Global Food Projections to 2020 - International Food Policy

GLOBAL FOOD

PROJECTIONS

TO 2020

GLOBAL FOOD

PROJECTIONS

TO 2020

EMERGING TRENDS AND

ALTERNATIVE FUTURES

MARK W. ROSEGRANT

MICHAEL S. PAISNER

SIET MEIJER

JULIE WITCOVER

International Food Policy Research InstituteAugust 2001

ISBN 0-89629-640-7

Library of Congress Cataloging-in-Publication Data Available.

CIP

Copyright © 2001 International Food Policy Research Institute

All rights reserved. Sections of this report may be reproduced without the express permission of but withacknowledgment to the International Food Policy Research institute.

Figures vii

Tables ix

Foreword xv

Acknowledgments xvii

1 1

INTRODUCTION

2 3

RECENT TRENDS IN FOOD SUPPLY AND DEMAND

Child Malnutrition: Sporadic Progress 4

World and Developed-Country Trends 5

Asia 21

Latin America 30

Sub-Saharan Africa 36

West Asia and North Africa 44

3 49

THE IMPACT MODEL

Baseline Assumptions 50

Commodity Prices and Trade Policy 56

Productivity and Area Growth 56

4 58

PROJECTIONS OF GLOBAL FOOD SUPPLY AND DEMAND AND CHILD MALNUTRITION

Baseline Projections to 2020 58

Land and Water: Limiting Factors to Global Food Supply? 76

CONTENTS

v

CONTENTS

5 82

ALTERNATIVE GLOBAL SCENARIOS FOR FOOD SUPPLY, DEMAND,

TRADE, AND SECURITY

Overview of Global Scenarios 83

Optimistic and Pessimistic Scenarios 101

6 112

ALTERNATIVE REGIONAL SCENARIOS

Asian Scenarios 112

African Scenarios 118

High Meat Demand in India—Turmoil in World Markets? 130

7 136

INVESTMENT REQUIREMENTS: WHAT WILL THE COSTS BE?

Requirements for the Baseline Scenario 136

Requirements for Optimistic and Pessimistic Scenarios 141

Regional Investment Requirements 143

8 146

SUMMARY AND CONCLUSIONS

Lessons from the Baseline Scenario 146

Lessons from Alternative Scenarios 147

Lessons for Investment 150

APPENDIX A 153

COUNTRIES AND COMMODITIES INCLUDED IN THE IMPACT MODEL

APPENDIX B 157

SUPPLEMENTARY PRODUCTION, DEMAND, AND TRADE DATA

APPENDIX C 171

REGIONAL FOOD SUPPLY AND DEMAND DATA

AND ANNUAL GROWTH RATES

APPENDIX D 177

PRODUCTION, DEMAND, AND TRADE DATA BY COMMODITY,

1997 AND 2020

Notes 195

References 199

vi CONTENTS

4.1 Total cereal demand by region, 1997 and 2020 58

4.2 Share of regions in cereal demand increase, 1997–2020 59

4.3 Cereal demand composition by crop, 1997 and 2020 59

4.4 Share of food, feed, and other uses in total cereal demand of developing countries, 1997 and 1997–2020 increase 61

4.5 Cereal production and demand increases by region, 1997–2020 62

4.6 Share of area and yield increase in regional cereal production growth, 1997–2020 62

4.7 Yield growth rates by region, all cereals, 1967–82, 1982–97, and 1997–2020 63

4.8 Cereal yields by crop in developing countries, 1997 and 2020 64

4.9 Net cereal trade by region, 1997 and 2020 65

4.10 Per capita meat demand by region, 1997 and 2020 67

4.11 Net meat trade by region, 1997 and 2020 68

4.12 Roots and tubers demand by region, 1997 and 2020 69

4.13 Share of roots and tubers demand increase by region, 1997–2020 70

4.14 Increase in roots and tubers demand by crop and region, 1997–2020 70

4.15 Net roots and tubers trade by region, 1997 and 2020 71

4.16 Net soybean trade by selected countries, 1997 and 2020 72

4.17 Net trade in edible oils by region, 1997 and 2020 73

4.18 Per capita egg demand by region, 1997 and 2020 74

4.19 Per capita milk demand by region, 1997 and 2020 75

4.20 Daily per capita calorie consumption by region, 1997 and 2020 75

4.21 Number of malnourished children by region, 1997 and 2020 76

4.22 Malnourished children as a percentage of total children under five years by region, 1997 and 2020 77

5.1 Wheat prices, alternative scenarios, 1997–2020 104

5.2 Maize prices, alternative scenarios, 1997–2020 105

FIGURES

vii

5.3 Rice prices, alternative scenarios, 1997–2020 105

5.4 Other coarse grain prices, alternative scenarios, 1997–2020 106

5.5 Meat production in 2020 by region, alternative scenarios 106

5.6 Value of world cereal trade under alternative scenarios, 1997–2020 107

5.7 Net cereal trade in 2020 by region, alternative scenarios 108

5.8 Net cereal trade value in 2020 by region, alternative scenarios 108

5.9 Per capita kilocalorie availability in 2020 by region, alternative scenarios 109

5.10 Malnourished children in 2020, alternative scenarios 110

viii FIGURES

2.1 Number of malnourished children since 1970 4

2.2 Per capita cereal production and annual growth rates in developing-countryregions, 1967–1997 5

2.3 Growth rates of cereal yields and production in developed countries, 1982–97 7

2.4 Growth rates of cereals, Former Soviet Union, 1982–97 7

2.5 Growth rates of population and total cereal demand, 1967–97 8

2.6 Net cereal trade, 1967–97 9

2.7 Net cereal trade by individual cereal exporters, 1967–97 10

2.8 Net cereal trade by major importing countries, 1967–97 11

2.9 Growth rates of per capita meat demand, 1967–97 12

2.10 Per capita meat demand, 1967 and 1997 13

2.11 Poultry’s share of total meat consumption, 1967 and 1997 14

2.12 Net meat trade, 1967–97 15

2.13 Growth rates of meat production and feed demand, 1967–97 16

2.14 Growth rates of roots and tubers demand, 1967–97 18

2.15 Growth rates of soybeans and meals demand, 1967–97 19

2.16 Net trade in meals, oils, and soybeans, 1967 and 1997 20

2.17 Growth of rice, wheat, and maize production and yields, Asia, 1967–97 22

2.18 Growth rates of cereal area, Asia, 1967–97 23

2.19 Per capita cereal demand, Asia, 1967–97 26

2.20 Per capita meat demand, Asia, 1982–97 28

2.21 Poultry production growth rates, Asia, 1967–97 29

2.22 Growth rates of meat production, Latin America, 1967–97 31

2.23 Per capita cereal demand, Latin America, 1967 and 1997 32

2.24 Share of feed in total cereal demand, Latin America, 1967 and 1997 33

2.25 Per capita meat demand, Latin America, 1967–97 33

TABLES

ix

2.26 Share of country’s cereals, meat, and roots and tuber production in total Latin Americaproduction, 1967 and 1997 34

2.27 Maize production, demand, and net trade, Latin America, 1967 and 1997 34

2.28 Growth rates of cereal yields, Latin America, 1967–97 35

2.29 Growth rates of cereal production, Sub-Saharan Africa, 1967–97 38

2.30 Cereal yields in Sub-Saharan Africa as a percent of cereal yields in developing countries,1967–97 39

2.31 Growth rates of maize yields, Sub-Saharan Africa, 1967–97 40

2.32 Growth rates of production and yields of other coarse grains, Sub-Saharan Africa, 1967–97 41

2.33 Net cereal trade, Sub-Saharan Africa, 1967–97 41

2.34 Roots and tubers production and yields, Sub-Saharan Africa, 1997 42

2.35 Growth rates of roots and tubers, Sub-Saharan Africa, 1967–97 43

2.36 Cereal and roots and tubers area, Sub-Saharan Africa, 1967–97 44

2.37 Growth rates of cereal production and yields, West Asia/North Africa, 1967–97 45

2.38 Growth rates of cereal demand, West Asia/North Africa, 1967–97 47

2.39 Net cereal trade, West Asia/North Africa, 1967–97 47

2.40 Growth rates of meat demand, West Asia/North Africa, 1967–97 48

3.1 IMPACT population and gross domestic product (GDP) assumptions, 1997 and 2020 52

4.1 Per capita food demand for cereals by crop and region, 1997 and 2020 60

4.2 Feed demand for cereals by region, 1997 and 2020 61

4.3 World prices by cereal crop, 1997 and 2020 64

4.4 Net cereal trade value by region, 1997 and 2020 66

4.5 Meat demand by region, 1997 and 2020 66

4.6 Soybean supply and demand, selected countries, 1997 and 2020 71

4.7 Value of regional agricultural trade, 1997 and 2020 73

4.8 Value of agricultural trade as a percentage of total agricultural production, 1997 and 2020 73

5.1 UN medium and low population projections in 2020, by region 85

5.2 Per capita cereal and meat demand under various UN population projections in 2020, byregion 86

5.3 Cereal prices under various UN population projections in 2020 87

5.4 Cereal yields in various regions under medium and low UN population projections in 2020 87

5.5 Total per capita kilocalorie availability under UN population projections in 2020, byregion 88

5.6 Number of malnourished children under UN population projections in 2020, by region 88

5.7 Percentage of baseline yield growth rates achieved and cereal yields realized under high-and low-yield scenarios, 2020 90

5.8 Cereal prices under baseline and various yield scenarios in 2020 91

5.9 Meat prices under the baseline and low feed ratio scenarios, 1997 and 2020 94

5.10 Meat and dairy demand under the baseline and low feed ratio scenarios in 1997 and 2020 94

5.11 Cereal prices under the baseline and low feed ratio scenarios in 1997 and 2020 94

x TABLES

5.12 Total cereal demand under the baseline and low feed ratio scenarios in 1997 and 2020 95

5.13 Net cereal trade under the baseline and low feed ratio scenarios in 1997 and 2020 95

5.14 Cereal food demand under the baseline and low feed ratio scenarios in 1997 and 2020 96

5.15 World prices under the baseline and full trade liberalization scenarios in 2020 97

5.16 Net cereal trade under the baseline and full trade liberalization scenarios in 2020 97

5.17 Regional cereal production and demand under the baseline and full trade liberalizationscenarios in 2020 98

5.18 Net meat trade under the baseline and full trade liberalization scenarios in 2020 99

5.19 Regional livestock production and demand under the baseline and full trade liberalization scenarios in 2020 100

5.20 Net welfare effects of global trade liberalization for IMPACT commodities 101

5.21 Percentage changes from baseline conditions under optimistic and pessimistic scenarios 102

5.22 Per capita cereal production in 2020 by region, alternative scenarios 111

6.1 Realized annual cereal yield growth rates under the baseline and India and Chinaslow-growth scenarios, 1997–2020 113

6.2 Net commodity trade under the baseline and India and China slow-growth scenarios in 2020 114

6.3 Crop prices under the baseline and India and China slow-growth scenarios in 1997 and 2020 115

6.4 Meat production under the baseline and India and China slow-growth scenarios in 2020 115

6.5 Meat prices under the baseline and India and China slow-growth scenarios in 2020 115

6.6 Meat production under the baseline and high Asian feed ratio scenarios in 1997 and 2020 117

6.7 Meat and dairy prices under the baseline and high Asian feed ratio scenarios in 1997 and2020 117

6.8 Cereal prices under the baseline and high Asian feed ratio scenarios in 1997 and 2020 117

6.9 Maize food demand under the baseline and high Asian feed ratio scenarios in 2020 118

6.10 Countries in Sub-Saharan Africa with significant amounts of remaining high-potential arable land 120

6.11 Number of countries achieving agricultural GDP growth of 4 percent or more inSub-Saharan Africa, 1980–97 123

6.12 Production assumptions for Sub-Saharan Africa under the optimistic and pessimisticscenarios 124

6.13 GDP growth rates in Sub-Saharan Africa under the baseline, optimistic, and pessimisticscenarios, 1997–2020 124

6.14 Cereal production and growth rates in Sub-Saharan Africa under the baseline, optimistic, and pessimistic scenarios in 2020 125

6.15 Roots and tubers production in Sub-Saharan Africa under the baseline, optimistic, andpessimistic scenarios in 2020 126

6.16 Sum of cereal and roots and tubers area cropped in Sub-Saharan Africa underthe baseline, pessimistic, and optimistic scenarios, 1997 and 2020 126

6.17 Cereal yield growth rates in Sub-Saharan Africa under the baseline, pessimistic, andoptimistic scenarios, 1997–2020 127

TABLES xi

6.18 Agricultural trade balances in Sub-Saharan Africa under the baseline, pessimistic, andoptimistic scenarios in 2020 127

6.19 Kilocalorie consumption in Sub-Saharan Africa under the baseline, pessimistic, and optimistic scenarios in 2020 128

6.20 Numbers of malnourished children in Sub-Saharan Africa under the baseline, pessimistic, and optimistic scenarios in 2020 128

6.21 Livestock production and demand assumptions underlying the high India meat demand and baseline scenarios, including comparison with historical growth in China 131

6.22 Ratio of cereal feed demand to livestock production (not including milk) under the high India meat demand scenario, 1997 and 2020 132

6.23 Total and per capita meat demand in India under high India meat demand and baselinescenarios, 1997 and 2020 133

6.24 Per capita meat demand in various countries under the high India meat demand scenario, 1997 and 2020 133

6.25 Net Indian cereal and meal trade under high India meat demand and baseline scenarios, 1997 and 2020 134

6.26 Prices under the baseline and high India meat demand scenarios, 1997 and 2020 135

7.1 Total projected investments, baseline scenario, 1997–2020 138

7.2 Total projected investments under the optimistic scenario, 1997–2020 141

7.3 Total projected investments under the pessimistic scenario, 1997–2020 142

7.4 Total projected investment in Sub-Saharan Africa under the baseline, pessimistic, andoptimistic scenarios, 1997–2020 144

B.1 Net maize and wheat trade, selected countries, 1967 and 1997 157

B.2 Growth rates of demand for various cereals, 1967–97 158

B.3 Global meat trade, 1967 and 1997 159

B.4 Roots and tubers demand, 1967–97 159

B.5 Cereal yields, Sub-Saharan Africa, 1967–97 159

B.6 Income demand elasticities, 1997 and 2020 160

B.7 Crop area (or animal slaughter numbers) elasticity with respect to own-crop price, average by region 161

B.8 Crop yield elasticity with respect to own-crop price, average by region 161

B.9 Crop yield elasticity with respect to wage of labor, average by region 162

B.10 Crop yield elasticity with respect to price of capital, average by region 162

B.11 Baseline IMPACT PSE estimates, livestock products 163

B.12 Baseline IMPACT PSE estimates, cereals and roots and tubers 164

B.13 Baseline IMPACT CSE estimates, livestock products 166

B.14 Baseline IMPACT CSE estimates, cereals and roots and tubers products 168

C.1 Food supply and demand indicators, East Asia, 1967–97 171

C.2 Annual growth rates in East Asia, 1967–97 171

C.3 Food supply and demand indicators, South Asia, 1967–97 172

C.4 Annual growth rates in South Asia, 1967–97 172

C.5 Food supply and demand indicators, Southeast Asia, 1967–97 172

C.6 Annual growth rates in Southeast Asia, 1967–97 173

xii TABLES

C.7 Food supply and demand indicators, Latin America, 1967–97 173

C.8 Annual growth rates in Latin America, 1967–97 173

C.9 Food supply and demand indicators, Sub-Saharan Africa, 1967–97 174

C.10 Annual growth rates in Sub-Saharan Africa, 1967–97 174

C.11 Food supply and demand indicators, West Asia/North Africa, 1967–97 174

C.12 Annual growth rates, West Asia/North Africa, 1967–97 175

D.1 Beef production, demand, and trade data, 1997 and 2020 177

D.2 Pork production, demand, and trade data, 1997 and 2020 178

D.3 Sheep and goat production, demand, and trade data, 1997 and 2020 178

D.4 Poultry production, demand, and trade data, 1997 and 2020 179

D.5 All meats production, demand, and trade data, 1997 and 2020 179

D.6 Wheat production, demand, and trade data, 1997 and 2020 180

D.7 Rice production, demand, and trade data, 1997 and 2020 181

D.8 Maize production, demand, and trade data, 1997 and 2020 182

D.9 Other coarse grains production, demand, and trade data, 1997 and 2020 183

D.10 All cereals production, demand, and trade data, 1997 and 2020 184

D.11 Potatoes production, demand, and trade data, 1997 and 2020 185

D.12 Sweet potatoes and yams production, demand, and trade data, 1997 and 2020 186

D.13 Cassava and other roots and tubers production, demand, and trade data, 1997 and 2020 187

D.14 All roots and tubers production, demand, and trade data, 1997 and 2020 188

D.15 Eggs production, demand, and trade data, 1997 and 2020 189

D.16 Milk production, demand, and trade data, 1997 and 2020 190

D.17 Meals production, demand, and trade data, 1997 and 2020 191

D.18 Oils production, demand, and trade data, 1997 and 2020 192

D.19 Soybeans production, demand, and trade data, 1997 and 2020 193

TABLES xiii

Over the years, it has become more and more apparent that any meaningful discussion of poli-cies and priorities for alleviating malnutrition and poverty and achieving economic growth dependson an accurate assessment of where we have been and where we are headed. But forecasting thefuture of food supply and demand requires more than just an examination of short-term trendsin global markets: it is essential to focus on long-term growth in income, population, agriculturaltechnology, and a host of other pressing potential changes.

IFPRI presented the first projections of global food supply and demand based on its IMPACTmodel in 1995 at its Washington, D.C., conference on a 2020 Vision for Food, Agriculture, andthe Environment. The 2020 discussion paper that presented those first projections stressed thatthe prospects for global food security greatly depend not only on population growth and eco-nomic development but also on emerging issues such as trade liberalization, urbanization, envi-ronmental degradation, water scarcity, the livestock revolution, and new technologies. Two yearslater, IFPRI updated its take on the global food situation, extending its baseline scenario forward,and in 1999 further updated it, adding commodities and again extending the baseline. Now, in2001, on the eve of another 2020 conference—”Sustainable Food Security for All”—to be heldSeptember 2001 in Bonn, Germany, IFPRI has once again fine-tuned its IMPACT model.

In this volume, which reports the results of IFPRI’s projection work in far more detail thanprevious publications, the authors give their best assessment of what the future food situationwill be in the baseline scenario. Then they examine the effects of changes in policy, technology,and life styles through two sets of alternative scenarios. One set explores changes at the globallevel; the other is regional, focusing on changes specific to Asia and Sub-Saharan Africa.

These scenarios point to one inescapable conclusion: even rather small changes in agriculturaland development policies and investments, made in both developed and developing countries,can have wide-reaching effects on the number of poor and undernourished people around theworld. The policy choices we make now will determine to a considerable degree what kind oflives the next generation will lead.

FOREWORD

xv

To further share the key findings from this pathbreaking research, IFPRI is publishing a morepopular version of this paper as a food policy report titled 2020 Global Food Outlook: Trends, Alter-

natives, and Choices.

Per Pinstrup-AndersenDirector General, IFPRI

xvi FOREWORD

This report benefited from the input of many colleagues. They are too numerous to mentionindividually, but several colleagues stand out because of the degree of their support for this col-laborative project and the depth of their insights on previous drafts. Per Pinstrup-Andersen andRajul Pandya-Lorch, director general and head of the 2020 Vision Initiative, respectively, of theInternational Food Policy Research Institute (IFPRI), provided an institutional framework for theproject, constant encouragement, and detailed and insightful comments. Pandya-Lorch provideddetailed intellectual and editorial input throughout the process.

A number of formal and informal reviewers and readers of earlier drafts of the report greatlyimproved the final product. Particular mention should be made of the very detailed and helpfulcomments by Nikos Alexandratos and Nurul Islam on an earlier draft. Finally, the authors com-mend Phyllis Skillman and Uday Mohan for their editorial work on the manuscript, Mary-JaneBanks and Sarah Cline for additional editorial assistance, and Maria Esteban for efficiently pro-cessing many drafts of the manuscript. Remaining deficiencies are the responsibility of the authors.

ACKNOWLEDGMENTS

xvii

1

1

INTRODUCTION

The world population is expected to grow from5.8 billion people in 1997 to 7.5 billion peoplein 2020. Although these latest population pro-jections represent a slowdown from past esti-mates, such a large absolute increase in popu-lation raises serious concerns about whetherthe world’s food production system will be ableto feed so many people, especially in the faceof a possibly stagnant or even declining stockof natural resources. These concerns escalatedsharply in the mid-1990s in the face of dramaticincreases in world cereal prices in 1996, declin-ing cereal stocks, and the appearance of sev-eral widely read publications starkly depictinga starving twenty-first century world unable tomeet growing food demands from a deterio-rating natural resource base (Brown 1995;Brown and Kane 1994).

