.0 -2.8 0.4 -0.7 8.1
Mining 0.9 1.6 10.6 0.5 2.0 2.3 10.0
Agro-industry and beverages 1.2 3.7 -7.7 0.3 3.3 5.4 3.6
Textile production 1.0 5.5 28.7 7.8 8.5 3.4 -1.9
Other manufacturing goods 1.0 2.0 11.5 1.4 1.7 2.1 11.1
Electricity, water and energy 0.8 5.9 12.8 8.1 5 .3 8.3 14.6
Construction and public works 1.0 1.4 1 1.2 0.7 1.2 0.8 9.8
Trade 0.8 5.8 16.9 9.5 4.1 4.3 13.3
Transport and telecom. 0.8 4.6 12.7 5.9 3.8 5.9 14.0
Other traded services 0.9 7.8 14.1 13.6 7.5 9.6 13.2
Bank and insurance services 0.8 4.9 15.5 8.2 2.6 3.8 16.2
Nontraded services 0.9 4.0 13.2 6.5 3.4 3.6 11.5

 

Source: Author

200

Appendix 6.4: SummaryMacroeconomic Indicators

 

 

 

Macro Indicators BASE FAPRI IFPRI] IFPRIZ OECD DFA EEP
Total real absorption 1443.2 2.3 2.0 0.9 2.4 1.2 -2.3
GDP, Market Price 1401.4 1.4 1.9 0.1 2.2 1.1 -1.6
GDP, Factor Cost 1272.4 1.8 1.0 1.3 0.7 0.1 -2.3
Total real household consumption 1056.8 3.2 2.7 1.3 3.3 1.7 -3.1
Total real exports 479.9 -0.2 0.3 -2.9 2.7 2.4 -0.1
Total real imports 567.0 3.1 0.8 0.2 3.1 2.7 -2.7
Real exchange rate 114.8 -5.4 8.9 -3.6 -3.1 -2.0 12.2
Nominal exchange rate 129.8 0.4 9.3 0.2 0.9 0.2 8.2
Export price index (FCU)——(1) 100.0 - - - - 0.7 -3.1
Irrrport price index (FCU)-—(2) 100.0 -2.2 -0.5 -2.0 -0.5 - -
World (tradables) price index 100.0 -1.2 -0.3 -1.1 -0.3 0.3 -1.4
Domestic (non-tradables) price index 100.0 4.8 0.1 2.9 3.8 2.6 -4.9
Terms of trade =(l)/(2) 100.0 2.2 0.5 2 0.5 0.7 -3.1

 

 

Source: Author

201

Appendix 6.5: Imports (Percent change from base)

 

 

Imported Commodities BASE FAPRI IFPRI] IFPRIZ OECD DFA EEP
Coarse grains 0.1 -0.2 -0.2 -0.2 ~0.2 -0.2 -0.2
Other food 2.0 -20.9 -30.4 -5.0 -11.8 -l3.7 -1.0
Rice 8.6 - - - - - -
Other industrial. ag products 20.5 -17.1 -33.2 8.6 -15.8 -22.7 19.5
Other livestock 19.4 -24.8 2.1 -9.1 -0.9 —2 .4 -28.7
Fish 2.2 32.3 -12.0 14.3 -24.1 -50.7 -15.7
Forestry & gathering products 2.1 - - - - - -
Agro-industry and beverages 39.9 46.3 -38.8 26.3 14.7 12.5 -18.1
Textile products 39.7 6.3 46.1 10.3 15.3 -2.6 -34.3
Other manufacturing goods 355.6 -3.5 1.4 -11.4 -2.1 -3.6 -1.6
Electricity water and energy 61.9 27.0 3.3 39.8 27.8 50.6 18.1
Transport & telecom 8.3 4.5 19.2 23.3 -0.5 -13.4 -0.7
Bank and insurance services 6.6 3.6 4.5 6.3 1.3 2.8 5.9
Total 566.9 3.1 0.8 0.2 3.1 2.7 -2.7

 

Source: Author

202

Appendix 6.6: Disaggregated Real Household Consumption (Percent change from

 

 

 

base)

BASE FAPRI IFPRI] IFPRIZ OECD DFA EEP
Bamako 139.5 12.2 -3.5 4.0 12.5 10.6 -8.2
Other urban 93.1 10.3 -0.7 6.3 10.3 7.3 -6.0
Rural S-K 321.5 0.0 10.2 -0.1 2.2 -0.5 0.1
Rural River 273.9 0.0 1.5 -0.2 -l .2 -1.5 -2.2
Other rural 228.7 3.2 -1.1 1.4 1.7 0.7 -4.6
TOTAL 1056.8 3.2 2.7 1.3 3.3 1.7 -3.1

 

Source: Author

203

Appendix 6.7: Implementation of the CET Scenario

 

 

Calibrated CET Percent Share of Changes to
Commodity import tariffs Tariffs change ROW imports import prices
Coarse grains 16.4 8.0 -51.2 92.1 -1.9
Rice 53.3 8.7 -83.7 100.0 -23.6
Other industrial. ag products 8.0 5.0 -37.5 98.4 2.5
Other livestock 17.2 4.8 -72.1 96.1 -5.2
Fish 14.6 10.1 -30.8 0.0 -6.8
Forestry & gathering product 16.5 10.1 -38.8 0.0 -8.6
Agra-industry and beverages 27.8 14.7 -47.1 1 1.3 -16.0
Other manufacturing goods 4.5 12.0 166.7 92.9 12.3
Electricity water and energy 6.7 8.7 29.9 0.5 -1.5
Transport & telecom 34.6 8.5 -75.4 98.3 -15.4

 

Source: Based on Chapter III and the calibrated CGE model database

N otes: The table shows the calibrated import tariffs from the CGE model, the CET rates from Chapter III
(aggregated to account for commodity aggregation), the percentage change from the calibrated to the CET
rates, the share of the rest of the world in total imports, and the transmitted changes in equilibrium import
prices. The effective CET rates applied in the CET scenario are weighted by share of ROW in total imports.
Commodities such as fish and forestry and gathering products are 100% source from within West Africa
(ROW’s import share is zero), so that the effective CET rates on these commodities are zero for these
commodities.

204

Appendix 6.8: Real Imports (Percent change from base)

 

 

Imported Commodities BASE CET BANS GINV
Coarse grains 0.1 -0.2 -0.2 -0.2
Other food 2.0 -16.9 -7.8 -17.2
Rice 8.6 - - -
Other industrial. ag products 20.5 -11.4 -7.1 22.2
Other livestock 19.4 -3 1.4 -28.8 -26.3
Fish 2.2 -17.8 -14.3 -17.5
Forestry & gathering products 2.1 - - -
Agra-industry and beverages 39.9 39.7 16.8 8.1
Textile products 39.7 -22.8 0.0 -5.2
Other manufacturing goods 355.6 3.8 -2.6 8.5
Electricity water and energy 61.9 15.9 17.6 -51.4
Transport & telecom. 8.3 -8.2 -0.5 8.5
Bank and insurance services 6.6 4.2 5 .0 3.5
Total 566.9 3.6 0.2 -0.2

 

Source: Author

205

Appendix 6.9: Real Exports (Percent change from base)

 

 

Exported Commodities BASE CET BAN GINV
Coarse grains 0.7 0.0 -100.0 0.0
Other food 6.1 14.3 8.3 8.9
Rice 0.1 -3.8 -100.0 1.0
Other industrial. ag products 6.3 10.9 6.1 -13.2
Cotton 163.3 - - -
Cattle and sheep and goats 51.9 13.1 6.5 9.9
Other livestock 2.4 85.1 59.6 53.4
Fish 6.8 29.4 15.4 18.3
Mining products 152, 5 - - -
Transport & telecom. 68.6 14.5 0.3 -5.1
Other traded services 21.2 -8.3 -27.0 ' -22.1
Total 479.9 4.3 0.3 -0.2

 

Source: Author

206

 

Appendix 6.10: Final and Intermediate Demand Prices (Percents changg from base)

 

 

 