Wheat and maize prices in mid-1996 were50 percent higher than a year before, while riceprices were 20 percent above 1994 levels. Ris-ing cereal prices were accompanied by declinesin cereal stocks between 1993 and 1997, withworld stocks dropping from an average of 18percent of total annual consumption in 1993to about 13 percent in 1996, the lowest level in

recent history. Some observers said that theserising cereal prices and falling stocks were indi-cators of a new reality for world food markets,with high prices, low stocks, and continuousfood shortages. Global cereal productionresponded to higher prices, hitting record lev-els in 1997 and 1998, while falling incomescaused by the East Asian economic crisisreduced the demand for food commodities.Prices for wheat and maize fell nearly 50 per-cent between 1996 and 1999, rice pricesdropped by 24 percent, and cereal stocks hadrisen again to 18 percent of consumption by1999/2000. The policy focus in much of theworld shifted from long-term food supply anddemand problems to providing subsidies forfinancially distressed farmers.

These recent fluctuations in cereal marketsshow how inappropriate it is to make judg-ments about long-term food security based onshort-term trends in global markets. Indeed,year-to-year variability in prices and produc-tion—and the influence that this variability hason the amount and type of attention devotedto the global food situation—may in fact con-tribute to long-term food problems by encour-

2 CHAPTER 1

aging complacency during periods of strongshort-run performance. In order to understandthe future of food supply and demand and foodsecurity, it is essential instead to focus on long-term fundamental drivers, such as income andpopulation growth, and technological changesin agriculture as influenced by investments inagricultural research, irrigation, roads, andother factors. In this volume we explore alter-native futures for global food markets, includ-ing both a baseline scenario that gives our bestestimates of the future and a number of alter-native scenarios that assess the flexibility ofworld food markets and the robustness of base-line results.

We have chosen to examine alternative sce-narios because the future world situation isdependent on a number of variables, many ofwhich are the result of policy decisions oninvestment in agricultural research, irrigation,clean water, and health; population programs;and the general economic policies that driveincome growth. Through alternative scenar-ios, we explore the effects of policy, technol-ogy, and lifestyle-driven changes on global foodmarkets and the poor.

The projections of long-term food supply,

demand, trade, and prices are based on assess-ment of the underlying factors driving globalfood markets, including likely future develop-ments for wheat, maize, rice, other coarsegrains, soybeans, roots and tubers, oils, meals,and meats. After a brief review of recent his-torical trends in the global and regional foodsituation, we describe the global food projec-tions model called the International Model forPolicy Analysis of Agricultural Commoditiesand Trade (IMPACT), developed by the Inter-national Food Policy Research Institute(IFPRI). We then present an overview of thebaseline demand and supply projections,including projections of crop area harvestedand crop yields, food demand, price and tradeprojections for these commodities, and theeffects of these projections on childhoodmalnutrition. Next we explore several alterna-tive regional and global scenarios, includingoptimistic and pessimistic paths for the futureworld food situation. Finally, we considerthe implications of these projections forfuture global food security and policy. A list ofall of the countries and regions and com-modities included in the model is presented inAppendix A.

3

2

RECENT TRENDS IN FOOD

SUPPLY AND DEMAND

The world food picture has undergone dramaticchanges in the three decades since the mid-1960s, when widespread food shortages in Asiacaused predictions of disastrous recurringfamines. A Green Revolution, featuring theadoption of high-yielding cereal varieties andrapid increases in irrigated area and fertilizeruse, dramatically improved productivity in Asiaand other developing regions, easing the fear ofendemic famine in Asia. The Green Revolutionpeaked in much of Asia in the 1970s and early1980s before slowing in the 1990s. Concurrentlywith the Green Revolution, many developingcountries experienced rising incomes and shift-ing consumption patterns, which led to strikingincreases in livestock consumption and produc-tion, particularly in Asia. The past few decadeshave also brought policy shifts—such as eco-nomic liberalization—that have improved effi-ciency in agriculture and eased trade flows.Other shifts, however, including debt crises,structural adjustment programs, the disinte-gration of the Soviet Union, and the recent EastAsian economic crisis, have negatively affectedthe ability of certain regions to maintain foodsecurity. Thus, despite the fact that most regions

have made substantial inroads against povertyand averted widespread famine in recent years,malnutrition persists, remaining intransigent insome developing regions and threatening resur-gence in others.

Food demand growth caused by expandingpopulations and shifting consumption pat-terns will necessitate future food productionincreases, but unexploited, available arable landis limited, placing the burden for these increaseson technologically driven yield improvements.The need for modern agricultural technologies,however, must be balanced against legitimateconcerns about environmental sustainability.Empirical evidence has demonstrated that neg-ative effects on the environment from inap-propriately applied technologies can translateinto productivity losses and threaten humanhealth, although assessing the precise extent ofthese effects is often difficult. Growing urbanand industrial demands on existing water sup-plies and the need for improved water qualityfurther complicate the situation.

In reviewing recent historical patterns in foodsecurity and global food markets, we first lookat trends in child malnutrition, followed by a

4 CHAPTER 2

broad overview of global commodity trendsbetween the mid-1960s and late-1990s, particu-larly cereal and livestock production and con-sumption. Finally, we focus on food supply anddemand trends in each of the developing regions:Asia, Latin America, Sub-Saharan Africa, andWest Asia and North Africa (WANA). In manycases, we have further subdivided Asia into EastAsia, South Asia, and Southeast Asia to capturethe diversity of these subregions.

CHILD MALNUTRITION: SPORADIC PROGRESS

Since the 1960s, developing countries have madeimpressive strides against malnutrition ratesamong children under five,1 declining from anaggregate rate of more than 46 percent in 1970to 31 percent in 1997.2 Nevertheless, as a resultof population growth rates in developing coun-tries averaging 2.1 percent annually between1967 and 1997 (despite declines in growththroughout the period), the percentage declinein child malnutrition has translated into anabsolute decline of 20 million malnourishedchildren since 1967 to approximately 167 mil-lion children in 1997.3

These aggregate declines mask alarmingregional trends (Table 2.1).4 In South Asia, child

malnutrition prevalence rates dropped from 72percent to just below 50 percent over this 30-year period. Despite these encouraging signs,South Asia’s progress against malnutrition hasbeen far from steady: while the region’s num-ber of malnourished children declined between1970 and 1980, it rose sharply in 1985. Improve-ments between 1990 and 1997, however, offsetthis backslide, and the number of South Asianmalnourished children in 1997 was 7.2 millionbelow 1970 levels.

Sub-Saharan Africa had many more mal-nourished children in 1997 than it did in the mid-1970s (prevalence rates also rose between 1985and 1997). While roughly 1 in 10 malnourishedchildren in developing countries resided in Sub-Saharan Africa in 1970, 1 in 5 did by the mid-1990s. Between 1985 and 1997, WANA alsoexperienced a worsening of child malnutrition:the ranks of the malnourished increased byapproximately 1 million children, more thanerasing gains made in the previous decade anda half.

The overall picture in Latin America has beenrelatively bright, with child malnutrition preva-lence dropping from 21 to 10 percent between1970 and 1997. Declines in child malnutritionwere more impressive (and malnutrition preva-

TABLE 2.1 Number of malnourished children since 1970

Region 1970 1975 1980 1985 1990 1995 1997

(millions of children under age 5)

Latin America and the Caribbean 9.5 8.2 6.2 5.7 6.2 5.2 5.1

Sub-Saharan Africa 18.5 18.5 19.9 24.1 25.7 31.4 32.7

West Asia/North Africa 5.9 5.2 5.0 5.0 n.a. 6.3 5.9

South Asia 92.2 90.6 89.9 100.1 95.4 86.0 85.0

East Asia 77.6 45.1 43.3 42.8 42.5 38.2 37.6

All regions 203.8 167.6 164.3 177.7 176.7 167.1 166.3

Source: Smith and Haddad (2000) 1970 through 1995. 1997 data are the IMPACT base-year valuesextrapolated from 1995 values using the IMPACT model.Note: n.a. is not available.

RECENT TRENDS IN FOOD SUPPLY AND DEMAND 5

lence rates among children lower) in South

America than in Central America and the

Caribbean (Garrett 1997). In East Asia, the

absolute number of malnourished children fell

dramatically between 1970 and 1997.

Although a downward trend is evident in the

number of malnourished children in develop-

ing countries, this trend does not demonstrate

a pattern of inevitable, steady progress. The

timing and size of gains have been uneven

and interspersed with periods of worsening or

stagnant malnutrition. The largest declines

occurred in Asia during the 1970s. Sub-Saharan

Africa was stable between 1970 and 1975,

steadily increasing thereafter; improvements in

WANA have emerged only recently, and

progress in Latin America slowed during the

late 1980s. Nevertheless, caloric availability per

capita rose in developing countries between the

1960s and the early 1990s by 400 kilocalories,

reaching nearly 2,700 kilocalories per day by

1997 (FAO 2000a).

WORLD AND DEVELOPED-COUNTRY TRENDS

Cereals

Over the 30-year period 1967–97, per capitacereal production worldwide rose substantiallyin the context of rapidly increasing cereal yields,slow growth in total harvested area, and declin-ing per capita harvested area. While per capitaproduction in the developed world rose from565 kilograms in 1967 to 660 kilograms in 1997,per capita cereal production in the developingworld rose from 176 kilograms in 1967 to 226kilograms in 1997, an increase of 28 percent(Table 2.2). (All values throughout this volumeare three-year averages centered on the identi-fied year, based on data downloaded from theFAOSTAT database of the Food and Agricul-ture Organization of the United Nations [FAO2000a]). Thanks to the Green Revolution, whichwas sparked by the development of high-yielding varieties of cereals and complementedby expansion of irrigated area and fertilizer use,these production increases in per capita pro-duction occurred despite an increase in total

TABLE 2.2 Per capita cereal production and annual growth rates in developing-country regions,

1967–97

Per capita cereal production Production growth rates

Region 1967 1982 1990 1997 1967–82 1982–90 1990–97

(kilograms/capita) (percent/year)

Latin America 225.3 262.0 222.1 253.4 1.0 −2.0 1.9

Sub-Saharan Africa 127.9 110.8 122.3 124.6 −1.0 1.2 0.3

West Asia/North Africa 255.8 231.5 245.5 245.6 −0.7 0.7 0.0

All Asia 163.6 206.9 224.4 236.4 1.6 1.0 0.7

South Asia 146.0 171.3 182.1 182.6 1.1 0.8 0.0

Southeast Asia 157.8 198.8 210.1 226.3 1.6 0.7 1.1

East Asia 188.7 248.7 276.5 295.8 1.9 1.3 1.0

Developed world 564.6 670.4 680.3 660.1 1.2 0.2 −0.4

Developing world 176.0 206.8 216.0 225.6 1.1 0.5 0.6

––––Source: Based on FAOSTAT data (FAO 2000a).

6 CHAPTER 2

harvested cereal area of 26 million hectaresworldwide between 1967 and 1997 and a declinein global per capita harvested area from 0.19hectares in 1967 to 0.12 hectares in 1997. At theGreen Revolution’s peak, worldwide cerealproduction moved sharply upward, especiallyin Asian developing countries, but also morebroadly in developing countries, with Sub-Saharan Africa lagging badly.

Cereal Production Trends. Worldwide cereal yieldgrowth has slowed since its peak in the mid-1970s, with cereal production growth in devel-oping countries slowing significantly and cerealproduction in developed countries practicallystagnating until the late 1990s. Recent trends indeveloping countries will be discussed further on.In those developed countries that produce a largeshare of cereal exports, dramatic declines in pro-duction growth were mostly the result of policychoices that led to area declines averaging 1.2 per-cent annually between 1982 and 1997. Trends inthe developed world are heavily influenced byagricultural subsidy policies in the EuropeanUnion countries (EU15),5 the United States, andCanada. As a result of high and increasing pro-ducer subsidies in the 1980s, cereal productionexpanded; stocks rose from an average of 20 per-cent of production during the late 1970s and early1980s to 27 percent by 1986–87. Subsidy policies,among other factors, caused international cerealprices to fall by 40 percent between 1981 and1987. As a result, cereal exporters responded bycutting subsidies and withdrawing land from pro-duction (Dyson 1996).

In the United States, producer subsidyequivalents (PSEs) declined from 26 percent to19 percent of production between 1987 and1990 but fell little between 1990 and 1997 (seeBox 3.1, p. 56). Consumer subsidy equivalents(CSEs) fell even more rapidly between 1987and 1990, from 8 percent to 3 percent and wereactually slightly positive by 1997. Reforms to

Canadian agriculture were slower to be imple-mented, but PSE estimates fell from 32 percentin 1990 to 15 percent in 1997 and CSEs fell from−21 in 1990 to −14 percent in 1997. In the Euro-pean Union, PSEs fell slightly, from 46 percentof production in 1987 to 39 percent of pro-duction in 1997; CSEs in the EU15 fell more—from 42 percent of production in 1987 to 25percent of production in 1997—with most ofthis reduction occurring between 1990 and1997 (OECD 1999). In addition to reducing themagnitude of subsidies, North American and European governments began to decou-ple subsidies from directly influencing farm production decisions by shifting from farm-price support programs to direct payments tofarmers, thus reducing the need to buy andhold large reserves. In 1996, the United Statesand the European Union together held lessthan one-half the cereal stocks they held in1993.

Cereal production grew slowly in the devel-oped countries during 1982–90, reflecting thenet effects of rising subsidies early on andsome subsidy and area reductions later in theperiod. Production growth stagnated during1990–97 except for Australia (Table 2.3),reflecting sharp reductions in the size andchanges in the form of farm subsidies, whichin turn led to a reduction in cereal area duringthe period. For the developed countries (otherthan the Former Soviet Union [FSU] and East-ern Europe), the policy-driven nature of theproduction slowdown, while yields continueto grow at a positive rate, implies that signifi-cant supply remains untapped. It is not sur-prising that cereal production growth returnedto these countries during the mid-1990s, whenincentives to increase production—in the formof high international cereal prices—alsoreturned.

Events that took place in the FSU countriesin the post-reform period form an important


component of the structure of world cerealdemand and production trends during the1990s, with production, demand, and totalimports all falling drastically in the region,while total exports rose (Table 2.4). Cereal pro-duction reversals in FSU probably began in 1988but became apparent only during the period1990–97, when production fell 5.3 percentannually and yields by 3.1 percent annually.

The reform process in the former Sovietbloc countries had drastically negative effectson agricultural production and yields for bothterms-of-trade and institutional reasons.Macours and Swinnen (2000a, 2000b) estimatethat price liberalization and subsidy reductionscaused a significant decline in the terms oftrade for FSU agriculture and have been

responsible for 40 to 50 percent of the fall inaverage crop output during the 1990s. An addi-tional 30 to 60 percent of the agricultural out-put decline was the result of institutional dis-ruptions to both the agricultural sector and thelarger economy, as the absence of contractenforcement mechanisms and information dis-tribution systems severely impeded investmentand growth in an increasingly liberalized pro-duction environment (Macours and Swinnen2000a, 2000b). The poor performance of theFSU agricultural sector in the post-reformperiod, while not as dismal as that of the econ-omy as a whole, became a major political issue;it prompted some retrenchment on agricul-tural trade liberalization in the mid-1990s in an

TABLE 2.3 Growth rates of cereal yields and production in developed

countries, 1982–97

Yield Production

Region/Country 1982–90 1990–97 1982–90 1990–97

(percent/year)

United States 1.5 2.3 0.1 2.2

EU15 2.8 2.1 1.9 1.1

Japan 1.0 0.7 0.1 −1.2

Australia 3.0 2.5 −1.0 6.0

Eastern Europe 1.3 −1.8 0.8 −2.1

Former Soviet Union 3.2 −3.1 1.6 −5.3

Other developed countries 1.0 1.6 0.8 0.0


TABLE 2.4 Growth rates of cereals, Former Soviet Union,

1982–97

Production Yields Demand Area per capita

(percent/year)

1982–90 1.6 3.2 1.4 −2.4

1990–97 −5.3 −3.0 −8.2 −2.5


effort to maintain food self-sufficiency (vonBraun et al. 1996).

Cereal Demand. Income and population growthhave driven expansion of cereal demand world-wide. The gradual slowing in populationgrowth rates over recent decades has resultedin a corresponding slowdown in the rate ofgrowth in cereal demand (Table 2.5). Cerealdemand growth in the developing worlddeclined from 3.8 percent annually in 1967–82to 3.2 percent in 1982–90, and to 2.7 percent in1990–97. In reality, however, such aggregategrowth rates obscure as much as they reveal;events in Asia have driven the overall demandgrowth decline, since that region accounts for54 percent of the world population and domi-nates trends at the global level. Demandgrowth has also slowed in other regions forwidely divergent reasons; thus, while satura-tion of consumer demand in most of the devel-oped countries and the collapse of the FSUeconomies after 1990 helped slow demandgrowth in the developed world, poverty anda scarcity of foreign exchange restrained de-

mand growth in Sub-Saharan Africa, despitepersistently high malnutrition in the region.

Cereal Trade. The dominant trend affectingworld cereal markets has been divergencebetween production and demand growth in anumber of regions, thus resulting in a dramaticincrease in world cereal trade from about 116million metric tons in 1967 to 257 million tonsin 1997. Most of this increase in cereal tradetook place between 1967 and 1982, when totalcereal trade expanded at a rate of 5.1 percentannually. Growth slowed between 1982 and1990 to 0.5 percent annually but picked up toa rate of 1.0 percent between 1990 and 1997.The growth in cereal trade has been influencedby rapid economic growth in developing coun-tries, as well as much improved communica-tion and transport capacities, improving tradeand macroeconomic policies, and changingpatterns of food demand (Dyson 1996). Totalworld cereal trade increased from 230.3 mil-lion tons in 1982 to 239.9 million tons in 1990.The composition of trade changed substan-tially during this period, however, as subsidy

8 CHAPTER 2

TABLE 2.5 Growth rates of population and total cereal demand, 1967–97

Population growth Total cereal demand

Region 1967–82 1982–90 1990–97 1967–82 1982–90 1990–97

(percent/year) (percent/year)

Latin America 2.4 2.0 1.7 4.0 1.7 3.6

Sub-Saharan Africa 2.8 2.9 2.6 2.8 4.1 4.0

West Asia/North Africa 2.8 3.0 2.2 4.3 4.1 2.4

Asia 2.2 1.9 1.6 3.8 3.2 2.6

South Asia 2.3 2.2 1.9 2.9 3.4 2.6

Southeast Asia 2.4 2.0 1.7 2.8 3.7 3.4

East Asia 2.0 1.5 1.1 4.3 3.0 2.4

Developing world 2.3 2.1 1.7 1.9 0.8 −1.3

Developed world 0.8 0.7 0.5 3.8 3.2 2.7


increases and resulting low cereal prices hadvarying effects on the major cereal exportingnations. Table 2.6 gives a breakdown of netcereal trade by region.

The EU15 countries benefited from the sub-sidy war, rapid technological progress in agri-culture, and the saturation of domestic con-sumption, shifting from net cereal imports of1.8 million tons in 1982 to net exports of 29.1million tons in 1990, as production growth sig-nificantly exceeded consumption growth(Koester and Tangermann 1990) (Table 2.7).The emergence of the European Union as amajor exporter certainly played an importantrole in keeping international cereal prices lowduring the early to mid-1980s, although a num-ber of other factors were at play as well, includ-ing the adoption of deflationary monetary pol-icy by many industrialized countries followingthe second oil crisis, worldwide economicrecession between 1981 and 1983, and macro-economic crises in a number of developingcountries during the early 1980s. Cereal pricesbegan recovering after 1986.

Despite the declining international priceenvironment, Viet Nam and India also man-aged to become net cereal exporters between1982 and 1990 (Table 2.7). Domestic policiesin both Asian nations limited cereal importsand favored the expansion of cereal exports, inpart as a result of policies that depressed con-sumer demand for food (Rosegrant and Hazell2000). Meanwhile, the United States—partlydue to EU15 competition and partly due toits own subsidy reductions—experienced adecline in net exports (Sanderson and Mehra1990).

Argentina also had a difficult decade, with netcereal exports falling by 48 percent between 1982and 1990, partly because the economic declineof the Soviet Union in the late 1980s softeneddemand for Argentina’s agricultural commodi-ties, while economic crises throughout LatinAmerica weakened regional demand. Moreover,in a high-subsidy environment, Argentina—inthe throes of economic crisis—simply could notafford to compete with developed world pro-ducers. Poor domestic macroeconomic and sec-toral policies and deficiencies in transport and


TABLE 2.6 Net cereal trade, 1967–97

Region 1967 1982 1990 1997

(million metric tons)

Latin America 3.1 −3.5 −11.4 −14.5

Sub-Saharan Africa −1.5 −8.3 −8.1 −12.0

West Asia/North Africa −5.9 −28.9 −38.7 −44.3

Asia −17.4 −28.0 −29.4 −28.0

South Asia −11.6 −2.9 −3.2 −1.7

Southeast Asia −0.1 0.8 0.1 −5.5

East Asia −5.8 −25.9 −26.3 −20.9

Developing world −21.7 −68.7 −87.6 −98.8

Developed world 24.6 73.8 93.2 105.9

––––Source: Based on FAOSTAT data (FAO 2000a).Note: Positive figures indicate exports; negative figures indicate imports.

marketing hurt Argentina more than other coun-tries in the low-price environment of the late1980s (Diaz-Bonilla 1999).