Item BASE CET BAN GINV

Commodities Consumer Prices (Percent change from base)
Coarse grains 0.7 -0.3 -3.3 1.8
Other food 0.9 -6.4 -7.1 -8.6
Rice 0.9 -3 .4 -3.0 1.0
Other industrial. ag products 2.1 -3.7 -5.8 15.0
Cattle and sheep and goats 1.0 -l6.2 -11.1 -15.5
Other livestock 2.4 -30.7 -27.8 -23.6
Fish 1.4 -15.0 -12.0 -1 1.9
Forestry & gathering products 0.8 -2.3 -4.0 0.2
Agro-industry and beverages 1.9 27.6 19.8 1 1.7
Textile products 1.9 -15.8 -1.1 -5.2
Other manufacturing goods 1.2 17.7 -4.5 12.2
Electricity water and energy 1.0 18.4 20.8 -58.9
Construction & public works 0.6 9.9 -0.1 3.3
Trade 0.4 7.9 0.4 1.0
Transport & telecom. 0.9 -10.9 -1.7 7.2
Other traded services 0.3 21.1 48.4 42.0
Bank and insurance services 0.5 2.1 4.9 5.9
Nontraded services 1.0 -5 .4 1.2 -1.4

Activities Price of Aggregate Intermediate Inputs
Food production, excl. rice 0.9 2.7 -2.0 0.1
Rice production 1.1 0.9 -0.4 1.4
Industrial ag, excluding cotton 1.1 7.4 -1.6 3.0
Cotton production 1.0 l 1.4 1.1 3.3
Livestock keeping and fishery 1.1 -4.4 -5.8 -9.5
Forestry and gathering 1.1 12.9 -1.9 3.8
Mining 0.9 11.5 -0.3 0.6
Agra-industry and beverages 1.2 10.0 6.4 6.2
Textile production 1.0 3.0 1.6 -2.6
Other manufacturing goods 1.0 10.9 0.9 3.1
Electricity, water and energy 0.8 8.0 3.7 -4.6
Construction and public works 1.0 9.9 -0.1 3.3
Trade 0.8 4.8 4.8 3.3
Transport and telecom 0.8 8.1 3.1 -0.7
Other traded services 0.9 0.8 3.0 -11.7
Bank and insurance services 0.8 2.1 4.9 5.9
Nontraded services 0.9 5.9 1.6 -O.9

 

Source: Author

207

BIBLIOGRAPHY

AFRISTAT (Observatoire Statistique et Economique de l’Afiique Subsaharienne).
2001. “Evolution des Prix dans les Etats Membres d’Afristat de 1970 a 1999”.
Series Etudes 4 (February 2001). Bamako: AFRISTAT.

Anand, S. and M. Ravallion. 1993. “Human Development in Poor Countries: On the Role
of Private Incomes and Public Services”. Journal of Economic Perspectives 7 (1):
13 3-1 50.

Anderson, J. E. 1997. “The Uruguay Round and Welfare Analysis in Some Distorted
Agricultural Economics”. Mimeo, Boston College and National Bureau of
Economic Research, Boston, MA.

Anderson, J. E. and W. Martin, 1996. “The Welfare Analysis of Fiscal Policy: A Simple
Unified Account” Boston College Working Papers in Economics 316. Boston
College Department of Economics

Anderson, J. E. and J. P. Neary. 1996. “A New Approach to Evaluating Trade Policy”.
Review of Economic Studies 63 (1996): 107-125.

Anderson, J. E. and J. P. Neary. 2000. “The Mercantilist Index of Trade Policy”. NBER
Working Paper Series 6870. Cambridge, MA: National Bureau of Economic
Research.

Armington, RS. 1969. “A Theory of Demand for Products Distinguished by Place of
Production” IMF Stafl Papers 16: 159-78.

Bacharach, M. 1970. “Biproportional Matrices and Input-Output Change”. Cambridge:
Cambridge University Press.

Bamou, E. 1998. “Libéralisation Commerciale et Performances Economiques du
Cameroun et du Gabon: Une Analyse Comparative a l’Aide d’un MCEG” .
Mimeo, Faculté des Sciences Economiques et de Gestion, Université de Yaounde,
Cameroun.

Barrett, C. “Food Security and Food Assistance Programs”. 2002. In B. Gardner and G.
Rausser (eds), Handbook of Agricultural Economics, Amsterdam: North Holland
Press.

Baustista, R. and M. Thomas. 1997. “Income Effect of Alternative Trade Policy
Adjustments on Philippine Rural Households: A General Equilibrium Analysis”.
Trade and Macroeconomic Division Discussion Paper 22, Washington DC:
International Food Policy Research Institute.

208

BCEAO (Banque Centrale des Etats de I’Afi'ique de l’Ouest). 2001. “Notes
d’Information et Statistiques”, No. 515 (June 2001), Dakar: BCEAO.