The 1990s saw a recovery of world cerealprices and the general erosion of the dominanceestablished by the United States and the EU15 inworld cereal markets during the 1980s. Totalworld cereal trade rose by 20 million tonsbetween 1990 and 1997, while net cereal exportsfrom the United States declined from 94 milliontons in 1990 to 78 million tons in 1997, and sig-nificant reforms to the Common AgriculturalPolicy (CAP) in the EU15 helped reduce exportsthere. Other traditional exporters gained at theexpense of the two leaders. Argentina, benefit-ing from the overall economic recovery of LatinAmerica and from expanded regional markets asthe result of MERCOSUR—a common marketagreement between Argentina, Brazil, Paraguay,and Uruguay—saw an increase in 1997 of 103percent above 1990 levels. Australia also increasedits net exports 32 percent.

Table 2.8 shows net cereal import trends forselected developing countries in more detail.Between 1967 and 1982, Japan, the otherWANA countries (excluding Egypt), and China

all emerged as major players in world cerealmarkets. However, this trend of rapidly increas-ing cereal imports had slowed by the end of the1980s. Chinese net cereal imports actuallydeclined by 3.5 million tons in 1990. Most othermajor importing regions and countries havenot shown such dramatic fluctuations: their lev-els of net imports increased slowly but steadilythroughout the 1980s and 1990s.

Growing wheat and maize import demandby developing countries has clearly been themajor source of growth for overall cereal trade(see Appendix B, Tables B.1 and B.2). World-wide wheat trade rose from 63 million tons in1967 to 122 million tons in 1997, and world-wide maize trade rose from 27 million tons in1967 to 75 million tons in 1997. While somecountries—including Argentina, China, India,and Pakistan—matched their high growth inwheat demand with rapid production growth,most developing countries experienced grow-ing import dependence (FAO 2000b). Feed useswere the dominant factor behind risingdemand for maize and other grains (in-cluding barley, millet, oats, rye, and sorghum),although the growing importance of compet-

10 CHAPTER 2

TABLE 2.7 Net cereal trade by individual cereal exporters, 1967–97

Region/Country 1967 1982 1990 1997


United States 43.5 104.1 93.6 76.8

EU15 −24.2 −1.8 29.1 20.0

Australia 7.3 13.1 14.9 21.7

Other developed countries 2.6 4.5 4.7 6.2

Argentina 8.1 18.5 9.7 19.4

India −9.1 −1.3 0.4 1.8

Thailand 2.7 6.2 5.8 4.5

Viet Nam −1.5 −0.6 1.2 2.8

––––Source: Based on FAOSTAT data (FAO 2000a).Note: Positive figures indicate net exports; negative figures indicate net imports.

ing feed sources such as oil crops, meals, andcassava, and the overall shift of meat produc-tion toward poultry, which has higher feedingefficiencies than other livestock, sloweddemand for these crops. The consumption ofother coarse grains increasingly shifted awayfrom food to feed uses, although some are stillimportant food crops in a number of poordeveloping countries in Sub-Saharan Africaand Latin America.

World Cereal Prices. Between 1982 and 1997,real world wheat prices declined by 28 percent,rice prices by 29 percent, and maize prices by30 percent. These trends combined to mitigatethe effects of higher cereal import demand onworld cereal prices in many developing coun-

tries. India, a significant net cereal importeralways on the verge of a food crisis during the1960s and 1970s, actually became a netexporter of cereals by the 1990s. And China,consistently defying predictions of impendingdisaster, experienced only modest increases incereal imports, with net imports standing atonly 7.6 million tons in 1997 (Table 2.8). Impor-tant questions remain as to the sustainabilityof these two trends, however. The high per-sistence of poverty in India has dampenedcereal demand growth there. And highly pub-licized degradation of the natural resourcebase in both countries may have serious impli-cations for future production growth. Never-theless, it is undeniable that the two most pop-ulous countries in the world have hadremarkable success in maintaining a high level


TABLE 2.8 Net cereal trade by major importing countries, 1967–97

Region/Country 1967 1982 1990 1997


Asia –18.9 –27.5 –30.2 –31.0

China −4.6 −19.1 −15.6 −7.6

Indonesia −0.6 −2.3 −2.1 −5.7

Japan −11.8 −24.7 −27.9 −27.8

Korea, Republic of −0.9 −6.4 −9.8 −12.0

Malaysia −0.9 −1.8 −2.8 −3.8

Philippines −0.8 −1.3 −2.2 −4.0

Latin America 2.7 –3.2 –10.6 –15.3

Brazil −1.8 −4.7 −4.4 −9.3

Colombia −0.3 −0.8 −0.8 −3.4

Mexico 1.3 −6.4 −6.6 −9.5

Other Latin American countries, −4.2 −10.1 −9.3 −12.0

excluding Argentina

West Asian/North African –6.3 –29.2 –39.7 –45.1

Egypt −2.0 −7.1 −7.9 −9.4

Other West Asian/North African −3.8 −22.6 −30.3 −33.8

countries, excluding Turkey


of cereal self-sufficiency under very difficult cir-cumstances. Another important developmentthat helped to keep prices low during the 1980sis the subsidization of Western European agri-culture, which led to the emergence of thatregion as a major net cereal exporter. WithoutWestern Europe’s remarkable turnaround fromlarge net importer to significant net exporter,the strain on traditional cereal exporters of sup-plying the growing import demands of thedeveloping world might have put strong upwardpressure on cereal prices (McDowell 1991).

Livestock

Demand: A Livestock Revolution? Some debatehas emerged over whether the term “revolu-tion” applies legitimately to the tremendousgrowth in livestock consumption that hasoccurred worldwide in recent decades (Delgadoet al. 1999; FAO 2000b). The analysis is some-what complicated by the fact that meat con-sumption has historically been concentrated dis-proportionately in industrialized countries,

where per capita consumption growth tends tobe slow because meat consumption is alreadyat such a high level. This slow growth in devel-oped countries has served to counteract theeffects of rapid growth in developing countries,led by Brazil and China. While developed coun-tries accounted for 30 percent of the world’spopulation and 71 percent of its meat con-sumption in 1967, they accounted for only 22percent of population and 47 percent of meatconsumption in 1997. Thus world per capitameat consumption appears at first glance tohave risen more slowly than the term “revolu-tion” would imply, with the fastest growth ratesachieved during 1982–90 at 1.5 percent annu-ally, slowing to 1.4 percent during 1990–97(Table 2.9).

Despite this apparent slowdown, severaltrends appear to justify the notion of a “revo-lution.” First, the magnitude of changes occur-ring in the developing world undoubtedly indi-cates an extraordinary change in the diet of theemerging middle class in developing countries

12 CHAPTER 2

TABLE 2.9 Growth rates of per capita meat demand,

1967–97

Region/Country 1967–82 1982–90 1990–97

(percent/year)

Former Soviet Union 1.8 2.2 −8.5

Eastern Europe 2.1 1.3 −2.4

United States 0.5 0.8 0.9

EU15 1.9 0.9 0.2

Latin America 1.4 0.9 3.5

Sub-Saharan Africa 0.2 −0.8 0.2

West Asia/North Africa 3.4 0.0 0.9

Developing Asia 2.4 6.0 7.1

Developing world 2.0 3.4 5.2

Developed world 1.5 1.2 −1.1

World 1.1 1.5 1.4


(Delgado et al. 1999). While per capita meatconsumption in the developed world rose from60 kilograms in 1967 to 76 kilograms in 1997,per capita consumption in the developingworld more than doubled from 11 kilogramsin 1967 to 24 kilograms per capita in 1997(Table 2.10). Between 1990 and 1997, per capitameat demand in the developing world actuallygrew faster than in any other period at 5.2 per-cent annually, representing a substantialincrease above the 3.4 percent annual growthrate achieved between 1982 and 1990 (Table2.9). Total demand increased at a ratherremarkable 7.1 percent annually during1990–97, substantially above increases duringthe earlier periods. Therefore, the slowdownin global per capita meat demand growth canbe squarely laid at the door of declines in percapita demand in the developed world.

There are two reasons for believing that thisslowdown will be far less important in the longterm than its effect on overall statistics during1990–97 would indicate. First, given thatdemand in the developed world has fallen from71 to 47 percent of total demand over the 30-year period, and is likely to continue to fall, thedeveloped world’s demand will have less effecton overall demand trends. Second, the othermajor factor behind the sharp fall in meatdemand growth in the developed world—the economic collapse of the transitioneconomies—was a one-time event. Totaldemand in the transition economies fell at anannual rate of 8.5 percent between 1990 and1997, declining from about 20 million tons in1990 to 12 million tons in 1997, representing asignificant portion of the developed world’stotal demand of 98 million tons.

The evidence that demand growth willprogress rapidly in the developing world isstrong. First, while 53 kilograms of per capitameat consumption in Latin America (withmuch higher consumption in Argentina and

Brazil) is fairly close to the developed worldaverage of 76 kilograms per capita, most of thedeveloping regions are still significantly belowconsumption levels in the developed world.Following Latin America, at 43 kilograms percapita, China’s unsatisfied demand for meatwill continue to rise as its population becomesincreasingly wealthy and urbanized. If China’smeat consumption and production figures areoverstated, which seems highly plausible giventhe almost impossibly high growth in theregion during 1990–97, the divide between percapita consumption in the developed anddeveloping worlds is even wider than itappears.6

Second, high latent demand in income-constrained regions should keep meat demandgrowth rates in developing countries quitehigh over the foreseeable future if these coun-tries realize sustained income growth. Esti-mates of income elasticities indicate that incountries with low but rising per capitaincomes, meat demand grows faster than percapita income (Bhalla, Hazell, and Kerr 1999;


TABLE 2.10 Per capita meat demand,

1967 and 1997

Region 1967 1997

(kilograms/capita)

Latin America 33.1 53.0

Sub-Saharan Africa 10.0 9.9

West Asia/North Africa 12.0 21.0

Asia 7.3 23.6

South Asia 3.9 5.6

Southeast Asia 8.5 18.1

East Asia 9.5 42.1

Developing world 11.0 24.0

Developed world 59.5 75.8

World 25.5 36.0


Delgado et al. 1999). While the issue of cul-tural constraints on meat consumption inIslamic countries is certainly an issue, there islittle doubt that the major restraining factor onmeat demand growth in Sub-Saharan Africa,WANA (after 1982), and much of South Asiahas been income related rather than cultural.Increasing urbanization and the greater expo-sure of developing-country populations todeveloped-world lifestyles has played a contin-uing role in driving meat demand growth(Bhalla, Hazell, and Kerr 1999).

In addition to rapidly rising consumption,the major meat demand trend over the 30-yearperiod in developing countries has been thegrowing role of poultry in total meat con-sumption, particularly at the expense of beef.Growth of the poultry sector has drivenincreases in meat demand, with poultry rais-ing its share of total meat production in thedeveloping world from 12 percent in 1967 to26 percent in 1997 and in the developed worldfrom 15 percent in 1967 to 30 percent in 1997(Table 2.11). The most radical shifts in poultrymeat consumption occurred in regions that aretraditionally large producers of beef or sheepand goat meat, including Latin America, wherepoultry’s share of total meat consumption rosefrom 10 percent in 1967 to 35 percent in 1997,and WANA, where poultry’s share went from20 percent in 1967 to 50 percent in 1997. Percapita poultry demand rose accordingly,increasing from 1 kilogram in 1967 to 6 kilo-grams in 1997 in the developing world andfrom 9 to 23 kilograms in the developed world.Beef consumption increased relative to othermeats in some individual countries, includingChile, Japan, Malaysia, and South Korea.Although overall consumption of beef world-wide increased from 85 million tons in 1967 to206 million tons in 1997, beef ’s share of totalmeat consumption declined from 41 percentin 1967 to 27 percent in 1997 (FAO 2000b).

Livestock Trade. A rapid expansion of tradeaccompanied the large increase in meat con-sumption between 1967 and 1997, rising from5 million tons in 1967 to 21 million tons in 1997(Appendix Table B.3). Much of this tradeoccurred between countries in the developedworld, where Japan and FSU were the top twoimporters of meat products in 1997, with netimports of 2.4 million tons each (Table 2.12).The United States, which was a net importerat 0.7 million tons in 1967, led exporters at 2.5million tons in 1997. This marked shift was theresult of rising poultry exports. The develop-ing world tended to be less involved in themeat trade, with total net imports of 0.6 mil-lion tons in 1997, although both Latin Amer-ica and South Asia were net exporters of meatproducts in that year.

Feed Demand. Worldwide use of cereals forlivestock feed increased from 369 million tonsin 1967 to 659 million tons in 1997, represent-ing 36 percent of total cereal consumption inthat year. Feed demand grew much faster than

14 CHAPTER 2

TABLE 2.11 Poultry’s share of total meat

consumption, 1967 and 1997

Region 1967 1997

(percent)




Asia 12.6 21.4

South Asia 5.9 15.0


East Asia 11.6 19.3




food demand in the developing world, with thegrowth of feed demand reaching 6.4 percentannually during 1990–97, a period when fooddemand grew only 1.8 percent annually (Table2.13). The situation was far different in thedeveloped world, where feed demand growthactually turned negative between 1990 and1997 at −1.5 percent annual growth. Feed usedeclined from 462 million tons in 1990 to 425million tons in 1997. This phenomenon againowes much to events in the FSU, where cerealfeed demand collapsed more than demand forany other commodity, falling 12.3 percentannually from 123 million tons in 1990 to 56million tons in 1997. Considering that demandin Eastern Europe alone declined from 56 mil-lion tons in 1990 to 48 million tons in 1997, itis clear that demand growth in the rest of thedeveloped world, while slow, was certainly pos-itive during this period.

Feed Demand and Meat Production. It is ratherdifficult to generalize about the relationshipbetween feed demand and meat productionbecause so many different factors around theglobe have been at work over the last 30 years.Feed demand and meat production trackedtogether between 1967 and 1982, at 2.4 percentannually in the developed world, but the twohave since diverged, with feed demand growthlagging meat production growth almost acrossthe board (Table 2.13). This apparent trend maybe somewhat misleading, however, because ofthe extraordinary circumstances of the declinein feed demand in the transition economies.Several factors may help to explain this phe-nomenon. Although demand for noncereal feedcrops such as cassava and meals grew rapidlybetween 1982 and 1990, the overall demand forfeed lagged almost certainly because ofimprovements in livestock feeding efficiencyresulting from advances in management, genet-


TABLE 2.12 Net meat trade, 1967–97

Region/Country 1967 1982 1990 1997


United States −0.7 −0.7 −0.3 2.5

Japan −0.1 −0.5 −1.3 −2.4

Former Soviet Union 0.1 −1.0 −1.1 −2.4

Australia 0.5 0.9 1.0 1.3

Latin America 0.8 1.2 0.9 0.6

Sub-Saharan Africa 0.1 −0.1 −0.2 −0.1

West Asia/North Africa −0.1 −1.3 −0.9 −0.9

Asia

South Asia 0.0 0.1 0.1 0.2

Southeast Asia 0.0 0.0 0.1 0.0

East Asia 0.1 0.1 0.2 −0.2

Developing countries 0.9 −0.1 0.1 −0.6

Developed countries −0.7 0.8 0.9 2.5


ics, and hormone use (Rosegrant et al. 1997).Smil (2000) reports that between the early 1960sand 1998, ratios for feed per kilogram of meatproduced in the United States improved 25 per-cent. A factor in this improvement in efficiencywas the growing shift from beef to poultry pro-duction, since poultry requires less feed thanother livestock, and the poultry sector has seenrapid efficiency gains (Smil 2000). Geneticimprovements and better management prac-tices account for most of these gains (FAO2000b). In the United States, poultry productiongrowth significantly exceeded other meat pro-duction growth in every period.

Between 1990 and 1997, however, the trendof lagging cereal feed demand growth slightlyreversed itself, with cereal feed demand grow-ing faster than meat production in every devel-oped region except the transition economies

and the United States. Overall trends in theEU15 seem to have been dominated by domes-tic price distortions. When price ceilings oncereals used as feed were lowered under the1992 McSharry reform of the CAP, feed userecovered from the negative growth rates ofthe period 1982–90. In general, rapid cerealfeed demand growth in the developed worldin the 1990s may represent the effects ofreduced distortions in cereal markets as aresult of agricultural trade liberalization (FAO2000b).

Cereal feed demand in the developing worldgrew 6.4 percent annually during 1990–97.Negative growth in South Asia and very slowgrowth in Latin America between 1982 and1990 gave way to relatively strong perform-ances during 1990–97. Feed demand increasedmost rapidly in Asia due to expanding livestock

16 CHAPTER 2

TABLE 2.13 Growth rates of meat production and feed demand, 1967–97

1967–82 1982–90 1990–97

Meat Feed Meat Feed Meat Feed

Region/Country production demand production demand production demand

(percent/year)

United States 1.5 0.8 1.9 0.9 2.9 2.0

EU15 2.9 1.7 1.4 −2.1 1.0 2.6

Former Soviet Union 2.2 4.5 2.8 1.3 −9.5 −12.3

Australia 2.7 2.4 1.7 5.3 1.5 4.5

Eastern Europe 2.9 3.9 1.2 −0.3 −2.4 −2.7

Latin America 3.7 5.3 2.3 1.2 4.5 5.0



Asia 4.6 7.6 7.2 5.6 7.4 7.9

South Asia 2.9 5.2 4.4 −0.6 3.9 3.2

Southeast Asia 3.7 6.8 5.6 9.3 5.9 6.1

East Asia 5.3 7.8 8.1 5.3 8.2 8.3

Developed world 2.4 2.4 1.7 0.2 −0.5 −1.5

Developing world 4.1 6.3 5.1 4.0 6.2 6.4


production and low per capita land ratios (Del-gado et al. 1999). At 8.3 percent annually, feeddemand growth was particularly strong in EastAsia in the 1990s, especially in contrast to fooddemand for cereals, which did not grow at allduring the period. Between 1990 and 1997,feed demand also grew rapidly in SoutheastAsia and Latin America.

Between 1967 and 1997, annual feed de-mand growth outpaced meat productiongrowth in East Asia, Southeast Asia, and Sub-Saharan Africa, indicating the growing inten-sification of feed use per unit of meat output.In all regions of the developing world, graz-ing areas, mixed-farming systems, and small-scale backyard operations are diminishing inimportance in the face of land shortages, lowreturns to labor, and heightened competitionfrom large-scale producers (Delgado et al.1999). Industrial production of pork, poultry,feedlot beef, and mutton has been the fastestgrowing form of animal production in recentyears worldwide; it supplied 43 percent ofglobal meat production (over half of pork andpoultry production) in 1996, compared with37 percent of total production in 1992. Whileindustrial systems concentrated in the devel-oped world accounted for 52 percent of globalindustrial pork production and 58 percent ofindustrial poultry production in 1996, Asia isthe region with the fastest growing industri-alized livestock sector. It is already responsi-ble for 31 percent of all industrialized porkproduction worldwide. Industrial systemsdepend on outside feed, energy, and otherinputs, and are able to achieve large economiesof scale and high production efficiencies interms of output per unit of feed (de Haan, Steinfeld, and Blackburn 1996). Despite theirgreater efficiency, however, large operationsmay underuse certain crop residue and house-hold food waste feed sources that have beentraditionally important for small-scale pro-

ducers (Delgado et al. 1999). Industrializedlivestock production also poses significantenvironmental threats: very large flocks orherds lead to high waste volumes and signifi-cant animal health risks (de Haan, Steinfeld,and Blackburn 1996).

Roots and Tubers

While cereals and livestock are the mostimportant staple foods at the global level, rootsand tubers form an essential component offood security for many of the poor and under-nourished in the developing world, contribut-ing a significant amount to overall caloric con-sumption. Worldwide demand for roots andtubers for food stood at 359 million tons in1997, and demand for feed consumed an addi-tional 148 million tons (Appendix B, Table B.4).The developing world accounted for 72 per-cent of worldwide food demand and 65 per-cent of feed demand, although trends for thesecategories are moving in opposite directions;food demand growth accelerated to an annualrate of 3.3 percent between 1990 and 1997,after stagnating at 0.3 percent a year during theprevious period; at the same time feed demandgrowth declined from a torrid rate of 7.0 per-cent a year during 1982–90 to 3.6 percent a yearduring 1990–97 (Table 2.14). While roots andtubers are most important in Sub-SaharanAfrica, supplying 20 percent of all caloric con-sumption in the region, they also serve as animportant supplemental source of carbohy-drates, vitamins, and amino acids in Asia andLatin America. Within developing countries,roots and tubers are generally consumed inpoorer regions, such as Sichuan, China, andNorthern Brazil (Scott, Rosegrant, and Ringler2000).