Beghin, J. C., J .-C. Bureau and S. J. Park. 2003. Food Security and Agricultural
Protection in South Korea”. American Journal of Agricultural Economics 85 (3)
(August 2003): 618-632.

Bouis, H. 2000. “Development of Modern Varieties of Rice: Impacts on Food Security
and Poverty”. Mimeo, International Food Policy Research Institute (IFPRI).
Washington DC: IFPRI.

Bourguignon, F., J. de Melo, and A. Suwa. 1991. “Modeling the Effects of Adjustment
Programs on Income Distribution”, World Development 19 (11): 1527—1544.

Bourguignon, F., A—S. Robilliard, and S. Robinson. 2003. “Representative Versus Real
Households in the Macro-Economic Modeling of Inequality”. DELTA Working
Papers 2003-05. Paris: Département et Laboratoire d’Economie Théorique et
Appliquée.

Brenton, P. 2003. “Integrating the Least Developed Countries into the World Trading
System: The Current Impact of EU Preferences under Everything But Arms”,
Working Paper no. 3018, International Trade Department. Washington DC.: The
World Bank.

Camara, O. 2004. “The Impact of Seasonal Changes in Real Incomes and Relative Prices
on Households’ Consumption Patterns in Bamako, Mali”. PhD Dissertation,
Michigan State University, East Lansing, Michigan.

Charles, J ., A. Longrigg and A. Tugend. 2001. “The Agreements on Agriculture: An
Impact Assessment”. Consumers International. Online,

www. agribusinessaccountability.org/pdfs/1 52_
Agreement%200n%20Agriculture%201mpact%2OAssessment.pdf, October 2003.

Charmes, J. 1994. “Enquéte Malienne de Conjoncture Economique et Sociale : Un
Instrument de Suivi des Groupes Vulnerable et des Principaux Indicateurs et
Sociaux en Période d’Ajustement Structurel”. Bamako : Direction Nationale de la
Statistique et de l’Information.

Cline, W. R. 2004. “Trade Policy and Global Poverty”. Washington DC: Institute of
International Economics.

Decaluwé, B., Y. Dissou and A. Patry. 1998. “Union Douaniere au Sein de l’UEMOA:
Une Analyse Quantitative”. CREFA Paper (June 1998), Montreal: Department of
Economics, Laval University, Quebec, Canada.

209

Decaluwé, B., Y. Dissou and V. Robichaud. 2000. “Regionalization and Labour Market
Rigidities in Developing Countries: A CGE Analysis of UEMOA”. CREFA
Paper, Montreal: Department of Economics, Laval University, Quebec, Canada.

Decaluwé, B. and A. Martens. 1988. “CGE Modeling and Developing Economics: A
Concise Empirical Survey of 73 Applications to 26 Countries”. Journal of Policy
Modeling 10(4): 529-568.

Dembélé, N. N., J. M. Staatz, and M. T. Weber. 2003. “Impact of the Malian Cereals
Market Reform Program on Farmers”. Policy Synthesis 68 (May 2003). East
Lansing: Michigan State University. Online,
www.msu.aec.edu/agecon/stlpsynindxhtm, September 24, 2004.

Devarajan, S., J. D. Lewis, and S. Robinson. 1990. “Policy Lessons from Trade-Focused,
Two-Sector Models”. Journal of Policy Modeling 12(4): 625-657.

Devarajan, S. and S. Robinson. 2002. “The Influence of Computable General Equilibrium
Models on Policy”. International Food Policy Research Institute’s T MD
Discussion Paper 98 (August 2002). Washington DC: IFPRI

Dervis, K., J. de Melo, and S. Robinson. 1982. “General equilibrium models for
development policy” . Cambridge: Cambridge University Press.

Diagana, B., F. Akindés, K. Savadogo, T. Reardon and J. Staatz. 1999. “Effects of the
CF A Franc Devaluation on Urban Food Consumption in West Africa: Overview
and Cross-Country Comparisons”. Food Policy 24(5): 465-478.

Diakite, L. and A. Samake. 2002. “Composition, Volume et Flux des Produits
Alirnentaires Echangés Entre le Mali et ses Voisins”. Rapport F ondation
SADAOC (October 2002). Bamako: Institut d’Economie Rurale.

Diaz-Bonilla, E., M. Thomas, S. Robinson and A. Cattaneo. 2000. “Food Security and
Trade Negotiations in the World Trade Organization: A Cluster Analysis of
Country Groups”. Trade and Macroeconomics Discussion Paper 59, Washington
DC: International Food Policy Research Institute.

Dixit, A. and V. Norman. 1980. “Theory of Intemational Trade: A Dual, General
Equilibrium Approch”. Cambridge: Cambridge University Press.

Dorosh, P. A. and D. E. Sahn. 2000. “A General Equilibrirun Model of the Effect of
Macroeconomic Adjustment on Poverty in Africa”. Journal of Policy Modeling
22 (6): 753-776.

Doumbia, A. and M. Kamate. “1997 Bilan de l'Emploi 1997: Etude sur le Secteur Rural”.

Bamako : Obervatoire de l'Emploi et de la Formation, Direction Nationale de la
Statistique et de l’Information.

210

Drud, A. (Not dated). “CONOPT”. ARKI Consulting and Development, Copenhagen,
Denmark.

Food and Agriculture Organization of the United Nations (F A0). 1996. “Rome
Declaration on World Food Security and World Food Summit Plan of Action”.
World Food Summit, Rome, Italy: FAO.

Food and Agriculture Organization of the United Nations (FAO). 2001. “Some Issues
Relating to Food Security in the Context of WTO Negotiations on Agriculture”,
Geneva Round Table on Food Insecurity in the Context of WTO Negotiations on
Agriculture (July 2001), Discussion Paper No. 1, Rome, Italy: FAO.

Food and Agricultural Organization of the United Nations (FAO). 2002. “Understanding
Food Insecurity and Vulnerability”, Food Insecurity and Vulnerability
Information Monitoring System, FAO Economic and Social Division, Rome:
FAO. Online,
www.fivims.org/images/'1‘ools&Tips/'l‘ oolsandTips/EnglisthnderstandingF S_EN
.pdf

Food and Agriculture Organization (F A0). 2003. “WTO Agreements on Agriculture: The
Implementation Experience—Developing Countries”. F AO Commodities and
Trade Division, Rome: FAO.

Food and Agriculture Policy Research Institute (F APRI). 2004. “FAPRI Agricultural
Outlook 2004”. FAPRI: Iowa State University, Ames.

Frakenberger, T.R, L. Frankel, S. Ross, M. Burke, C. Cardenas, D. Clark, A. Goddard, K.
Henry, M. Middleberg, D. O’Brien, C. Perez, R. Robinson and J. Zielinski. 1995.
“House Livelihood Strategy: A Unifying Conceptual Program for CARE
Programs”. In Proceeding of the USAID Workshop on Pefi'ormance Measurement
for Food Security, Washington DC: USAID

F risch, R. 1959. “A Complete Scheme for Computing All Direct and Cross Demand
Elasticities in a Model with Many Sectors”, Econometrica 27: 177-196.

Frongillo, E. A., M. de Onis and K. M. P. Hanson. 1997. “Socioeconomic and
Demographic Factors are Associated with Worldwide Patterns of Stunting and
Wasting of Children”. Journal of Nutrition 127 (December): 2302—2309.

Helpman, E. and P. R. Krugman. 1985. “Market Structure and Foreign Trade”.
Cambridge: The MIT Press.

Hoekrnan, B., C. Michalopoulos and L. A. Winters. 2003. “Differential and More
Favorable Treatment, Reciprocity and Fuller Participation of Developing

211

 

Countries: Objectives, Instruments and Options for the WTO”, Draft. The World
Bank, CEPR and University of Sussex.

Goldin, I., O. K. Knudsen and D. Van der Mensbrugghe. 2003. “For Whom the Bell
Tolls: Incomplete Trade Liberalization and Developing Countries”. Presentation
at the International Conference Agricultural policy reform and the WTO: Where
Are We Heading?, Capri (Italy), June 23-26, 2003.

Golan, A., G. Judge and S. Robinson. 1994. “Recovering Information from Incomplete or
Partial Multisectoral Economic Data”. Review of Economics and Statistics 76:
541-549.

Golan, A., G. Judge and D. Miller. 1996. “Maximum Entropy Econometrics, Robust
Estimation with Limited Data”. John Wiley and Sons.

Ianchovichina, E., A. Mattoo, and M. Olarreaga. 2001. “Unrestricted Market Access for
Sub-Saharan Africa: How Much is it Worth and Who Pays?” World Bank Policy
Research Working Paper No. 2595, Washington DC: The World Bank.