Not much research has been done todevelop yield-enhancing technologies for rootsand tubers. Developing world yields grew at arate of only 1.0 percent annually during


1967–97, with this rate slowing slightly in recentyears (Scott, Rosegrant, and Ringler 2000). Areaexpansion in the developing world has also notbeen particularly rapid, averaging 0.9 percentannually between 1967 and 1997. One caveat tothese general trends is that developing worldpotato production, yields, and area have allexpanded much more rapidly than those ofother root and tuber crops; potato productiongrew at a rate of 3.9 percent annually during1967–97, with yields rising 1.9 percent and areaexpanding 2.0 percent annually over the sameperiod. Potato production and demand haveexpanded rapidly in Asia, which increased itspercentage of global consumption from 8.2 per-cent in 1967 to 27.8 percent in 1997.

Soybeans, Meals, and Oils

Production and consumption of edible oils,meals for livestock feed, and soybeans expandedrapidly between 1967 and 1997. In the develop-ing world, total demand for soybeans rose 6.9percent a year, demand for meals rose 6.4 per-cent, and demand for oils rose 5.1 percent (Table2.15). Feed demand has been the most dynamic

source of growth. Feed demand for meals stoodat 67 million tons in 1997, with growth averag-ing 6.7 percent annually over the period as awhole, with growth accelerating to 8.3 percentannually during 1990–97 (FAO 2000b). Processeduses, such as oils for human consumption andmeals for feed, have dominated growth in soy-bean demand, which averaged a torrid 9.1 per-cent annually between 1967 and 1997 to reach59.4 million tons in 1997 (from only a few mil-lion in 1967), although demand growth wasmost rapid between 1967 and 1982. Nonfood orfeed uses for oilseeds—averaging growth of 8.1percent annually between 1990 and 1997—havealso increased in importance, especially in Chinaand the European Union, where oilseeds serveas inputs for a large number of industrial prod-ucts with rapidly growing demand (FAO 2000b).

Oil crops are also essential components offood security. According to FAO (2000b), theyhave accounted for one out of every five calo-ries added to developing-world diets since 1976.Consumption of edible oilseeds reached 44.9million tons in 1997, with growth averaging 4.8percent annually between 1967 and 1997; direct

18 CHAPTER 2

TABLE 2.14 Growth rates of roots and tubers demand, 1967–97

1967–82 1982–90 1990–97

Region Food Feed Food Feed Food Feed

(percent/year)

Latin America −0.9 −2.0 −0.8 −2.3 0.9 −3.8


West Asia/North Africa 6.9 6.3 5.6 31.5 3.9 −15.5

Asia 0.8 4.4 −2.0 5.5 3.1 3.8

South Asia 3.8 3.8 2.8 −6.3 4.0 −6.4

Southeast Asia 1.7 2.9 −1.1 10.5 2.0 1.0

East Asia 0.2 4.5 −3.3 5.4 3.1 3.9


Developed world −0.4 −1.2 0.2 −0.8 1.4 −4.7


soybean food consumption grew at a slower rateof 2.9 percent annually during the period, reach-ing 11.8 million tons by 1997. Per capita fooddemand for oils is relatively high in Latin Amer-ica at 15 kilograms per capita, East Asia at 9 kilo-grams per capita, and South Asia at 9 kilogramsper capita. Per capita food demand growth hasbeen particularly fast in East Asia at 4.6 percent

annually and in Latin America at 4.1 percentannually, although China, India, and a few othercountries are effectively driving overall oil cropfood demand trends in the developing world.

As demand for oil crops has grown overrecent years, trade has expanded significantly.Total worldwide trade in meals rose from 9.7million tons in 1967 to 49.0 million tons in


TABLE 2.15 Growth rates of soybeans and meals demand, 1967–97

Soybeans

1967–82 1982–90 1990–97

Other Other Other

Region Food Feed uses Food Feed uses Food Feed uses

(percent/year)

Latin America 0.4 6.8 24.3 6.8 12.8 4.6 7.8 10.7 6.4

Sub-Saharan Africa 4.5 n.a. 12.2 13.4 n.a. 8.9 6.8 n.a. 2.8

West Asia/North Africa 26.7 14.5 24.9 12.3 11.7 1.0 1.3 8.9 11.3

Asia 1.1 0.5 4.9 2.3 −3.0 5.6 7.1 14.2 7.5

South Asia 27.9 41.7 25.7 −4.3 −10.4 27.4 30.3 −17.9 13.6

Southeast Asia 6.0 n.a. 8.7 7.8 n.a. 17.6 2.2 n.a. 2.8

East Asia 0.2 0.4 4.4 1.1 −3.0 2.1 7.6 14.2 6.0

Developing world 1.1 3.3 12.5 2.6 7.9 4.9 7.2 6.7 6.8

Developed world 1.1 6.5 5.3 2.3 16.0 0.1 0.4 0.4 3.1

Meals

1967–82 1982–90 1990–97

Other Other Other

Region Food Feed uses Food Feed uses Food Feed uses

(percent/year)

Latin America n.a. 9.5 −2.5 n.a. 2.5 56.8 n.a. 8.4 −44.3

Sub-Saharan Africa −2.2 7.8 12.9 8.4 4.3 −1.1 10.8 5.3 −166.7

West Asia/North Africa n.a. 10.5 −1.2 n.a. 5.3 14.8 n.a. 6.8 0.5

Asia n.a. 6.1 2.7 n.a. 5.0 −0.8 n.a. 8.7 7.2

South Asia n.a. 4.3 13.1 n.a. 7.5 0.3 n.a. 3.9 −1.5

Southeast Asia n.a. 8.7 9.5 n.a. 10.5 2.8 n.a. 7.3 53.7

East Asia n.a. 6.9 2.6 n.a. 1.8 − 0.9 n.a. 12.6 6.7

Developing world –2.2 7.2 2.8 8.5 4.4 –0.7 10.7 8.3 6.9

Developed world n.a. 4.7 0.8 n.a. 1.8 −13.3 n.a. 0.6 31.0


Note: n.a. indicates that data were not available.

1997, trade in edible oilseeds rose from 8.5 mil-lion tons in 1967 to 41.4 million tons in 1997, andsoybean trade rose from 8.1 million tons in 1967to 37.1 million tons in 1997. The United Statesand the EU15 are the two main developed-country players in world oil crop markets, withthe United States exporting 24 million tons ofsoybeans and almost 6 million tons of meals, andthe EU15 importing 14 million tons of soybeansand 15 million tons of meals in 1997 (Table 2.16).

The developing countries remained netexporters of meals and edible oilseeds between1967 and 1997—with net exports of meals asubstantial 12.7 million tons in 1997—but theyhave become net importers of soybeans inrecent years, importing 2.7 million tons in1997. Among developing countries, a few coun-tries dominate oil crop exports: Brazil is a netexporter of meals and soybeans, Argentina of

all three oil crops, Malaysia and Indonesia ofedible oilseeds, and India of meals. A signifi-cant number of developing countries are alsonet importers of these crops, led by China,with 4.7 million tons of net soybean imports,3.3 million tons of net edible oilseed imports,and 2.9 million tons of net meals imports.Other big importers include Mexico, with 3.3million tons of net soybean imports; India,with 1.8 million tons of net edible oilseedimports; and Pakistan, with 1.3 million tons ofnet edible oilseed imports. As developing coun-try demand for oil crops increases, the netexport status of the developing world will con-tinue to erode.

In the remaining sections of this historicalassessment chapter, we focus on the majordeveloping regions. Tables summarizing basicindicators for food supply and demand and

20 CHAPTER 2

TABLE 2.16 Net trade in meals, oils, and soybeans, 1967 and 1997

Meals Oils Soybeans

Region/Country 1967 1997 1967 1997 1967 1997


United States 3.0 5.9 1.5 2.3 7.3 24.0

EU15 −6.3 −14.6 −2.7 −1.1 −4.5 −14.4

Former Soviet Union 0.3 −0.1 0.8 −1.1 0.0 0.0

Australia 0.0 0.0 0.2 0.3 0.0 −0.1

Eastern Europe −0.9 −2.2 −0.1 −0.2 −0.1 −0.2

Latin America 1.6 19.6 0.2 2.4 0.1 5.4

Sub-Saharan Africa 0.9 0.5 0.5 −0.9 0.0 0.0

West Asia/North Africa 0.9 −3.2 −0.6 −4.5 0.0 −0.6

Asia 1.4 −2.1 0.3 4.8 0.2 −7.5

South Asia 0.9 4.2 −0.2 −4.1 0.0 0.0

Southeast Asia 0.4 −1.3 0.6 12.7 0.0 −1.2

East Asia 0.0 −5.1 0.0 −3.9 0.2 −6.2

Developing world 4.3 12.7 0.3 1.9 0.4 –2.7

Developed world −4.1 −12.0 −0.8 −0.4 −0.3 3.9


annual growth rates for each region are givenin Appendix C. For Asia, see Tables C.1 to C.6.

ASIA

Cereals

Slow Cereal Production Growth. The Green Rev-olution had a dramatic effect on food securityin Asia. It enabled the two most populouscountries in the region, China and India, toescape rising import dependence and periodicfood shortages. In the last decade, however,concerns have arisen about the health of cerealproduction systems in Asia. Recent signs indi-cate that phenomenal Green Revolutiongrowth in wheat and rice productivity is slow-ing, especially in the intensively cultivated low-lands. Since the early 1990s, rising unit pro-duction costs have led to a decline in farmerprofits in both India and China. Slackening ofinvestments in infrastructure and research andreduced policy support partly explain the slug-gish growth. Degradation of the lowlandresource base from long-term, intensive usealso has contributed to declining productivitygrowth rates.

Rice yield growth rates declined steadily inChina throughout 1967–97, falling from 2.8percent annually between 1967 and 1982 to 2.1percent between 1982 and 1990, and to 1.6 per-cent annually between 1990 and 1997 (Table2.17). Rice yields took off in India later than inChina, with annual growth rates rising from2.0 percent annually in 1967–82 to 3.4 percentin 1982–90, declining to 1.3 percent in 1990–97.The precipitous drop must be a cause for con-cern for a country that still has a massiveamount of food insecurity despite overallcereal self-sufficiency. Southeast Asia followeda similar trajectory to China. Its rice yieldgrowth in 1990–97 was the lowest of the threeregions.

Wheat yield growth throughout Asiadeclined from 4.7 percent annually in 1967–82to 2.9 percent in 1982–90, with a further declineto 2.5 percent annual growth in 1990–97 (Table2.17). Wheat yield growth in East Asia slowedfrom an annual rate of 5.5 percent in 1967–82to 3 percent thereafter. The slowdown in maizeyield growth has been less dramatic, sincemaize in Asia has not been subject to GreenRevolution–type yield growth. For Asia as awhole, maize yield growth dropped from 3.4percent a year during 1967–82 to 2.3 percent ayear during 1990–97. Although maize yieldgrowth expanded significantly in SoutheastAsia in the latter period, with the increasedadoption of hybrid varieties, it slowed dramat-ically in China and East Asia.

At the same time that cereal yield growthrates were declining in Asia, the contributionof area expansion to cereal production growthalso declined dramatically, as countries in theregion ran up against limits to remaining landsuitable for cultivation. For Asia as a whole,total growth in cereal area virtually stagnatedafter 1990, dropping from an already slow rateof 0.1 percent annually between 1982 and1990. Only maize area showed significantexpansion after 1990, growing at a rate ofnearly 1 percent annually. Wheat and rice areacontinued to expand after 1982, but at muchslower rates, and area planted to other grainsdeclined sharply. Wheat area grew at a rate ofonly 0.4 percent annually between 1990 and1997, while rice area expansion declined from0.7 percent per year to 0.4 percent over theperiod (Table 2.18). Area growth varied con-siderably across regions within Asia. East Asiaand South Asia showed declines in area plantedto cereals after 1982, with declines in rice andcoarse grain area offsetting expansion in maizearea (East Asia) and wheat area (South Asia).Cereal area continued to expand slowly inSoutheast Asia, mainly because rice area con-


TA

BL

E 2

.17

Gro

wth

of r

ice,

wh

eat,

an

d m

aize

pro

du

ctio

n a

nd

yie

lds,

Asi

a, 1

967–

97

Pro

duct

ion

gro

wth

Yie

ld g

row

thY

ield

s

Reg

ion/

Co

untr

y1

96

7–

82

19

82

–9

01

99

0–

97

19

67

–8

21

98

2–

90

19

90

–9

71

96

71

98

21

99

01

99

7

(per

cen

t/yea

r)(p

erce

nt/

yea

r)(m

etri

c to

ns/

hec

tare

)

Ric

e Chi

na3.

41.

81.

02.

82.

11.

62.

13.

23.

74.

2In

dia

2.7

4.2

1.7

2.0

3.4

1.3

1.0

1.3

1.7

1.9

Eas

t Asi

a3.

41.

80.

82.

82.

01.

52.

13.

33.

84.

2So

uth

Asi

a2.

63.

71.

61.

93.

11.

31.

11.

31.

71.

9So

uthe

ast A

sia

3.9

2.5

2.7

3.1

1.8

1.2

1.1

1.8

2.0

2.2

All

Asi

a3.

22.

51.

52.

52.

11.

11.

42.

02.

42.

6W

heat

Chi

na6.

53.

92.

75.

53.

03.

01.

12.

53.

13.

8In

dia

7.7

4.0

3.2

4.1

3.3

2.0

0.9

1.7

2.2

2.6

Eas

t Asi

a6.

53.

92.

65.

53.

03.

11.

12.

43.

13.

8So

uth

Asi

a7.

43.

63.

24.

12.

82.

00.

91.

72.

02.

4So

uthe

ast A

sia

3.7

0.3

−5.3

5.3

−2.9

−0.7

0.6

1.2

1.0

0.9

All

Asi

a6.

83.

82.

94.

72.

92.

51.

02.

02.

63.

1M

aize

Chi

na5.

14.

94.

14.

13.

41.

91.

83.

34.

35.

0In

dia

1.6

2.8

2.2

1.1

2.7

1.6

1.0

1.2

1.5

1.7

Eas

t Asi

a5.

14.

83.

64.

13.

41.

61.

83.

34.

44.

9So

uth

Asi

a1.

53.

12.

00.

92.

41.

41.

11.

21.

51.

7So

uthe

ast A

sia

4.1

4.1

3.2

2.6

2.7

4.0

1.0

1.5

1.8

2.4

All

Asi

a4.

54.

63.

43.

43.

32.

31.

52.

53.

23.

8––

––S

ou

rce:

Bas

ed o

n FA

OST

AT

dat

a (F

AO

200

0a).

Note

:n.

a. in

dica

tes

that

dat

a w

ere

not a

vaila

ble.

TA

BL

E 2

.18

G

row

th r

ates

of c

erea

l are

a, A

sia,

196

7–97

Eas

t A

sia

So

uth

Asi

aS

out

heas

t A

sia

All

Asi

a

Cer

eal

19

67

–8

21

98

2–

90

19

90

–9

71

96

7–

82

19

82

–9

01

99

0–

97

19

67

–8

21

98

2–

90

19

90

–9

71

96

7–

82

19

82

–9

01

99

0–

97

(per

cen

t/yea

r)

Whe

at0.

90.

9−0

.43.

20.

81.

1−1

.43.

3−4

.62.

00.

90.

4

Mai

ze1.

01.

32.

00.

60.

70.

61.

51.

4−0

.81.

01.

21.

1

Ric

e0.

6−0

.2−0

.70.

70.

60.

30.

80.

71.

40.

70.

40.

4

Oth

er c

oars

e gr

ains

−4.4

−4.5

−3.5

−0.9

−2.1

−2.5

5.4

−2.1

−0.3

−1.8

−2.6

−2.6

––––

Sou

rce:

Bas

ed o

n FA

OST

AT

dat

a (F

AO

200

0a).

tinued to grow. One trend that will exacerbatethe problem of Asian land shortages in thefuture is the ongoing removal of land fromagricultural use to satisfy the needs of expand-ing urban and periurban areas. Since 1979,China’s annual farmland losses have averaged500,000 hectares per year, with rice field lossesalone averaging 100,000 hectares per year(enough to feed half the country’s annual pop-ulation increase) (Smil 1998).7

Declining world cereal prices and factorsrelated to the increasing intensification ofcereal production have caused cereal produc-tion growth to slow in developing Asian coun-tries since the early 1980s. Declining cerealprices caused a direct shift of land out of cere-als into more profitable cropping alternativesand slowed growth in input use, thus hurtingyields. More important over the long run,declining world prices have also slowed invest-ment in crop research and irrigation infra-structure, with consequent effects on yieldgrowth (Rosegrant and Pingali 1994; Rose-grant and Svendsen 1993). Green Revolutiongrowth in cereal crop productivity resultedfrom an increase in land productivity; itoccurred through strong policy support andgood market infrastructure in areas of grow-ing land scarcity or high land values or both.High investment in research and infrastruc-ture—especially in irrigation infrastructure—resulted in the rapid intensification of agricul-ture in the lowlands, with the result that bothirrigated and high-rainfall lowland environ-ments became the primary source of food sup-ply for Asia’s escalating population.

The use of high levels of inputs andachievement of relatively high wheat and riceyields in parts of Asia have made it more dif-ficult to sustain the same rate of yield gains,as family farm yields in these regions approacheconomically optimum yields. By 1990, mod-ern varieties of rice occupied 74 percent of

rice area in Asia, accounting for all irrigatedarea plus about one-third of the rainfed low-lands. Opportunities for further expansion ofmodern variety use are essentially exhaustedin existing irrigated areas, and the risk ofdrought or flooding is severely constrainingdissemination in rainfed environments (Pin-gali, Hossain, and Gerpacio 1997). The declinein the potential for yield growth has been par-ticularly evident in India, where both the fulldiffusion of modern technologies in thenorthwest and the stagnation of agriculturalproductivity in the rest of the country con-tribute to the decline (Hopper 1999). Moregenerally, both China and India have under-gone significant shocks to their cereal pro-duction systems and major reforms to insti-tutions affecting agricultural performance. Itseems likely that very high yield growth inChina between 1970 and 1974 at least partiallyreflected recovery from the famine of1959–64, and major reforms undertaken inresponse to periodically recurring food crisesspurred rapid yield growth in India during the1980s (Dyson 1996). Gains from structuralchanges in response to crises are one-offeffects, although room remains for more eco-nomic reform throughout the region.

Environmental and resource constraintshave also contributed significantly to the slow-down in yield growth evident over the last twodecades. Increased intensity of land use has ledto increasing input requirements in order tosustain current yield gains. Moreover, Pingali,Hossain, and Gerpacio (1997) argue that thepractice of intensive rice monoculture itselfcontributes to the degradation of the paddyresource base and hence declining productivi-ties. Declining yield growth trends can bedirectly associated with the ecological conse-quences of intensive rice monoculture, includ-ing buildup of salinity and waterlogging, useof poor quality groundwater, nutrient deple-

24 CHAPTER 2

tion and mining, increased soil toxicities, andincreased pest buildup, especially soil pests.Salinization affects an estimated 4.5 millionhectares in India, and waterlogging affects afurther 6 million hectares (Abrol 1987; Cham-bers 1988; and Dogra 1986 as cited in Pingali,Hossain, and Gerpacio 1997. Many of thesedegradation problems are also prevalent in theirrigated lowlands, where farmers grow wheatin rotation after rice (Hobbs and Morris 1996).Experimental evidence from India shows thatconstant application of too low a level ofinputs over an extended period has led todeclining yields in rice-wheat systems (Paroda1998).

However, intensification per se is not theroot cause of lowland resource base degrada-tion; rather, a policy environment that encour-ages monoculture systems and excessive orunbalanced input use is to blame. Trade poli-cies, output price policies, and input subsi-dies—particularly for water and fertilizer—have all contributed to the unsustainable useof the land base. The dual goals of food self-sufficiency and sustainable resource man-agement are often mutually incompatible.Policies designed for achieving food self-sufficiency tend to undervalue goods nottraded internationally, especially land, water,and labor resources. As a result, food self-sufficiency in countries with an exhaustedland frontier came at a high ecologicaland environmental cost. Appropriate policyreform—at both macro and sectoral levels—will go a long way toward arresting and pos-sibly reversing the current degradation trends,but the degree of degradation in many regionswill pose severe policy challenges (Pingali andRosegrant 1998). But even if environmentaldegradation in intensive Asian cropping sys-tems were stabilized, it is unlikely that previ-ous crop yield growth rates will be restored,

as long as research and infrastructure invest-ments continue to decline.

Cereal Demand: Slowing Growth. The extent towhich all these potentially worrisome declinesmay represent a demand phenomenon morethan a supply shock is dependent on the regionunder discussion. China’s total per capita cerealdemand is shared fairly equitably among rice,maize, and wheat, all of which had per capitatotal consumption of approximately 100 kilo-grams in 1997 (Table 2.19). Nevertheless, maizedominated growth in cereal demand between1990 and 1997 to an extent not seen in priorperiods. Maize accounted for 22 kilograms ofthe total 24-kilogram increase in per capitacereal demand in China between 1990 and1997, whereas maize and wheat each accountedfor 13 kilograms of the 23 kilograms increasein per capita cereal demand between 1982 and1990.8 This increase in maize demand is anentirely feed-driven phenomenon, with thedemand for maize for animal feed actuallyincreasing a remarkable 33 kilograms per capitabetween 1990 and 1997 as food demand de-clined. The rising demand for maize as a feedcrop is due not only to the rapid expansion of theChinese livestock sector during this period, butalso to structural changes within the livestocksector leading to replacement of traditional feedswith cereals (Steinfeld and Kamakawa 1999).Despite this massive increase in per capita maizedemand, however, Chinese maize imports roseonly 1.6 million tons between 1990 and 1997,although they reached a peak of 5.0 million tonsin 1995. Meanwhile, rice demand has essentiallyleveled off.

These changes in Chinese diets are largelya function of increasing urbanization. Huangand Bouis (1996) show that diets change aspopulations move from rural to urban areas.Urban markets offer a wider choice of foods,and urban dwellers are exposed to the dietary


TA

BL

E 2

.19

P

er c

apit

a ce

real

dem

and

, Asi

a, 1

967–

97

Chi

naIn

dia

So

uthe

ast

Asi

aE

ast

Asi

aS

out

h A

sia

Cer

eal

19

67

19

82

19

90

19

97

19

67

19

82

19

90

19

97

19

67

19

82

19

90

19

97

19

67

19

82

19

90

19

97

19

67

19

82

19

90

19

97

(kil

ogra

ms/

cap

ita

)

Whe

at42

8194

9636

5462

696

1011

1641

7993

9439

5765

71

Mai

ze38

6578

100

1110

1011

2027

3847

3767

9394

3957

6571

Ric

e80

102

105

106

6674

8585

129

154

161

166

8210

310

610

573

7885

85

Oth

er c

oars

e 29

1913

1242

3326

221

12

231

2013

1234

2720

18

grai

ns

––––

Sou

rce:

Bas

ed o

n FA

OST

AT

dat

a (F

AO

200

0a).

patterns of foreign cultures. Urban lifestylesalso place a premium on foods that require lesstime to prepare (inducing, for example, a shiftfrom rice to wheat bread) as employmentopportunities for women improve and theopportunity cost of their time increases. Urbanoccupations tend to be more sedentary thanrural ones. People engaged in more sedentaryoccupations require fewer calories to maintaina given body weight. In addition, urban resi-dents typically do not grow their own food.Thus, their consumption choices are not con-strained by the potentially high cost of sellingone food item at farmgate prices (say, rice) tobuy another food item (say, bread) at retailprices (a choice faced by semi-subsistence pro-ducers). And while changes in food demandpatterns that cannot be attributed to increasesin household incomes and changes in foodprices may first be noticed in urban areas, asstructural transformation proceeds to a moreadvanced level, these same changes in fooddemand patterns may eventually move to ruralareas (Huang and Bouis 1996).

The demand trends have been slightly dif-ferent in India, however. Per capita cerealdemand has slowed significantly in the 1990s,only rising 4 kilograms per capita during1990–97 to 187 kilograms per capita, comparedwith gains in per capita consumption of 12kilograms during 1982–90. Increases in percapita cereal consumption in India over the last30 years have been unimpressive, especiallyconsidering the low initial level of consump-tion. Both wheat and rice contributed to theincrease in cereal consumption between 1982and 1990, with rice consumption rising 11 kilo-grams per capita and wheat consumption rising 8 kilograms per capita (Table 2.19). Consumption of other coarse grains declinedsignificantly during this period, from 33 kilo-grams per capita to 26 kilograms per capita.Between 1990 and 1997, however, rice con-

sumption remained constant, while wheat con-sumption continued its modest increases. Incontrast to China, maize consumption in Indiahas risen a bare 1 kilogram per capita between1982 and 1997, indicating that it has not yetbecome an important feedgrain in that coun-try.

Rice—at 166 kilograms per capita of con-sumption in 1997—dominates Southeast Asiandiets far more than it dominates Indian andChinese diets. Wheat only accounted for 16kilograms per capita of consumption in South-east Asia in 1997. Southeast Asian rice demandhas increased steadily since 1982, while wheatseems to be emerging as an increasinglyimportant crop, with demand rising 45 percentbetween 1990 and 1997.

These trends indicate that while the slow-down in Chinese rice production probably hasa significant demand component associatedwith it—mainly the shift to maize productionin response to rising demand for animal feed—the trend of slowing yields and production inIndia is more of a supply phenomenon andthus more of a cause for concern. Per capitarice and wheat consumption in India are still well below Chinese levels, despite somegrowth in per capita wheat consumptionbetween 1990 and 1997. Hopper (1999) calcu-lates that energy and protein supplies in Indiaonly grew at a rate of 1 percent per yearbetween 1980 and 1995, and that the develop-ing world as a whole has increased per capitaenergy supplies twice as fast as India since1960. Thus, despite overall food self-sufficiency, Hopper (1999) concludes that“average supplies of energy and protein inIndia are insufficient to meet average needs.”Protein in particular fell 21 percent below exist-ing per capita needs, and several million peo-ple did not receive the energy intake necessaryto escape wasting and stunting.


Livestock: Structural Changes Affect Both

Demand and Supply

A phenomenon oft noted in the literature con-cerns the dietary shift away from cereals androots and tubers to meat products as incomesrise and populations in developing countriesbecome increasingly urbanized (Bhalla, Hazell,and Kerr 1999). China’s per capita gross domes-tic product (GDP) in purchasing power parityterms rose from $509 in 1982 to $2,963 in 1997,9

while the percentage of the population inurban areas rose from 21 to 31 percent over thesame period. India’s per capita GDP in pur-chasing power parity terms rose from $736 in1982 to $2,036 in 1997, while the percentage ofthe population in urban areas only rose from24 to 27 percent (World Bank 2000b). Rapidchange in the Chinese society and economybrought about a steep rise in per capita meatconsumption (Table 2.20) from 15.2 kilogramsper capita in 1982 to 42.3 kilograms in 1997.Southeast Asian meat consumption also rosesubstantially during this period. Indian meatconsumption, however, only rose from 3.7kilograms to 4.5 kilograms. Thus, while meatconsumption in China has clearly risen in tan-dem with the significant shifts occurring in theChinese society and economy, India, withslower growth and levels of urbanization andsignificant cultural constraints against meatconsumption, has barely seen meat consump-tion rise at all.

Rising livestock demand in Asia has led tosurging livestock production accompanied bystructural changes to the sector. According toSteinfeld and Kamakawa (1999), the three mainchanges sweeping the Asian livestock sector inrecent years are the concentration of produc-tion near large cities favored by cheap input sup-plies and good output markets, the gradual shiftin production away from land-based systemstoward large-scale industrial operations, and the

vertical integration of primary production, pro-cessing, and marketing that has generated largeeconomies of scale. These changes to the live-stock sector, while supplanting the traditionalfunctions of livestock as assets, insurance, andobjects of sociocultural importance, have per-mitted a rapid expansion of livestock produc-tion that simply would not have been possibleunder low-intensity production (Steinfeld andKamakawa 1999). Between 1967 and 1997 live-stock production increased at an annual rate of6.7 percent in East Asia, 4.7 percent in South-east Asia, and 3.5 percent in South Asia. Pro-duction grew fastest in both East and SoutheastAsia during the period 1990–97, rising 8.3 per-cent annually in East Asia and 5.9 percent annu-ally in Southeast Asia. Within the livestock sec-tor, production has increasingly shifted towardpigs and poultry, which offer better feed con-version than ruminants, require less space, andprovide for flexible production (Steinfeld andKamakawa 1999). In East Asia, poultry produc-tion increased at an overall rate of 8.5 percentannually during 1967–97, but from 1990 to 1997the pace of growth was a torrid 14.4 percentannually, as production became increasinglyindustrialized (Table 2.21). The trend of rapidlyexpanding poultry production during the 1990swas similar in South and Southeast Asia. Egg

28 CHAPTER 2

TABLE 2.20 Per capita meat demand,

Asia, 1982–97

Region/Country 1982 1990 1997

(kilograms/capita)

India 3.7 4.2 4.5

China 15.2 25.5 42.3

Southeast Asia 10.3 13.6 18.1

East Asia 15.6 25.5 42.8

South Asia 4.2 4.9 5.6


production rose more slowly than poultry pro-duction in all Asian regions during the 1990s,but it actually grew faster between 1982 and1990 in both South and Southeast Asia. This wasnot the case in East Asia, where egg productiongrew fastest between 1990 and 1997 at 12.7 per-cent annually.

Modern, demand-driven, and capital-inten-sive production is thus dominating growth inAsian livestock production—particularly poul-try, eggs, pork, and occasionally milk—at theexpense of a more traditional, resource-driven,and labor-intensive sector. While this changehas brought with it rapidly expanding produc-tion, it has had its costs as well. Industrial live-stock production generates relatively littleemployment, poses severe environmental haz-ards due to its tendency to cluster near largeurban areas, and increases the potential sever-ity of animal health problems (Steinfeld andKamakawa 1999). Nevertheless, as long as ris-ing incomes and urbanization continue to gen-erate high demand for livestock products,structural change in the Asian livestock sectorwill be sure to follow.

Roots and Tubers: The Rise of the Potato

Demand for roots and tubers as a food sourceexpanded at a rate of only 0.6 percent annually

between 1967 and 1997 in Asia, but this slowoverall increase masks sharp differences amongregions, especially the growing importance ofpotatoes throughout East and South Asia.Potato demand grew at a modest 3.7 percentannually in East Asia during the period as awhole, but growth exploded to 11.2 percentannually during 1990–97, reaching 28.6 milliontons of food demand in 1997. South Asiandemand grew from 3.5 million tons in 1967 to17.7 million tons in 1997, a 5.6 percent annualrate of growth. Rapid economic developmentand rising incomes have driven potato demandin Asia, where consumers desire to diversifytheir diets and where potatoes are viewed as apreferred luxury good, in contrast to their infe-rior status in the developed world. In addition,potato demand has often increased at theexpense of less preferred, alternative com-modities such as sweet potatoes. And an increas-ingly urban population desires processed foodsgenerally associated with Western diets. Finally,expanding Asian production and resulting lowprices have further stimulated demand (Scott,Rosegrant, and Ringler 2000). Along with otherroots and tuber crops, sweet potatoes are animportant feed crop in China, with demand ris-ing at an overall rate of 4.4 percent annually to56.9 million tons in 1997. Sweet potatoes areused mainly as feed for pigs, and more than 80percent of pig production in China takes placeat the village and household level, much of it inSichuan Province, a geographically isolated areawithout easy access to feed imports (Scott,Rosegrant, and Ringler 2000). Recent shifts inthe livestock sector toward industrialized pro-duction of pigs and poultry slowed growth insweet potato feed demand in China to 2.7 per-cent annually between 1990 and 1997.

Conclusion

Asian agricultural production systems haveundergone dramatic changes over the last 30


TABLE 2.21 Poultry production growth

rates, Asia, 1967–97

Region/Country 1967–82 1982–90 1990–97

(percent/year)

India 3.8 12.5 6.6

China 5.0 9.9 15.1

East Asia 5.2 10.0 14.4


South Asia 4.1 10.3 8.9


years, but much work remains if the gainsalready achieved are to be extended to the vastnumbers who remain food insecure. In China,the two main problems confronting policy-makers over the next several decades will bepreservation of the natural resource base—withsustainable use of water resources and conflictover competing land uses the big issues—anddemand-driven structural change in the live-stock sector as intensive, industrial productiongradually replaces extensive, small-scale tradi-tional production. More generally, politicalchange in China remains the wild card in anyassessment of future agriculture performance.Indian policymakers currently face the difficultchallenge of dismantling a heavily state-centered food production system that has his-torically involved heavy input and consumptionsubsidies as well as import and export controls.Beyond the wastefulness and inefficiencies asso-ciated with such policies, a record of slow percapita consumption growth and persistent foodinsecurity on a massive scale speaks for itself.Given the realities of local production capacity,Indian policymakers must acknowledge the factthat food self-sufficiency is not a viable optionif the nation wishes to achieve true food secu-rity in the foreseeable future (Hopper 1999).Democratic accountability renders the processof gradual liberalization inevitable over the longrun, but the resistance of those who benefitfrom wasteful and expensive policies, as well asthe reluctance among much of the leadershipto return to high levels of cereal imports, hasrendered reform difficult over the short run.

LATIN AMERICA

Basic indicators for food supply and demandand annual growth rates for Latin America arepresented in Appendix C, Tables C7 and C8.

Macroeconomic Cycling and Recovery

The overarching story of the agricultural sec-tor in Latin America over the last 30 years hasbeen that of a policy-induced macroeconomiccycle, which led to expansion of the agricul-tural sector during the 1960s and 1970s,retrenchment during the 1980s, and subse-quent rapid growth in a low-inflation, liberal-izing environment during the 1990s. Duringthe 1960s and 1970s, Latin American countriesbenefited from the overall strength of theworld economy and were able to weather thefirst oil shock through a combination of highcommodity prices and accessible financing.Agricultural production accelerated rapidlyduring the 1970s, with high worldwide andincome-driven domestic demand fueling theexpansion of exports and supporting highprices. Nevertheless, the agricultural sectorgrew slightly slower than the overall economy,reflecting the region’s policy emphasis onimport-substituting industrialization and basicstructural shifts in economies undergoingrapid development.

In the early 1980s, Latin America’s unsus-tainable macroeconomic policies finally broughtthe region to crisis, as the second oil shockcaused declining terms of trade, falling exportvolumes, and skyrocketing interest rates. Theregion suffered another major trade shock in1986, when prices for a number of its majorexport commodities collapsed worldwide. Themacroeconomic crisis affected both privateinvestment and the banking sectors, with infla-tion rising from 45 percent during the 1970s to190 percent during the 1980s, and per capitaGDP falling 10 percent between 1980 and 1990(Reca and Díaz-Bonilla 1997; Garrett 1997). Thecrisis hit the agricultural sector hard throughoutthe region, particularly through reduced domes-tic demand. Governments across the regionwere forced to terminate support programs to

30 CHAPTER 2

heavily subsidized import-substituting sectorsof agriculture. Fiscal crisis also reduced govern-mental capacity to invest in agricultural researchand infrastructure development (Díaz-Bonilla1990). As a result, strong cereal productiongrowth of 3.5 percent between 1967 and 1982actually turned negative, with −0.1 percentannual growth in cereal production between1982 and 1990. Per capita cereal productiondeclined from about 262 kilograms per capita in1982 to 222 kilograms per capita in 1990, thenrose to 253 kilograms per capita in 1997.

While the 1980s is often called “the lostdecade” for Latin America, crisis did forceregional governments to undertake majorreforms during the mid- and late 1980s, begin-ning with initial efforts at economic stabiliza-tion through devaluation and cuts in govern-ment expenditures and eventually shifting intostructural reforms that included liberalizationof markets and reduction of trade barriers(Garrett 1997). Reforms led to improvement ofexternal indicators and strong prospects forlong-term recovery despite the fact that inter-nal economic and social indicators still reflectedthe difficulties faced by the region. The agri-cultural sector benefited tremendously fromthe reform efforts of the late 1980s, with deval-uation of the exchange rate and the advance oftrade liberalization removing the policy biasagainst agriculture and mitigating the negativeeffects of the overall deterioration of theregion’s infrastructure and the widespreadscarcity of inputs and credit (Díaz-Bonilla1999). Ultimately, the reforms of the late 1980sand early 1990s—combined with strong invest-ment inflows, rising world prices, and a rapidlyliberalizing regional trade system—led to atremendous rebound in the region’s agricul-tural sector (Díaz-Bonilla 1999).

Agricultural Production

Agricultural production growth was extremelyrapid throughout the region during the 1990s,with cereal production growth averaging 3.7percent annually—the fastest in the world—and meat production growth averaging 4.5percent annually between 1990 and 1997(Table 2.22). Latin America’s growth in percapita cereal production also led all otherdeveloping regions during 1990–97 at a rate of1.9 percent annually. Maize—the region’s dom-inant crop, representing 59 percent of totalcereal production in 1997—led the overall pro-duction increase with 5.3 percent annualgrowth, recovering from production growthof only 0.4 percent annually between 1982 and1990. In the meat sector, poultry productiongrew rapidly at 8.6 percent annually, althoughthis growth rate was still slower than the 9.4percent annual growth achieved in the sectorbetween 1967 and 1982.

While overall agricultural production per-formance was strong between 1990 and 1997,this period was also one of accelerating diver-


TABLE 2.22 Growth rates of meat

production, Latin America, 1967–97

Region/Country 1967–82 1982–90 1990–97

(percent/year)

Argentina 0.6 −0.1 1.0

Brazil 5.4 4.2 6.4

Colombia 3.6 4.5 2.2

Mexico 6.0 −0.3 4.7

Other Latin America 3.5 2.2 3.9

All Latin America 3.7 2.3 4.5

––––Source: Based on FAOSTAT data (FAO 2000a).Note: The four countries listed have the largestagricultural sectors in Latin America. The rest ofthe countries are combined under Other LatinAmerica.

gence between major agricultural producingcountries and the rest of the countries in theregion. Concentration of agricultural produc-tion in Latin America within a few economiesincreased substantially between 1965, when agri-culture in the three largest economies accountedfor 58 percent of value added in the sector, and1995, when the corresponding figure was 77 per-cent (Díaz-Bonilla 1997). On the trade side, tradeliberalization, appreciation of the real exchangerate, low world prices during the early 1990s, andthe termination of internal support measures ledto larger imports in many sectors, while funda-mental supply-side restructuring led to a signifi-cant increase in exports from other sectors.

Agricultural Demand

On the demand side, Latin American cereal con-sumption exhibited strong growth in 1967–82and 1990–97. The economic downturn duringthe 1980s, however, reduced cereal demandgrowth to 1.7 percent annually, by far the low-est in the developing world during that period.As shown in Table 2.23, per capita cerealdemand also rose significantly. The increasingimportance of feed demand as a component of

total cereal demand has been a major trend

(Table 2.24). All countries in the region except

Argentina experienced rapid rises in the feed

component of total cereal demand, with the

most dramatic increases occurring in Colom-

bia, where the share of total demand accounted

for by feed demand rose from 9.2 to 27.5 per-

cent. The anemic performance of Argentina’s

livestock sector explains its declining share of

the feed component.

Regional meat demand has expanded par-

ticularly rapidly in the 1990s, rising at a rate of

5.6 percent annually. Dairy demand rose 3.9

percent annually. During 1967–97, per capita

meat demand rose from 33 to 53 kilograms per

capita, while per capita dairy demand rose

from 96 to 125 kilograms per capita. Brazil had

a particularly large increase in per capita meat

demand, rising from 28 to 69 kilograms per

capita; meat demand in Argentina, on the

other hand, declined from 103 kilograms per

capita in 1967 to 91 kilograms per capita in

1997 (Table 2.25).

32 CHAPTER 2

TABLE 2.23 Per capita cereal demand, Latin America,

1967 and 1997

1967 1997

Region/Country Total Food Total Food

(kilograms/capita)

Argentina 427.4 133.2 372.4 130.5

Brazil 215.4 97.1 316.2 106.6

Colombia 88.9 71.1 146.3 96.4

Mexico 256.4 165.7 394.6 173.5

Other Latin America 155.4 108.3 189.8 112.7

All Latin America 212.5 113.8 280.9 122.3

––––Source: Based on FAOSTAT data (FAO 2000a).Note: The four countries listed have the largest agricultural sectorsin Latin America. The rest of the countries are combined underOther Latin America.

Brazil Emerges

Perhaps the biggest story in Latin America overthe 30-year period has been Brazil’s emergenceas the dominant agricultural force on the con-tinent. While Brazil expanded its share of pro-duction of cereals from 30 percent in 1967 to 33

percent in 1997 (Table 2.26), it expanded its shareof total regional meat production from 27 to 46percent, mainly through rapid growth in poul-try production of 9.8 percent annually and beefproduction of 4.6 percent annually. Major gov-ernment subsidy campaigns from the late 1960sto early 1970s and during the crisis period of themid-1980s helped achieve these gains (Díaz-Bonilla 1999). Brazil’s agricultural surge cameat the expense of Argentina, whose share oftotal regional meat production declined from32 percent in 1967 to 13 percent in 1997.Argentina’s beef sector, in particular, grewslowly in 1967–97, with production expandingonly 0.3 percent annually, with no growth at allbetween 1990 and 1997.