International Monetary Fund (IMF). 1998. “Mali: Selected Statistics”. Washington DC:
IMF.

International Monetary Fund (IMF). 1999. “Mali: Selected Statistics”. Washington DC:
IMF. .

International Monetary Fund(IMF). 2002. “Mali: Selected Statistics”. Washington DC:
IMF.

International Monetary Fund (IMF). 2004. “Mali: Selected Issues and Statistical Annex”
IMF Country Report 04/10 (January 2004). Washington DC: IMF

International Monetary Fund and The World Bank. 2001. “Market Access for Developing
Countries’ Exports”. Prepared by the Staffs of the IMF and the World Bank, April
27, 2001. Online, www.imforg/extemal/np/madc/eng/O42701.pdf, December 21,
2002.

International Monetary Fund and The World Bank. 2002. “Market Access for Developing
Country Exports-Selected Issues”. Prepared by the Staffs of the IMF and the
World Bank, September 26, 2002. Online,
www.imf.org/extemal/np/pdr/ma/Z002/eng/O92602.pdf, December 21, 2002.

Kullback, S. and R. A. Leibler. 1951. “On Information and Sufficiency”, Annale
Mathématique et Statistique 4: 99-111.

212

Khan, F. C. 1998. “Household Disaggregation”. In J. F. Francois and K. A. Reinert (eds),
Applied Methods for Trade Policy Analysis: A Handbook. Cambridge: Cambridge
University Press.

Kirman, A.1992. “Whom or What does the Representative Individual Represent?”
Journal of Economic Perspectives, 6(2): 1 17-136. ‘

Krugrnan, P. R. 1981. “Intraindustry Specialization and the Gains fiom Trade”. Journal
of Political Economy 89 (1991): 959-973.

Lee, J. E. 2002. “Can Agricultural Economists Contribute to Good Public Policy”.
Choices (Fall 2002): 11-14.

Lipsey, RG., and K Lancaster. 1956. “The General Theory of the Second Best”. Review
of Economic Studies (1956) (24): 11-32.

Lloyd, P. J. 1998. “A Generalization of the Stolper-Samuelson Theorem with Diversified
Households: A Tale of Two Matrices”. CREDIT Research Paper No. 98/6.
Nottingham: Centre for Research in Economic Development and International
Trade, University of Nottingham.

Lofgren, H., R. L. Harris and S. Robinson (with assistance from M. Thomas and M. El-
Said). 2002. “A Standard Computable General Equilibrium (CGE) Model in
GAMS”, Microcomputers in Policy Research 5 (2002). Washington DC:
International Food Policy Research Institute.

Madeley, J. 2000. “Hungry for Trade: How the Poor Pay for Free Trade”. London and
New York: Zed Books.

Martin, F. 1988. “Food Security and Comparative Advantage in Senegal: A Micro-Macro
Approach” PhD Dissertation, Department of Agricultural Economics, Michigan
State University.

Mas-Colell, A., M. D. Whinston, and J. R. Green. 1995. “Microeconornic Theory”. New
York: Oxford University Press.

Mattoo, A. and A. Subramanian. 2004. "The WTO and the Poorest Countries: The Stark
Reality”. IMF Working Paper 04/ 81. Washington DC: International Monetary
Fund.

Maxwell, S. 1996. “Food Security: A Post-Modem Perspective”. Food Policy 21 (6):
1 5 5-1 70.

Mbabazi, J. 2002. “A CGE Analysis of the Short-rim Welfare Effects of Tariff
Liberalisation in Uganda”. Discussion Paper 2002/114, United Nations
University (UNU), World Institute for Development Economics Research
(W IDER).

213

McDougall, R. A and A. Elbehri. 1998. “Protection and Support Data”, In R.A.
McDougall, A. Elbehri and T.P. Truong (eds), Global Trade Assistance and
Protection: The GTAP 4 Data Base, Chapter 4, Center for Global Trade Analysis,
Purdue University.

Metzel, J ., A. Doumbia, L. Diakite, and N. A. Diarra. 1997. “Prospects for Developing
Malian Livestock Exports”. Cambridge, MA: AIRD/EAGER.

Mollard, W. 2004. “Cotton Outlook to 2008-09”. Australian Commodities 11 (1) (March
Quarter 2003): 44-46.