Agricultural Trade

Latin America’s emergence as a major cerealimporter over the last several decades has beena major development in world cereal markets.The region exported 3.1 million tons of cerealsin 1967, but by 1982, it was importing 3.5 mil-lion tons, and by 1997, a substantial 14.5 milliontons. These figures actually disguise the extentto which most countries in the region hadbecome heavily dependent on cereal imports by1997, since Argentina was a net exporter of 19.7million tons by that year, much of it to othercountries in the region. Argentina’s dependenceon regional export markets has increased sig-nificantly as it has lost market share in the restof the world, particularly the FSU. For instance,while total wheat exports to Brazil representedless than 10 percent of Argentina’s total exportsin 1985, that percentage jumped to between 60and 79 percent from 1993 to 1996. Other impor-tant Argentine cereal markets are Chile, Colom-bia, and Peru, all of which have grown in impor-tance since the mid-1980s (Díaz-Bonilla 1999).

Both Brazil and Mexico had emerged asheavy importers on world cereal markets by1997, with Brazil importing 9.3 million tonsand Mexico 9.5 million tons. Imports into Latin


TABLE 2.24 Share of feed in total cereal

demand, Latin America, 1967 and 1997

Region/Country 1967 1997

(percent)

Argentina 49.9 38.1

Brazil 41.3 53.7

Colombia 9.2 27.5

Mexico 22.4 38.9

Other Latin America 20.3 29.4

All Latin America 32.9 41.8


TABLE 2.25 Per capita meat demand, Latin

America, 1967–97

Region/Country 1967 1982 1990 1997

(kilograms/capita)

Argentina 103 99 92 91

Brazil 28 39 51 69

Colombia 23 27 33 38

Mexico 25 40 36 46

Other Latin America 26 31 30 36

All Latin America 33 41 43 53


America, excluding Argentina, represented 38percent of total regional production in 1997.These figures represent a sharp increase inimport dependence over a short period oftime—a trend spurred in part by low worldcereal prices in the second half of the 1980s,acceleration of economic growth and con-sumption, the liberalization of the region’strade regime, and the appreciation of domes-

tic currencies as capital flows returned to theregion (Díaz-Bonilla 1997). Perhaps the mostimportant factor in this rising import depend-ence, however, has been the growing role ofmaize as a feed crop (Table 2.27). Overall, theregion imported only 1.9 million tons of maizein 1997 because Argentina is a net exporter of9.9 million tons. Both Mexico and the coun-tries in the Other Latin America category have

34 CHAPTER 2

TABLE 2.26 Share of country’s cereals, meat, and roots and tuber production

in total Latin America production, 1967 and 1997

Cereals Meat Roots and tubers

Region/Country 1967 1997 1967 1997 1967 1997

(percent)

Argentina 30 27 32 13 5 7

Brazil 30 33 27 46 69 54

Colombia 3 2 5 5 4 10

Mexico 23 23 12 14 1 3

Other Latin America 15 14 23 22 20 27

––––Source: Based on FAOSTAT data (FAO 2000a).Note: The four countries listed have the largest agricultural sectors in Latin America.The rest of the countries are combined under Other Latin America.

TABLE 2.27 Maize production, demand, and net trade, Latin America, 1967 and 1997

Feed demand as

Production Demand percent of total Net trade

Region/Country 1967 1997 1967 1997 1967 1997 1967 1997

(million metric tons) (percent) (million metric tons)

Argentina 7.4 15.1 3.4 4.5 87.2 55.9 3.7 9.9

Brazil 12.3 32.1 11.4 33.7 69.1 79.5 0.7 −0.6

Colombia 0.9 0.9 0.9 2.7 10.0 40.3 0.0 −1.8

Mexico 9.0 18.0 7.8 21.5 12.8 21.9 1.0 −4.4

Other Latin America 4.2 7.9 4.6 13.2 33.4 50.6 −0.4 −5.0

All Latin America 33.8 74.1 28.1 75.6 48.0 55.3 5.0 –1.9

––––Source: Based on FAOSTAT data (FAO 2000a).Note: The four countries listed have the largest agricultural sectors in Latin America. The rest of thecountries are combined under Other Latin America.

become heavily dependent on maize importsover the 30-year period, however.

Crop Yields: Long–Term Growth

with Regional Disparities

Although the macroeconomic boom-bustcycle examined earlier has contributed greatlyto short-term variability in yield growth overthe 30-year period, the long-term trend hasbeen sustained productivity growth driven bythe adoption of Green Revolution technolo-gies. Use of fertilizer, irrigation, and improvedseeds all expanded rapidly. Adoption of newtechnologies actually progressed most rapidlyin the post-reform period, when a less-dis-torted environment encouraged heightenedinvestment and increasing production scale. Incontrast to the somewhat worrying yield signscoming out of Asia during the 1990s, the mainstory of domestic cereal production between1990 and 1997 in Latin America is the rapidyield increase of 3.4 percent annually that theregion achieved over this period. Nevertheless,recent declines in national research investmentand funding for extension agencies—andincreased reliance on nongovernmental organ-izations (NGOs) and the private sector—bodesill for future technology improvements, par-ticularly for smallholders who are generally ill-served by the private sector (Trigo 1995).

Despite strong trends at the regional level, aparsing of overall yield growth figures revealsthat the giants of Latin American cereal pro-duction—Argentina and Brazil—have been themain sources of yield increases, with annualyield growth rates of 4.9 percent and 4.2 per-cent, respectively, during 1990–97 (Table 2.28).Medium and large mechanized farms increas-ingly dominate agricultural production inArgentina and Brazil (mainly Southern Brazil),bringing with them strong economies of scaleand rapidly increasing use of agrochemicals andmonocultures (Garrett 1997). For instance, fer-

tilizer use in wheat production in Argentina rosefrom 25 percent of area planted in 1991 to 64percent of area planted by 1996 (Díaz-Bonilla1999). In Argentina and Brazil, maize has beenthe driving force behind high yield growth rates,averaging 5.0 percent annually in Argentina and4.3 percent in Brazil during 1990–97. Adoptionof hybrid seeds and intensive use of chemicalinputs help explain the increase of yields inthese countries during the 1990s, although thelong-term effects of the expanding scale andintensification of agricultural production maybe less positive than recent yield growth wouldindicate, with soil erosion, declining water qual-ity, and disease resistance becoming increasinglyproblematic (Garrett 1997).

The technological transformations thathave boosted Brazilian and Argentine agricul-tural productivity do not seem to have dra-matically affected other countries in the regionso far. All other countries or country group-ings in the Latin American region were belowthe developing world average of 1.9 percent


TABLE 2.28 Growth rates of cereal yields,

Latin America, 1967–97

Region/Country 1967–82 1982–90 1990–97

(percent/year)

Argentina 3.2 −0.3 4.9

Brazil 1.2 2.2 4.2

Colombia 3.3 0.3 1.6

Mexico 3.5 0.4 1.6

Other Latin America 2.2 2.3 1.0

All Latin America 2.5 0.8 3.4

Developed world 1.7 2.2 1.9



annual yield growth between 1990 and 1997,although all exceeded 1.0 percent. Historically,the period 1967–82 was one of strong growthin Latin American yields, with Argentina,Colombia, and Mexico all experiencing yieldgrowth of more than 3 percent per year. How-ever, performance was mixed in 1982–90, withyields stagnating in much of the region andturning slightly negative in Argentina.

Oils and Meals

Oils and meals, particularly meals, emerged asan important agricultural export earner in LatinAmerica between 1967 and 1997. Argentina andBrazil dominated the sector, with Argentinaincreasing its net meal exports from 1.0 milliontons in 1967 to 10.9 million tons in 1997, andBrazil from 0.4 million tons in 1967 to 10.5 mil-lion tons in 1997. Brazil and Mexico alsobecame world players in soybean markets, withBrazil increasing its net exports to 6.0 milliontons in 1997 from only 0.2 million tons in 1967,while Mexico became a net importer of 3.3 mil-lion tons of soybeans. Argentina also became amoderate player in oils markets, exporting a netof 3.5 million tons in 1997. Latin America beganto produce oils and meals during the 1940s,when export of olive oil from Europe wasembargoed, but it did not achieve extraordinarysuccess until Argentina and Brazil widelyadopted soybean production in the 1960s. By1997, soybean production had reached 27.1 mil-lion tons in Brazil and 14.1 million tons inArgentina. Strong export demand, develop-ment of a system for supplying technical inputssuch as seeds and agrochemicals, and a vibrantentrepreneurial class combined with animproving macroeconomic policy environ-ment all contributed to their success (Díaz-Bonilla 1997). Expanding soybean productionhas not been without its share of problems: inthe Argentine pampean region, the replace-ment of traditional extensive farming with

intensive production systems led to soil com-paction and erosion (Ekboir and Parellada2000).

The Smallholder Challenge

In many ways, the strong performance of LatinAmerican agriculture as a whole in the 1990sobscures a widening of the gap between theagricultural superpowers of the region andother countries continuing to struggle withfairly slow yield growth rates. One of the rea-sons behind this trend is undoubtedly the con-tinuing predominance of smallholders through-out the hillsides of Central America, Centraland Southern Mexico, and Northeast Brazil.These smallholders tend to diversify productionto reduce risks and income variation, but theylose the benefits of specialization and scaleeconomies as a result (Garrett 1997). Small-holder production systems are becomingincreasingly dependent on their ability to adoptnew technologies and participate in labor andcredit groups that enable them to take advan-tage of economies of scale (Díaz-Bonilla 1999).An important question regarding future agri-cultural development in Latin America iswhether smallholders can enhance their pro-ductive efficiency in order to compete withlarge-scale agriculture, while maintaining theintegrity of the natural resource base. Alterna-tively, large numbers of smallholders may haveto be smoothly integrated into the nonagricul-tural sectors of the economy (Garrett 1997).

SUB-SAHARAN AFRICA

Basic indicators for food supply and demandand annual growth rates for Sub-Saharan Africaare given in Appendix C, Tables C.9 and C.10.

Introduction

Sub-Saharan Africa’s agricultural performanceduring 1967–97 was the worst in the developing

36 CHAPTER 2

world, and the region now finds itself in muchthe same position as India in the 1960s, with highpopulation and food demand growth exceedingmodest production growth (Byerlee and Eicher1997). The reasons behind Africa’s poor agri-cultural performance are myriad. The continenthas been afflicted with the triple curse of poorresource endowments (including poor land qual-ity, large landlocked areas, endemic livestock dis-ease, and human diseases), a colonial legacy ofextraction and exploitation, and a policy envi-ronment that consistently undermined agricul-ture and the institutions that served it (WorldBank 2000a). In addition to problems of climateand geography, the pernicious influence of yearsof exploitation and colonialism lingered into thepost-colonial period, contributing to underde-velopment of agriculture. These factors havebeen intensified in recent years by adverse agroclimatic conditions including significantdroughts in 1983, 1984, and 1992. Nevertheless,the blame for lagging agricultural sector per-formance rests mainly with poor developmentstrategies and policy choices, including the over-all unwillingness of many national leaders to rec-ognize the importance of agriculture to overalleconomic growth (Abdulai and Hazell 1995).

Despite common problems across the conti-nent, it is probably useful to disaggregate Sub-Saharan African production systems into land-constrained systems (prevalent in Eastern andSouthern Africa as well as humid parts of West-ern Africa) and labor-constrained systems(mainly prevalent in West and Central Africa). Inthe land-constrained systems, primary impedi-ments to agricultural growth have historicallybeen structural, with high transport costs, under-financing of research systems, and lack of sup-port for smallholder agriculture. In the labor-con-strained systems, the challenge has traditionallybeen to find technologies and a policy environ-ment to permit increases in labor productivity,mainly through mechanization (Delgado 1996).

Post-Independence Agricultural Policy

and Recent Reforms

In the 1960s, 1970s, and early 1980s, mostAfrican governments attempted to acceleratethe process of industrial development and toensure cheap urban food prices by taxing agri-culture through measures including over- valued exchange rates and price-depressingmarketing board interventions in food markets(Kherallah et al. 2000). However, unlike EastAsian governments, which invested consis-tently and generously in smallholder agricul-tural and rural development even while theytaxed it heavily, African governments havetended to invest little in the sector, with subsi-dies for fertilizer and credit generally benefit-ing entrenched interests—usually larger,export-oriented farmers—capable of exercis-ing political power (Kherallah et al. 2000;World Bank 2000a; Killick 1994). While gov-ernments in East and South Asia encouragedthe adoption of productivity-enhancing tech-nologies through significant investments in therural infrastructure necessary to enable thecommercialization of smallholder agriculture,African road and communications infrastruc-tures remain undeveloped and inadequate.The high transfer costs, limited information,and general fragmentation that resulted ledDelgado (1996) to the conclusion that much ofthe rural economy in Sub-Saharan Africa isprobably demand constrained, meaning thateven if macroeconomic and trade reformsreduced price distortions, supply might notincrease for a long time without strong locallygenerated demand. Severe underutilization ofresources and low productivity characterizedemand-constrained local economies.

Input markets have been inefficient and prop-erty rights weak throughout the region sinceindependence, thus limiting long-term entre-preneurial planning and undermining both the


will and ability of farmers to invest in the prof-itability of their land. Low-intensity agriculturehas been the norm. Fertilizer application ratesin 1997 were only 8 kilograms per hectare ofarable and permanent cropland, compared withthe Asian total of 135 kilograms per capita,although these figures mask regional differencesand the far greater prevalence of irrigated sys-tems in Asia. Fertilizer policy in Sub-SaharanAfrica has been disastrous overall, with costlysubsidies distorting markets during the 1970sand 1980s (World Bank 2000a).

Economic crises during the 1980s forcedmany Sub-Saharan African countries to under-take long-run structural adjustment programsembodying a wide range of market liberaliza-tion and public sector reforms, including majoragricultural sector reforms. Guided by a beliefthat the introduction of market forces to theagricultural sector could jump-start growth,reforms generally focused on liberalizing inputand output prices, removing regulatory controlson input and output markets, and restructuringpublic enterprises (including the elimination ofthe regulatory functions of marketing boards).Reforms to food crop markets were more com-prehensive than reforms to export crop marketsin Central and Western Africa, but actualreforms in Southern and Eastern Africa havebeen limited, with state trading and price bandsremaining in effect in a number of countries,including Kenya, Malawi, and Zimbabwe (Kher-allah et al. 2000). Input market reforms have beensignificantly less comprehensive throughout theregion, as state-owned enterprises continue todominate fertilizer, seed, and agrochemical mar-kets, despite the penetration of some multina-tionals and private traders. All reform effortshave been subject to backsliding and the resist-ance of entrenched groups that do not wish tolose their access to rents and privileges. Mostregional governments—generally lacking strongpolitical legitimacy—have been unwilling or

unable to obtain strong indigenous support formajor reform efforts, with donor demands andprescriptions tending to dominate policymak-ing (Kherallah et al. 2000).

Cereals

Cereal Production. Despite the many problemsfaced by agriculture in Sub-Saharan Africa, the30-year period has seen a certain amount of suc-cess on the cereal production side. Cereal pro-duction more than doubled from 31 milliontons in 1967 to 69 million tons in 1997, rising abrisk 4.2 percent annually in 1982–90 and 2.9percent in 1990–97. However, rapid populationgrowth actually led to declining per capita cerealproduction, which fell from 128 kilograms in1967 to 124 kilograms in 1997 (Appendix TableC.9). Nevertheless, production growth wasstrong during 1990–97 in Northern Sub-Saha-ran Africa (rising 4.9 percent annually) (Table2.29). Production growth rates slowed moder-ately in Central and Western Sub-SaharanAfrica, increased slightly in Southern Sub-Saharan Africa, and dropped sharply in EasternSub-Saharan Africa (Table 2.29).

While conflict has afflicted a number of coun-tries in both Northern Sub-Saharan Africa andCentral and Western Sub-Saharan Africa during

38 CHAPTER 2

TABLE 2.29 Growth rates of cereal

production, Sub-Saharan Africa, 1967–97

Region/Country 1967–82 1982–90 1990–97

(percent/year)

Northern 2.3 2.0 4.9

Central and Western 1.9 3.4 3.0

Eastern 3.0 2.8 0.1

Southern 0.8 1.9 2.0

Nigeria 1.1 9.9 2.7

All Sub-Saharan Africa 1.8 4.2 2.9


most of the 1990s, strong production growthduring this period may be an indication that mar-ket reforms are finally having a positive effect onproduction performance. Devaluation in West-ern Africa has aided local food markets by boost-ing prices for the imported food that local pro-duce competes against. There is also someindication that increased use of fertilizers bycash crop producers has spilled over onto foodcrop production. However, it is far from certainthat the fuller implementation of reforms inCentral and Western Sub-Saharan Africa hasboosted their performance vis-à-vis the rest ofthe region; while Jaeger (1992) and Faini (1992)note a positive cereal production response toadjustment policies, Seppala (1997) concludesthat marketing reforms had no effect onfood production. According to Seppala (1997),food production performance varied greatlybetween countries with similar amounts of gov-ernment intervention, with differences mainlydue to varying levels of food self-sufficiency,crop choice, and drought prevalence. In anyevent, as was noted earlier in the discussion ofdemand constraints on rural producers, mostobservers agree that structural and institutionalconstraints continue to exert a severe drag onagricultural performance, despite some degreeof market liberalization (Kherallah et al. 2000).

Cereal Yields. Unlike other regions, where yieldand production performance tend to closelytrack each other, Sub-Saharan Africa has hadoccasionally strong production performanceaccompanied by dismal yield performance (seeAppendix B, Table B.5). Cereal yields have fallensteadily behind those in other developingregions, partly because African yields have stag-nated and partly because cereal yields havegrown rapidly in regions that have successfullyadopted Green Revolution technologies on alarge scale (Table 2.30). However, maize andother grains—the region’s two main cereal

groups—were not targeted by the Green Revo-lution. Therefore, the lagging performance ofthese crops can mainly be attributed to Sub-Saha-ran Africa’s poor agricultural performance. Infact, cereal yield growth rates in the region havelagged well behind those of every other devel-oping region in all three periods, and they wereeven negative during 1982–90 (–0.2 percent annu-ally). The 1990–97 period has not been much bet-ter, with yield growth of only 0.3 percent annu-ally. Market reforms were expected to boostproductivity by increasing the availability ofmodern inputs and encouraging their use, butthe limited evidence of productivity improve-ments in the post-reform period mainly comesfrom countries that heavily taxed agriculturalproduction in the pre-reform period (mainlyWest Africa). In many other countries—particu-larly in Eastern and Southern Africa—subsidyreductions in the absence of credit markets haveboth limited farmer access to inputs such as fer-tilizer and encouraged the use of scarce inputson export crops. Market reforms have thus failedto have the dramatic impact on productivityexpected of them (Kherallah et al. 2000).

Maize. Maize shows the variability in perfor-mance that characterizes production systems in


TABLE 2.30 Cereal yields in Sub-Saharan

Africa as a percent of cereal yields in

developing countries, 1967–97

Cereal 1967 1982 1990 1997

(percent)

Maize 69.4 56.0 50.0 44.0

Wheat 89.4 75.9 69.6 62.3

Other coarse 80.5 76.2 71.3 68.5

grains

Rice 64.4 47.3 48.2 43.1


Sub-Saharan Africa (Table 2.31). Because maizeis a politically important crop, particularly inSouthern and Eastern Africa (and increasinglyin parts of Western Africa as well), it has beenthe focus of a series of government initiatives toraise productivity through the adoption of seed-fertilizer technology, input and credit subsidies,price supports, and investments in marketingand infrastructure. Governments in Eastern andSouthern Africa have devoted a considerableshare of their miniscule agricultural researchbudgets to adopting technological packages orig-inally developed for commercial farmers to theneeds of smallholders (Byerlee and Eicher 1997).

Efforts to improve maize technologies havemet with some success, as improved maize vari-eties are now grown on approximately 40 per-cent of maize area in Sub-Saharan Africa,although fertilizer use remains low even onimproved varieties (Byerlee and Eicher 1997).Varieties of maize responsive to increased use ofchemical fertilizers, herbicides, and pesticideshave particularly benefited highland areas (Del-gado 1996). Despite these technological devel-opments, however, maize yields region-widedeclined from 69 percent of the developingworld average in 1967 to 44 percent in 1997. Cen-tral and Western Africa have had the most con-

sistent maize yield growth rates, even thoughmaize in these regions has been the least affectedby new technologies. Southern Africa had thestrongest maize yield growth during 1990–97 at1.9 percent annually, rebounding from negativegrowth of −0.2 percent during the conflict-rid-den period between 1982 and 1990. The worstmaize yield performers in the 1990s have beenNorthern and Eastern Sub-Saharan Africa: yieldsin Eastern Sub-Saharan Africa actually declined0.9 percent annually between 1990 and 1997,while yields in Northern Sub-Saharan Africarose only 0.4 percent annually. The miserableperformance in these regions between 1982 and1997 stands in stark contrast to the impressivemaize yield growth rates of approximately 3 per-cent annually that they achieved between 1967and 1982. Nigeria has been somewhere inbetween these extremes, with severe negativegrowth in maize yields of −1.1 percent between1982 and 1990, but with growth reboundingsomewhat annually between 1990 and 1997.