Njinkeu, D. and E. Barnou. 1996. “Trade and Exchange Rate Policy Options for the CFA
Countries: Simulation with a CGE Model for Cameroon”. Research Report
Submitted to The African Economic Research Consortium (AERC), Nairobi,
Kenya.

Nouve, K. and J. Staatz. 2003a. “The Food Security Debate in West Africa Following the
WTO Agreements on Agriculture”. Stafl Paper no. 2003-05, Department of
Agricultural Economics, Michigan State University, East Lansing.

Nouve, K. and J. Staatz. 2003b. “AGOA and the Latent Agricultural Supply Response in
Sub-Saharan Africa”, Stafl Paper 2003-08, Dept. of Agricultural Economics,
Michigan State University. Selected Paper for the 2003 AAEA Annual Meeting,
Montreal (July 2003).

Organization for Economic Cooperation and Development (OECD). 2003. “Agricultural
Outlook, 2003-2008: Highlights”. Paris: OECD

Organization for Economic Cooperation and Development (OECD). 2004. “Agricultural
Outlook: 2004—2013”. Paris: OECD

Osmani, S. R. 1997. “Poverty and Nutrition in South Asia”. In Nutrition and Poverty
Papers from the ACC/SCN 24th Session Symposium, Katmandu.

Partnership to Cut Hunger and Poverty in Africa. 2002. “Now is the Time: A Plan to Cut
Hunger and Poverty in Afiica”. Washingtion DC: The Partnership to Cut Hunger
and Poverty in Africa. Online,
www.africanhunger.org/files/partnership_report_final_web_may.pdf

Posner, L. 2001. “Unequal Harvest: Farmers’ Voice on International Trade and the Right
to Food”. International Center for Human Rights and Democratic Development
(ICHRDD). Online, www.ichrdd.ca/english/commdoc/publications/globalization/
unequalHarvest.html, October 2003.

214

Powell, A. A. and F. H. G. Gruen. 1968. “The Constant Elasticity of Transformation
Production Frontier and Linear Supply Systems”. International Economic Review
9: 315-328.

Pritchett, L. and L. H. Summers. 1996. “Wealthier is Healthier”. Journal of Human
Resources 31 (4): 841—868.

Robinson, 8., A. Cattaneo and M. El-Said. 2000. “Updating and Estimating a Social
Accounting Matrix Using Cross-Entropy Methods”. Trade and Macroeconomic
Division Discussion Paper 58 (August 2000). Washington DC: International
Food Policy Research Institute.

Robinson, 8., M. El-Said. 2000. “GAMS Code for Estimating a Social Accounting
Matrix (SAM) Using Cross-Entropy (CE) Methods”. Trade and Macroeconomic
Division Discussion Paper 64 (December 2000). Washington DC: International
Food Policy Research Institute.

Rosegrant, M.W., M.S. Paisner, s. Meijer and J. Witcover.’ 2001. “Global Food
Projections to 2020: Emerging Trends and Alternative Futures”. Washington
DC: International Food Policy Research Institute.

Sadoulet, E. and A. de J anvry. 1995. “Quantitative Policy Analysis”. Baltimore and
London: The John Hopkins University Press.

Sato, K. 1972. “Additive Utility Functions with Double-Log Consumer Demand
Functions”, Journal of Political Economy 80: 102-124.

Sen, A. 1981. “Poverty and Famine: An Essay on Entitlement and Deprivation”. Oxford:
Clarendon Press.

Shannon, C. E. 1948. “A Mathematical Theory of Communication”. Bell System
Technical Journal 27: 379-423.

Singare, K, T. Reardon, Y. Camara, M. Wanzala, B. Témé and O. Sanogo. 1999.
“Household Consumption Responses to the Franc CFA Devaluation: Evidence
from Urban Mali” Food Policy 24(5): 517-534.

Smith, L. C. 1998. “Can FAO’s Measure of Chronic Undemourishment Be
Strengthened?” Food Policy 23 (5): 425-445.

Smith, L. C. and L. Haddad. 2000. “Explaining Child Malnutrition in Developing

Countries: A Cross-Country Analysis”. IF PRI Research Report 111. Washington
DC: Intemational Food Policy Research Institute.

215

 

Staatz, J. M., V. C. d’Agostino and S. Sundberg. 1990. “Measuring Food Security in
Africa: Conceptual, Empirical, and Policy Issues”. American Journal of
Agricultural Economics 72 (5) (December 1990): 1311-1317.

Subbarao, K. and L. Raney. 1995. “Social Gains from Female Education: A Cross-
national Study”. Economic Development and Cultural Change 44 (1): 105-128.

Suranovic, S. 1999. “International Trade Theory and Policy Analysis”, Online
http://internationalecon.com/vl .0/index.htm1, May 2003.

Sumner, D. A. 2002. “Food Security, Trade, and Agricultural Commodity Policy”. In L.
Tweeten and S. R. Thompson (eds), Agricultural Policy for the 21“ Century,
Ames: Iowa State Press. pp. 231-244.

Theil, H. 1967. Economics of Information Theory. North-Holland.

Thurlow, J. and P. Wobst. 2003. “Poverty-focused Social Accounting Matrices for
Tanzania”, Trade and Macroeconomic Division Working Paper 112 (March
2003). Washington DC: Intemational Food Policy Research Institute.

Timmer, C. P., W. P. Falcon, and S. R. Pearson. 1983. “Food Policy Analysis”. Baltimore
MD: Johns Hopkins University Press.

Tongeren, F. and H. Meijl. 1999. “Review of Applied Models of International Trade in
Agriculture and Related Resource and Environmental Modelling”. The Hague:
Agricultural Economics Research Institute (LEI), December 1999.

Toure, A. 1995. “EMCES 1994: Analyse des Themes: Avoirs Agricoles des Menages;
Activities des Entreprises non Agricoles par les Menages”. Draft Report
(November 1995), Enquéte Malienne de Conjoncture Economique et Sociale.
Bamako : Direction Nationale de la Statistique et de l’Informatique.

Trueblood, M., and A. Somwaru (2002). ‘Trade Liberalization and the Least Developed
Countries: Modeling the EU’s Everything but Arms Initiative’, paper presented at
the 2002 Conference on Global Economic Analysis, Taipei.

Tweeten, L. 1999. “The Economics of Global Food Security”. Review of Agricultural
Economics 21 (2) (Fall/Winter 1999): 473-478.

Uebayashi, A. 2000. “OECD Agricultural Outlook and Its Baseline Process using
AGLINK Model”. Paris: Organization for Economic Cooperation and
Development (OECD), Online:
www.unece.org/stats/documents/ces/sem.44/wp.7.e.pdf, April 20, 2004.

216

United Nations Conference on Trade, Agriculture and Development (UN CTAD). 2001.
“Improving Market Access for Least Developed Countries”. UNCTAD Document
No. UNCTAD/DITC/TNCD/4, Geneva: UNCTAD.

United States Assistance for International Development-Mali(USAID-Mali). 2003.
“Country Strategy Plan: FY 2003-2012”. Bamako: USAID-Mali.

Valdés, A. and P. Konandreas. 1981. “Assessing Food Insecurity Based on National
Aggregate in Developing Countries”. In A. Valdés (ed), Food Security for
Developing Countries, Boulder, Colorado: Westview Press.

Varian, H. 1978. “Microeconomic Analysis”, Third Edition. New York: Norton.

von Braun, J ., H. Bouis, S. Kurnar, and R. Pandya-Lorch. 1992. “Improving Food
Security for the Poor: Concepts, Policies and Programs”. Washington DC:
International Food Policy Research Institute.

West African Economic and Monetary Union (W AEMU). 2002. “Nomenclature Tarifaire
et Statistique (NT S) du Tarif Extérieur Commun (TEC)”. Ouagadougou :
WAEMU. Online, www.uemoa.int/actes/2002/I‘aritExterieurCommunhtm

Wobst, P. 2001. “Structural Adjustment and Intersectoral Shifts-A Computable General
Equilibrium Analysis”. IFPRI Research Report 117, Washington DC: IFPRI.

Wobst, P. 2002. “The Impact of Domestic and Global Trade Liberalization on FiVe
Southern African Countries”. T MD Discussion Paper 92 (March 2002),
Washington DC: IF PRI.

World Bank, The. 2001. “Global Economic Prospects and the Developing Countries
2002”. Washington DC: The World Bank.

World Bank, The. 2004. “World Development Indicators Online 2004”. Washington
DC: The World Bank. Online,
http://pub1ications.worldbank.org/subscriptions/WDU, May 4, 2004.

Zellner, A. 1988. “Optimal Information Processing and Bayes Theorem”. American
Statistician 42: 278-284.

217