The remarkable disjuncture between researcheffort and technological diffusion on the onehand and yield performance on the other indi-cates that technological development has notbeen the main factor behind short-term maizeyield trends. This cautionary tale serves as awarning about the perils of relying on magic bul-lets to raise African cereal yields, especially GreenRevolution technology that has largely suc-ceeded in an Asian setting. Byerlee and Eicherblame the sluggish yield response on three mainfactors: first, the lack of technical progress inmajor producing regions of Ethiopia and Tan-zania; second, the shift of much production fromhigh-yielding, large commercial farms to low-yielding, small-scale farms;10 and third, limitedadoption of complementary inputs such as fer-tilizer and other soil fertility–related practicesto accompany new seed varieties (Byerlee andEicher 1997). These factors have all played outin an environment characterized by frequent

40 CHAPTER 2

TABLE 2.31 Growth rates of maize yields,

Sub-Saharan Africa, 1967–97

Region/Country 1967–82 1982–90 1990–97

(percent/year)

Northern 3.1 −0.1 0.4


Southern 0.1 −0.2 1.9

Eastern 3.0 0.0 −0.9

Nigeria 2.7 −1.1 1.0



intense conflicts, poorly designed and im-plemented policies, and weak institutionalarrangements. Besides having an independenteffect on yields, the environment also has a closerelationship with technological development:technological progress in large maize produc-ers such as Ethiopia, Tanzania, and Zaire hasbeen quite slow precisely because these coun-tries have either been racked by violence orexperienced severe institutional decline.

Other Coarse Grains. Other coarse grains11—amajor food source in Northern Sub-SaharanAfrica, where they represented 70 percent oftotal cereal production, and Nigeria, 63 percentin 1997—had particularly dismal yield growthof −1.6 percent annually across Sub-SaharanAfrica during 1982–90 (Table 2.32). This overall

poor performance was driven by declines inother coarse grain yields of −5.5 percent annu-ally in Nigeria, from 1,598 kilograms per hectarein 1982 to 1,014 kilograms per hectare in 1990.12

Yields did turn around between 1990 and 1997,rising 1 percent annually in both Nigeria andNorthern Sub-Saharan Africa. Nevertheless,other coarse grain yields in Sub-Saharan Africafell from 80 percent of the developing worldaverage in 1967 to 69 percent by 1997.

Cereal Trade. Cereal imports into Sub-SaharanAfrica have remained low during the last threedecades because of persistent poverty and thelow import capacity of the region. Thus, despitewidespread and persistent malnutrition, theregion only increased net cereal imports from1.5 million tons in 1967 to 12.0 million tons in1997, with Central and Western Africa account-ing for 36 percent of all cereal imports in 1997(Table 2.33). One key component of Africancereal import patterns is the dominance ofwheat, historically purchased cheaply throughovervalued exchange rates to support volatileand growing urban populations and thus under-mining domestic prices and production. Sub-Saharan Africa imported a net 6.5 million tons


TABLE 2.32 Growth rates of production

and yields of other coarse grains, Sub-

Saharan Africa, 1967–97

Region/Country 1967–82 1982–90 1990–97

(percent/year)

ProductionNorthern 3.1 −0.1 0.4

Central 0.4 2.2 1.1

Southern 0.1 −0.2 1.9

Eastern 3.0 0.0 −0.9

Nigeria 2.7 −1.1 1.0


Yields

Northern 0.0 −1.7 1.0

Central 1.0 0.7 0.6

Southern −1.7 0.0 0.5

Eastern 2.0 −0.6 −0.2

Nigeria 6.6 −5.5 1.0

All Sub-Saharan Africa 1.7 −1.6 0.8


TABLE 2.33 Net cereal trade, Sub-Saharan

Africa, 1967–97

Region/Country 1967 1982 1990 1997


Northern −0.3 −1.4 −2.3 −2.2

Central and Western −1.0 −2.8 −3.5 −4.3

Eastern 0.0 −0.5 −0.3 −1.3

Southern 0.0 −1.4 −1.4 −2.2

Nigeria −0.2 −2.1 −0.6 −1.9

All Sub-Saharan Africa −1.5 −8.3 −8.1 −12.0

––––Source: Based on FAOSTAT data (FAO 2000a).Note: Positive figures indicate net exports; negativefigures indicate net imports.

of wheat and 3.5 million tons of rice in 1997.Another important component of cereal tradewith ramifications for food security is the factthat both Eastern and Southern Africa movedfrom positions of maize self-sufficiency to netimport status during the period, although inboth cases these imports amount to less than 1million tons (Byerlee and Eicher 1997).

Roots and Tubers

Roots and tubers are a major contributor to foodsecurity in Sub-Saharan Africa, particularly inyears of bad cereal harvests. A recent studyfound that the most important reasons cited bySub-Saharan African farmers for increasing pro-duction of roots and tubers were famine,hunger, and drought; the crop’s low inputrequirements mesh well with regional resourceendowments, a shortage of chemical inputsand organic products, and limited irrigation(Scott, Rosegrant, and Ringler 2000). Per capitademand for roots and tubers for food in Sub-Saharan Africa was 163 kilograms in 1997, oralmost three times higher than the developingworld average. Cassava was by far the dominantroots and tubers crop with per capita con-sumption of 119 kilograms in 1997. Roots andtubers consumption is concentrated in Centraland Western Africa and Nigeria, where demandin 1997 averaged 264 and 216 kilograms percapita, respectively. Production averaged 45 mil-lion tons in the Central and Western region and58 million tons in Nigeria in 1997 (Table 2.34)Despite the importance of roots and tubers tofood security and the prevalence of relativelyfavorable agro-climatic conditions throughoutmuch of the region, roots and tubers yields inall countries and subregions in 1997 were lowerthan the developing world average of 11,777kilograms per hectare (Table 2.34).

Regional roots and tubers productiongrowth has been relatively high over the past30 years, averaging 4.6 percent in 1982–90 and

3.9 percent in 1990–97 (Table 2.35). Higherprices, increased farmer access, and risingcommercialization have all played a role inexpanding production throughout the region.Production growth has been particularly rapidin Nigeria, averaging 9.9 percent in 1982–90and 7.6 percent in 1990–97. According toOuraga-Djoussou and Bokanga (1998), a num-ber of factors have played a role in the strikingincrease in the importance of roots and tubersto Nigerian diets between 1982 and 1997,including a ban on cereal imports between1987 and 1990, the multiple uses of the crop asa contributor to food security, and the treat-ment of cassava as a high-value good by a rap-idly expanding urban population. Productionhas kept up with rising demand through incen-tives provided by the high profitability of com-mercial sales of processed and fresh roots,technical improvements in processing, and theintroduction of high-yielding varieties (Scott,Rosegrant, and Ringler 2000).

Outside of Nigeria, the yield performance ofroots and tubers in the region has been relativelyunimpressive, except for Southern Africa, whereyields increased at an annual rate of 2.9 percentin 1990–97. Nutrient-poor soil, low irrigation

42 CHAPTER 2

TABLE 2.34 Roots and tubers production

and yields, Sub-Saharan Africa, 1997


(million (kilograms/

metric tons) hectare)

Northern 5.5 6,217

Central and Western 45.4 7,611

Eastern 16.0 6,107

Southern 14.7 5,991

Nigeria 58.3 9,981

All Sub-Saharan Africa 139.9 7,876

Developing world 461.6 11,777


coverage, and weak infrastructure have all ham-pered yield performance regionwide. Sub-Saha-ran African governments have also neglectedresearch into the challenges posed by roots andtubers production, preferring to expend scarceresearch dollars on cash crop production (Scott,Rosegrant, and Ringler 2000).

Area Trends

Given the generally dismal yield performancefor both cereals and roots and tubers, areaexpansion has had to account for most pro-duction growth in Sub-Saharan Africa over thelast 30 years. Cereal area expanded by 31 mil-lion hectares, while roots and tubers areaexpanded by 8 million hectares (Table 2.36).Because population growth was so high dur-ing this period, averaging 2.8 percent annually,per capita cereal area harvested declined from0.17 hectares per capita in 1967 to 0.13 hectaresper capita in 1997. However, the slowing ofpopulation growth and the 3.4 percent expan-sion of area combined to keep per capita cerealarea harvested constant during 1990–97.

Such high levels of area expansion areclearly unsustainable over the long run, andpoor yield performance in most countries and

country groupings may reflect in part theextension of agricultural production systemsinto increasingly less favored areas. Already,efforts to expand cultivation onto previouslyunused land are running up against someharsh realities of the continent: almost 50 per-cent of all land is too arid to sustain direct rain-fed cultivation and 60 percent is subject todrought (World Bank 2000a). The trend ofarea expansion driving overall cereal produc-tion growth seems to have slowed slightlybetween 1990 and 1997, with the rate of areaexpansion declining from 4.4 percent annuallyin 1982–90 to 2.5 percent annually in 1990–97(Table 2.36). The slowing of area expansion inNigeria from 14.5 percent to 2.2 percent in thelatter half of the period accounts for much ofthis decline.

Conclusion

Ultimately, area expansion cannot be the solu-tion to Africa’s agricultural production prob-lems. Yields must grow at relatively high rates,and smallholders must move more rapidly fromsubsistence to commercial production. The lastthree decades have left a legacy of increasingdegradation and desertification, with insuffi-


TABLE 2.35 Growth rates of roots and tubers, Sub-Saharan Africa, 1967–97

Production Yield Area

Region/Country 1967–82 1982–90 1990–97 1967–82 1982–90 1990–97 1967–82 1982–90 1990–97

(percent/year)

Northern 4.4 1.5 3.6 2.7 0.5 0.5 1.6 1.0 3.1

Central and Western 1.8 3.8 2.3 0.3 2.0 0.6 1.5 1.7 1.7

Eastern 5.0 1.2 −1.7 2.4 −1.4 −2.3 2.6 2.7 0.6

Southern 1.8 2.2 4.8 0.7 0.7 2.9 1.1 1.5 1.8

Nigeria −0.6 9.9 7.6 −1.0 4.2 −0.6 0.5 5.5 8.1

All Sub-Saharan Africa 1.8 4.6 3.9 0.4 1.9 0.6 1.4 2.6 3.4


cient investment in soil improvement causing

severe nutrient depletion. Poor policies and

institutional failures have removed the incentive

for farmers to care for the land and to invest

long-term capital in yield-boosting technologies

(World Bank 2000a). Despite the attention given

to Asia over the last decade, the loss of the nat-

ural resource base in Sub-Saharan Africa is of

far greater concern because of the seeming

inability of governments in the region to

address the growing problems they face. Over-

all, the structural and macroeconomic reforms

of the 1980s and 1990s have been a disappoint-

ment as far as food production is concerned, and

greater agricultural research expenditures com-

bined with far-reaching reforms of both infra-

structure and institutions will be necessary to

jumpstart production in the future.

WEST ASIA AND NORTH AFRICA

Basic indicators for food supply and demandand annual growth rates for WANA are pre-sented in Appendix C, Tables C.11 and C.12.

Introduction

The WANA region is highly disparate, with the5 percent of the region’s population who live inoil-exporting nations enjoying far higher percapita incomes than everyone else: outside ofthese oases of wealth, poverty is high and wide-spread, particularly in low-rainfall areas. Allcountries in the region have faced severe chal-lenges in expanding their agricultural sectorsover the 30-year period, including a limitedresource base of arable land and water, low anderratic rainfall with frequent drought, growingpopulations, low productivity growth, andincreased rural-urban migration (Oram et al.

44 CHAPTER 2

TABLE 2.36 Cereal and roots and tubers area, Sub-Saharan Africa, 1967–97

Total area Annual growth rate

Region/Country 1967 1982 1990 1997 1967–82 1982–90 1990–97

(million hectares) (percent/year)

Cereals area

Northern 13.4 17.8 22.1 29.2 1.9 2.7 4.0

Central and Western 6.3 7.9 9.0 10.2 1.4 1.7 1.7

Eastern 4.8 5.2 6.4 6.8 0.5 2.6 0.9

Southern 6.8 7.5 8.4 9.0 0.7 1.5 0.9

Nigeria 10.5 5.2 15.5 18.1 −4.5 14.5 2.2

All Sub-Saharan Africa 41.8 43.6 61.4 73.1 0.3 4.4 2.5

Roots and tubers area

Northern 0.5 0.7 0.7 0.9 1.6 1.0 3.1

Central and Western 3.7 4.6 5.3 6.0 1.5 1.7 1.7

Eastern 1.4 2.0 2.5 2.6 2.5 2.7 0.6

Southern 1.6 1.9 2.2 2.5 1.1 1.5 1.8

Nigeria 2.1 2.2 3.4 5.8 0.4 5.5 8.1

All Sub-Saharan Africa 9.3 11.5 14.1 17.8 1.4 2.6 3.4


1998). Governments have attempted to over-come long-term impediments to higher agri-cultural production mainly through a set of dis-tortional policies that provided for significantproduction growth at the expense of deterio-ration of the natural resource base. Policies toexpand cereal and meat production haveincluded cheap fuel and credit for machinery,high producer and consumer subsidies, andhigh tariffs on imported food products. Whilethe negative effects of these policies on the envi-ronment have become increasingly apparent,economic reform has progressed very slowlyover the past 20 years because of the perceivedneed to keep agricultural import dependenceat manageable levels (Karshenas 1999).

Cereals

Cereal Production. WANA’s agricultural growthperformance over the last three decades hasbeen impressive, although substantial invest-ments made by oil-producing states flush withpetrodollars during the 1980s skewed overallregional growth upward. For example, SaudiArabia has expanded its cereal production sub-stantially through massive investments in irri-gation, using nonrenewable water supplies andheavy application of subsidized fertilizers.Because of the economic unsuitability of theseprojects, costs of production in Saudi Arabia

are several times higher than world prices, evenwithout counting the scarcity price of wateror subsidy costs. Partly as a result of theseunsustainable policies, cereal production per-formance was particularly strong in the OtherWANA country group13 and in Egypt between1982 and 1990 (Table 2.37). Growth slowed inOther WANA between 1990 and 1997, butremained strong in Egypt. Iraq has been amajor drag on the production performance ofOther WANA, with agricultural growthdepressed by constraints linked to the imposi-tion of sanctions. The Saudi Arabian agricul-tural sector has also performed poorly becausefarm subsidies were significantly reduced inthe 1990s (Nordblom and Shomo 1995).

Weather conditions in the region as a wholewere particularly variable during the 1990s,leading to sharp fluctuations in year-to-yearproduction. In the 1990s, Algeria, Egypt,Morocco, the Syrian Republic, Tunisia, andTurkey all adopted aggressive national policiesto expand irrigated area and reduce depen-dence on highly variable rainfed agriculture.Irrigated agriculture does play an importantrole throughout the region, and links betweenirrigated systems, rangelands, and rainfedfarming that operate through the movementof food, feed, livestock, and labor are essential


TABLE 2.37 Growth rates of cereal production and yields, West Asia/North Africa,

1967–97

Production Yields

Region/Country 1967–82 1982–90 1990–97 1967–82 1982–90 1990–97

(percent/year)

Egypt 1.5 5.3 4.6 1.2 3.7 2.4

Turkey 3.1 1.4 1.2 2.7 1.1 0.9

Other West Asia/North Africa 1.5 5.8 1.8 1.0 3.6 2.6

All West Asia/North Africa 2.2 3.8 2.1 1.7 2.3 2.4


to overall agricultural development (Nordblomand Shomo 1995).

Cereal Yields. Between 1982 and 1990, pro-duction not only grew rapidly because ofrapid area expansion, but also because yieldgrowth rates were strong. Between 1990 and1997, yield growth rates continued to rise inEgypt and Other WANA but lagged signifi-cantly in Turkey. Despite this strong yieldgrowth, however, the 1,973 kilograms perhectare average cereal yield in WANA in 1997remained well below the developing worldaverage of 2,312 kilograms per hectare. Rea-sons for continued low yields include the arid-ity of the climate and high variability of pre-cipitation, risk-aversion to adoption of newcultivars and fertilizer application, and con-tinuing resource degradation (Oram et al.1998). In addition, WANA generally possessesvery low stocks of human capital in compari-son with other countries at similar income lev-els: for example, while the region has a simi-lar land-to-labor ratio as Latin America, laborproductivity was significantly higher in LatinAmerica (Karshenas 1999).

Cereal Area. The fact that production in OtherWANA slowed significantly between 1990 and1997 despite strong yield growth indicates thatthese countries have reached the effective lim-its of unutilized area for cereal and livestock pro-duction. The problem is exacerbated by exten-sive environmental degradation that is takingits toll on existing crop- and livestock-produc-ing areas. Wind and water erosion and seedbank loss have been particularly severe on thesteppes, as mechanized production of cereals—particularly barley for feed uses—has expandedonto increasingly marginal land. Rapid urban-ization and the lack of appropriate technologiesfor rangeland management have also taken theirtoll. Population pressures and incentives to

expand cereal production have led to overgraz-ing and cultivation of marginal areas, mono-culture of annual crops, reduced soil fertility,and significant erosion. Additionally, over-pumping of groundwater in many WANAcountries has led to significant soil degradationthrough salinization of irrigation water andlowered water tables (Taimeh 1999).

WANA’s extremely high rates of populationgrowth over the last three decades have createdmany problems for the region and led to rapiddeclines in per capita cereal production through-out the period.14 Massive declines in per capitaharvested area were the main force behind pro-duction declines. Continued decreases in percapita area negatively affected per capita pro-duction levels even as yield growth rates seemedto be gradually increasing (Dyson 1996).

Cereal Demand. Cereal demand growth wasactually most rapid regionwide between 1967and 1982 (Table 2.38). Other WANA had a par-ticularly rapid decline in annual demand growth,falling from 5.6 percent in 1982–90 to 2.3 per-cent in 1990–97. Cereal demand growth ratesalso declined in Turkey, while remainingstrong in Egypt throughout the period. Declin-ing population growth rates—falling to 2.2 percent annually between 1990 and 1997—saturation of demand, and declining oil revenueshelped slow the overall growth of cereal demandin WANA during the 1990s (FAO 2000b).

Cereal Trade. WANA had net cereal imports of5.9 million tons in 1967, placing it toward themiddle of the pack of developing regions interms of cereal import dependence. By 1997,exploding domestic demand had drivenWANA’s net imports to 44.3 million tons, or 52percent of domestic production. Net cerealimports rose from 3.8 million tons in 1967 inOther WANA to 33.8 million tons by 1997(Table 2.39). Egypt also had dramatic import

46 CHAPTER 2

growth. Despite the slowing of demandgrowth, net imports continued to increase inthe 1990s. It is interesting to note that thegrowth rate of WANA’s imports was notstrongly related to the performance of the agri-cultural sector. Cereal imports increased mostrapidly during 1982–90, a period when agricul-tural production, driven by oil-financed invest-ments, was growing at a rate of 3.8 percent peryear (although population growth was alsorapid). This period was extraordinary in that amassive influx of petrodollars permitted bothlarge-scale development of local productioncapacity and large quantities of cereal imports.

Livestock Sector

Cultural constraints on meat consumption aswell as poverty historically have kept meat con-sumption to low levels in WANA. Total meatdemand has actually grown substantially in theregion over the last 30 years, but rapid popula-tion growth has been responsible for much ofthis increase. Fastest growth in total demandoccurred in 1967–82, averaging 6.3 percent forthe region as a whole, slowing to 3.3 percentthereafter (Table 2.40). Turkey and Egypt fol-lowed somewhat different paths, with meatdemand in Turkey growing more rapidly during


TABLE 2.38 Growth rates of cereal demand, West

Asia/North Africa, 1967–97

Region/Country 1967–82 1982–90 1990–97

(percent/year)

Egypt 4.1 3.9 3.5

Turkey 3.0 1.4 1.8

Other West Asia/North Africa 5.2 5.6 2.3

All West Asia/North Africa 4.3 4.1 2.4


TABLE 2.39 Net cereal trade, West Asia/North Africa,

1967–97

Region/Country 1967 1982 1990 1997


Egypt −2.0 −7.1 −7.9 −9.4

Turkey −0.1 0.8 −0.5 −1.1

Other West Asia/North Africa −3.8 −22.6 −30.3 −33.8

All West Asia/North Africa −5.9 −28.9 −38.7 −44.3

––––Source: Based on FAOSTAT data (FAO 2000a).Note: Positive figures indicate net exports; negative figuresindicate net imports.

1982–90, whereas it did not take off in Egypt until1990–97. Per capita meat demand regionwideincreased from 12.1 kilograms per capita in 1967to 21.0 kilograms per capita in 1997. In general,growth has slowed significantly since 1967–82.

Conclusion

The problems confronting agricultural pro-duction in WANA are certainly substantial, butit would be a mistake to see them as inevitable.Poor government policies have worsened a sit-uation already characterized by unfavorableinitial endowments of land and water. Policiesencouraging misuse of rangeland areas and theextension of crop production onto fragile rain-

fed soils have been guided by a desire toimprove food security, but they will destroy thenatural resource base over the long run. Askewed system of land distribution and inse-cure property rights are also important imped-iments to long-term sustainable agriculturalgrowth (Oram et al. 1998). While governmentsin the region have taken some steps to improveagricultural policymaking over the last twodecades, the region’s high level of importdependence renders further major reformpolitically problematic. The backsliding of eco-nomic reform efforts in Libya and Turkey dur-ing the 1990s is evidence that the success ofongoing reform is in no way guaranteed.

48 CHAPTER 2

TABLE 2.40 Growth rates of meat demand, West

Asia/North Africa, 1967–97

Region/Country 1967–82 1982–90 1990–97

(percent/year)

Egypt 4.8 3.0 6.8

Turkey 2.8 6.3 0.7

Other West Asia/North Africa 7.9 3.0 3.2

All West Asia/North Africa 6.3 3.5 3.3


49

3

THE IMPACT MODEL

IFPRI’s IMPACT model offers a methodologyfor analyzing alternative scenarios for globalfood demand, supply, and trade. IMPACT cov-ers 36 countries and regions, accounting forvirtually all of the world’s food production andconsumption, and 16 commodities, includingall cereals, soybeans, roots and tubers, meats,milk, eggs, oils, oilcakes, and meals. (For acomplete list of countries and commodities,see Appendix A.) IMPACT represents a com-petitive agricultural market for crops and live-stock. It is specified as a set of country orregional submodels, within each of which sup-ply, demand, and prices for agricultural com-modities are determined. The country andregional agricultural submodels are linkedthrough trade, a specification that highlightsthe interdependence of countries and com-modities in global agricultural markets. Themodel uses a system of supply and demandelasticities, incorporated into a series of linearand nonlinear equations, to approximate theunderlying production and demand functions.World agricultural commodity prices aredetermined annually at levels that clear inter-

national markets. Demand is a function ofprices, income, and population growth.Growth in crop production in each country isdetermined by crop prices and the rate of pro-ductivity growth. Its component sources,including crop management research, con-ventional plant breeding, wide-crossing andhybridization breeding, and biotechnologicaland transgenic breeding contribute to theestimate of future productivity growth. Othersources of estimated growth include privatesector agricultural research and development,agricultural extension and education, roads,and irrigation.

In order to examine the effects on food secu-rity, the percentage and number of malnour-ished preschool children (0 to 5 years old) areprojected for developing countries. The pro-jected numbers of malnourished children arederived from the estimated relationshipbetween the percentage of malnourished chil-dren and average per capita calorie consump-tion, the percentage of females with access tosecondary education, the quality of maternaland child care (proxied by the status of women

relative to men as captured by the ratio offemale to male life expectancy at birth), andhealth and sanitation (proxied by the percent-age of the population with access to treatedsurface water or untreated but uncontami-nated water from another source).

A wide range of factors with potentially sig-nificant effects on future developments in theworld food situation can be modeled based onIMPACT. They include population and incomegrowth, the rate of growth in crop and live-stock yield and production, feed ratios for live-stock, agricultural research, irrigation andother investment, price policies for com-modities, and elasticities of supply anddemand. For any specification of these under-lying factors, IMPACT generates projectionsfor crop area; yield; production; demand byfood, feed, and other uses; prices; and trade;and for livestock numbers, yield, production,demand, prices, and trade. (Supplementarydata for each crop and livestock category for1997 and 2020 are presented in Appendix D,Tables D.1 to D.19.)

BASELINE ASSUMPTIONS

All simulations in this study begin from a baseyear of 1997. The base-year data on produc-tion and utilization, extracted from the FAOAgrostat database (FAO 2000a), is a three-yearaverage centered on 1997. Key parametersdefined in the model include the price andincome elasticities of demand; price elastici-ties of area and yield; animal feed ratios; andgrowth rates of population, income, croparea, yield, and livestock numbers and yields.These parameters are the primary drivers inthe projections of global and regional foodmarkets.

Population and Income Growth

and Urbanization

Population and income growth will remainimportant determinants of food supply anddemand balances. A world population of 3 bil-lion people in 1960 doubled to 6 billion by1999, with population growth rates peaking at2.1 percent annually between 1965 and 1970,and declining progressively since then to 1.4percent annually between 1997 and 1998 (UN1990). Further declines in global populationgrowth rates are likely, with birth rates fallingin many regions and declines in mortality ratesleveling off.15 The population of Sub-SaharanAfrica, estimated at 561 million in 1997, is mod-eled over separate five-year periods, increasingat a rate of 2.4 percent per year between 1997and 2020—with rates declining in later peri-ods, as they do for all regions—to a populationin 2020 of about 959 million people. In otherregions, Pakistan is expected to have a partic-ularly high population growth rate, projectedat 2.8 percent per year, while growth rates inthe two world giants—India and China—willaverage only 1.2 and 0.7 percent per year,respectively. Nevertheless, because the popu-lation bases of these countries are so large,they will still account for about 32 percent ofthe total world population increase during1997–2020. Population growth rate assump-tions for all countries and regions covered inthe model are shown in Table 3.1.

GDP growth rates between 1997 and 2020are also shown in Table 3.1. GDP growth ratedisparities among countries in the developingworld are projected to remain high. Growthrates will be highest in East and Southeast Asia,ranging from 3.5 to 6.0 percent per year.Strongly reformist initiatives undertaken inIndia involving strong implementation ofmacroeconomic stabilization and marketreforms account for a projected improvement

50 CHAPTER 3

in income growth of 5.8 percent per year.China’s economy, which has shown signs ofslowing in recent years from a GDP growthrate that averaged nearly 10 percent per yearover the last few decades, is projected to stabi-lize at an annual rate of growth of 6 percent.

A number of reforming countries in LatinAmerica, including Chile and Argentina, haveshown improved economic prospects thatshould continue into the future, through theattraction of large inflows of foreign invest-ments. However, the potential for renewedmacroeconomic instability in the region leadsto a cautious estimate of annual regional GDPgrowth of 3.6 to 4.5 percent over the projec-tion period.

GDP growth rates in Sub-Saharan Africanhave been dismal over the last few decades:they have also been highly variable because somany countries in the region depend on rain-fed agriculture and uncertain governance.Nevertheless, regional GDP growth hoveredaround 3 percent between 1990 and 1997, andthe region possesses considerable scope forrecovery. Many governments in the regionhave adopted reformist agendas to improvemacroeconomic stability and foster develop-ment of the infrastructure and institutions nec-essary to generate sustained, long-termgrowth. Nigeria’s return to democratic ruleprovides hope for renewed growth led by private sector development. Integration of theSouth African economy as a regional marketis an ongoing process, which should continueto have beneficial spillover effects on the restof Southern Africa. Nevertheless, the preva-lence of HIV/AIDS and continued social andpolitical strife give plenty of reason for cautionregarding long-term growth rates. GDPgrowth for Sub-Saharan Africa is projected tobe 3.2 to 3.8 percent during 1997–2020.

Signs of progress toward more economicstabilization, restructuring, and recovery are

mixed in Eastern Europe and FSU. Some coun-tries in Eastern Europe (Poland and Hungary,for example) have experienced a turnaroundfrom the negative GDP growth rates of theearly 1990s to strongly positive rates in the late1990s. Their progress stands in stark contrastto Russia and most other countries in FSU,where prospects remain rather grim. Pro-jected GDP growth rates are 4.0 percent peryear in Eastern Europe and 2.0 to 3.0 percentper year in the remaining FSU countries.Growth rates in the rest of the developedworld are projected to range from 2.2 to 2.7percent per year.

Closely related to population and incomechanges is the transformation of demographicpatterns. The most vital of these demographiccharacteristics, particularly in terms of project-ing future food needs in fast-growing econo-mies, is the rate of urbanization. Rural-to-urbanmigration—and its attendant significant effecton demand structures—has increased quite rap-idly over the last few decades throughout thedeveloping world and will continue to do so overthe projections period (World Bank variousyears). Approximately 46 percent of the popu-lation of developing countries resided in urbanareas in 1998, up from 22 percent in 1960 and 30percent in 1980 (World Bank 2000b). Past stud-ies have shown that urbanization accelerateschanges in diet away from basic staples such assorghum, millet, maize, and root crops, to cere-als requiring less preparation (such as wheat),fruits, livestock products, and processed foods.As described in the previous chapter, in the moredeveloped countries of Asia, urbanization hasresulted in substantial increases in demand formeat and other livestock products and a shiftfrom rice to wheat. In Sub-Saharan Africa,urbanization should encourage a shift awayfrom staple foods such as coarse grains and rootsand tubers toward higher consumption of wheatand rice. While the concept of “urbanization” is

THE IMPACT MODEL 51

TA

BL

E 3

.1

IM

PA

CT

po

pu

lati

on

an

d g

ross

do

mes

tic

pro

du

ct (

GD

P)

assu

mp

tio

ns,

199

7 an

d 2

020

Po

pula

tio

nG

DP

gro

wth

rat

esP

er c

apit

a G

DP

Regi

on/C

ount

ry1

99

72

02

0(1

99

7–

20

20

)2

02

0

(mil

lion

s)(p

erce

nt/

yea

r)(U

S$

)

Dev

elop

ed w

orld

1,29

7.3

1,36

0.6

2.4

29,2

08

Uni

ted

Stat

es27

1.6

317.

02.

745

,585

EU

1537

3.1

371.

12.

237

,032

Japa

n12

5.6

123.

52.

255

,990

Aus

tral

ia18

.222

.32.

530

,914

Oth

er d

evel

oped

cou

ntri

es94

.411

1.2

2.2

19,7

08

Form

er S

ovie

t Uni

on29

3.1

296.

82.

12,

910

Eas

tern

Eur

ope

121.

211

8.7

4.0

7,47

4

Cen

tral

Asi

a54

.969

.53.

01,

465

Res

t of F

orm

er S

ovie

t Uni

on23

8.2

227.

32.

03,

352

Dev

elop

ing

wor

ld4,

490.

66,

095.

74.

62,

455

Latin

Am

eric

a48

6.0

652.

13.

86,

869

Mex

ico

94.1

124.

93.

66,

659

Bra

zil

163.

220

9.3

3.6

8,52

6

Arg

entin

a35

.745

.44.

517

,438

Col

ombi

a37

.252

.63.

84,

652

Oth

er L

atin

Am

eric

a15

5.8

219.

93.

63,

759

Sub-

Saha

ran

Afr

ica

560.

995

8.6

3.6

443

Nig

eria

103.

916

8.2

3.8

544

Nor

ther

n14

2.7

253.

03.

321

9

Cen

tral

and

Wes

tern

13

7.8

246.

43.

853

2

Sout

hern

83

.513

5.9

3.2

648

Eas

tern

93

.015

5.0

3.5

377

Wes

t Asi

a/N

orth

Afr

ica

338.

850

5.0

3.8

3,47

5

Egy

pt64

.590

.24.

02,

057

Tur

key

62.9

83.6

4.5

6,27

2

Oth

er W

est A

sia/

Nor

th A

fric

a21

1.4

331.

23.

53,

155

Asi

a3,

101.

33,

974.

45.

52,

082

Sout

h A

sia

1,28

8.5

1,78

0.3

5.5

1,05

2

Indi

a96

0.1

1,26

6.0

5.8

1,20

2

Paki

stan

144.

124

4.3

4.5

715

Ban

glad

esh

122.

416

9.8

4.8

715

Oth

er S

outh

Asi

a61

.910

0.2

4.5

549

Sout

heas

t Asi

a49

2.4

648.

74.

92,

515

Indo

nesi

a20

3.2

262.

34.

51,

879

Tha

iland

59.3

70.5

5.2

6,70

8

Mal

aysi

a20

.929

.25.

210

,012

Phili

ppin

es70

.710

1.6

5.0

2,31

9

Vie

t Nam

76.1

102.

55.

891

4

Mya

nmar

46.2

58.8

4.0

335

Oth

er S

outh

east

Asi

a15

.923

.74.

01,

027

Eas

t Asi

a1,

320.

41,

545.

55.

73,

087

Chi

na1,

249.

01,

461.

46.

02,

330

Sout

h K

orea

45.6

51.9

5.0

26,0

15

Oth

er E

ast A

sia

25.8

32.3

3.5

555

Res

t of w

orld

36.5

5.6

4.0

5127

Wor

ld5,

787.

97,

456.

42.

97,

337

––––

Sou

rce:

GD

P as

sum

ptio

ns b

ased

on

Wor

ld B

ank

(200

0b);

popu

latio

n as

sum

ptio

ns b

ased

on

UN

(199

8).

not explicitly incorporated in the model, itseffects are reflected in the assumptions ofincome and price elasticities, as described in thenext section.

Demand and Supply Elasticities

The income elasticities of demand clearly vin-dicate the underlying assumption of the modelthat there will be a gradual shift in the demandstructure from the main staples to high-valueproducts such as meat and other livestockproducts, particularly in the developing coun-tries. Several factors give rise to this hypothe-sis, including expected increases in per capitaincomes from economic growth, rapid urban-ization, and the continued commercializationof agricultural production (Bouis 1994). TheIMPACT income demand parameters repre-sent a synthesis of average, aggregate incomeelasticities for each country, given the incomelevel and distribution of population betweenurban and rural areas as they evolve over theprojection period (Table B.6).

In 1997, income food demand elasticities forlivestock products were particularly high forpoultry and beef products in Asia, with theexception of beef in South Asia. Poultry elas-ticities will decline by a third or more in Asiaby 2020. Beef income elasticities in 1997 werealso very high in Sub-Saharan Africa. In LatinAmerica, where demand for beef is alreadyquite high in many countries, the income elas-ticity is already low and expected to decreasefurther by 2020.

Pork and sheep and goat meat income elas-ticities of demand are expected to be lowerthan beef demand elasticities across almost allregions in both 1997 and 2020. Pork elastici-ties are relatively high in East Asia and Sub-Saharan Africa in 1997, declining slightly by2020, while they will remain low in South Asia.Sheep and goat elasticities are about the sameacross all developing regions except East Asia,

where they are fairly high. Income demandelasticities for poultry are significantly higherthan those for pork and beef in developedcountries. Income demand elasticities in theFSU countries are comparable to those in thedeveloped countries for all meats except forsheep and goat meat, for which they are sig-nificantly higher.

Cereal food demand elasticities will besignificantly lower than meat elasticitiesthroughout the developing world, reflecting the ongoing increase in the share of meat in developing-country diets. Wheat and rice elas-ticities are low throughout Asia, but relativelyhigh in Sub-Saharan Africa, with virtually nodecline by 2020. In the rest of the world, wheatand rice income elasticities are lower than inSub-Saharan Africa, with rice income elastici-ties slightly higher than wheat elasticities inLatin America, WANA, and most of the devel-oped world. Maize and other grains are infe-rior goods throughout the world, with low ornegative income elasticities. Sub-SaharanAfrica is the exception.

Among the roots and tubers, marginaldemand for potatoes will remain high amongthe developing countries, with slightly lowerbut still moderate elasticities in the developedcountries. Income elasticities for sweet pota-toes and cassava and other roots and tubers aremuch lower, with sweet potatoes inferiorgoods in South Asia and in East Asia. Sweetpotato elasticities are low in all other develop-ing regions and negative in most developedcountries. The trends for cassava and otherroots and tubers are similar, with income elas-ticities in 1997 negative and declining in LatinAmerica and in East Asia. Demand elasticitiesfor cassava will also be negative in most devel-oped countries.

Egg income demand elasticities remainstrong in all regions, with particularly highvalues in 1997 in South Asia, East Asia, and

54 CHAPTER 3

Southeast Asia, declining by 2020. Incomeelasticities in the rest of the developing worldin 1997 range from 0.21 in Latin America to0.46 in Sub-Saharan Africa, declining to 0.11and 0.32 by 2020 and actually turning nega-tive in a number of developing countries. Sub-Saharan Africa also has the highest milkdemand elasticity.

Supply Response

Supply elasticities include yield elasticities withrespect to the crop price, labor price, and capi-tal price (including fertilizer and other re-current input prices), and crop area and live-stock numbers elasticities with respect to own-commodity price and competing commodityprices. The absolute value of cross-price areaelasticities with respect to other crops generallysums to an average across all regions of betweenone-quarter and one-half of the own-price elas-ticity for the given crop (Tables B.7–B.10).

Own-price crop area elasticities (and live-stock number elasticities) are generally fairlysmall. In the developing world, pork probablyhas the most widely varying elasticity, rangingfrom 0.13 in WANA to 0.42 in Latin America.Pork number elasticities in the developedworld range from 0.33 in Japan to 0.45 in theEU15. In the developing world, beef numberelasticities are also high, but somewhat lessvariable than pork elasticities. Beef elasticitiesin the developed world are similar to those forpork. Sheep and goat and poultry own-priceelasticities of supply are clustered between 0.2and 0.4, with elasticities in the developed worldslightly higher in both cases.

Cereal own-price area elasticities areexpected to be lower than livestock numberelasticities. Wheat area elasticities will rangefrom 0.10 in South Asia to 0.21 in Sub-SaharanAfrica and Latin America; maize area elastici-ties will range from 0.12 in Southeast Asia andEast Asia to 0.2 in WANA; other grain area elas-

ticities will range from 0.11 in Southeast Asiato 0.26 in Sub-Saharan Africa; and rice elastic-ities will range from 0.10 in Southeast and EastAsia to 0.16 in WANA. Elasticities in the devel-oped world will, for the most part, lie at thehigh end of these ranges.

Own-price elasticities for the area of rootsand tubers will be similar to cereal elasticities,ranging from 0.10 to 0.18. Egg number elas-ticities range from 0.28 in Sub-Saharan Africato 0.37 in WANA, and milk supply elasticitiesvary from 0.32 in Latin America to 0.40 inWANA, with similar elasticities for both cropsin the developed world.

Own-price area elasticities of supply formost products in the FSU countries are approx-imately two-thirds of those in the developedcountries, reflecting the difficulties that pro-ducers in the FSU will continue to have inrationalizing agricultural production in theaftermath of the collapse of centralized pro-duction systems.

Crop prices also affect yields, reflecting thegreater incentive to farmers of generatinghigher yields at higher crop prices. Own-priceyield elasticities for wheat range from 0.13 inSoutheast Asia to 0.19 in Sub-Saharan Africa,while maize and rice elasticities have a slightlylower if similar spread. Other grain (coarsegrain) yields are less responsive to pricechanges than the other three cereal crops, witha range of elasticities from 0.09 in SoutheastAsia to 0.14 in Sub-Saharan Africa. Own-priceyield elasticities for roots and tubers crops willbe similar to those for cereals.

Capital and labor are the two other mainvariables for which yield elasticities are speci-fied. The absolute values of yield elasticitieswith respect to capital and labor sum to thecrop price elasticity. The yield elasticity withrespect to capital is generally higher than thatof labor in the developed countries, while theopposite holds true in the developing countries.

THE IMPACT MODEL 55

56 CHAPTER 3

COMMODITY PRICES AND TRADE POLICY

Agricultural commodity prices are endoge-nous in the IMPACT model. Domestic pro-ducer and consumer prices are a function ofworld prices (expressed in the respective country-group currencies via the exchangerate to the U.S. dollar), producer subsidyequivalents (PSEs), consumer subsidy equiv-alents (CSEs), and marketing margins. Theeffects of country- and region-specific tradeand price policies are thus expressed in termsof trade-distorting PSEs and CSEs, and mar-keting margins between the world price andproducer and consumer prices. PSEs and CSEsmeasure the level of taxation or subsidy borneby producers or consumers relative to worldprices and account for the wedge betweendomestic and world prices (see Box 3.1). Mar-keting margins reflect factors such as trans-port costs. In the model, PSEs, CSEs, and mar-keting margins are expressed as percentagesof the world price.

The world price of a commodity is theequilibrating mechanism such that when anexogenous shock is introduced in the model,the world price will adjust and each adjustment

is passed back to the effective producer andconsumer prices via price transmission equa-tions. Changes in domestic prices subsequentlyaffect commodity supply and demand, neces-sitating iterative readjustments until worldsupply and demand balance and world nettrade is again equal to zero.

Comprehensive estimates of trade-distortingPSEs and CSEs for agricultural commoditiesare available only for OECD countries. Forother countries and regions, PSEs and CSEs areproxied by estimates of nominal or effectiveprotection rates or tariffs. The estimated PSEsand CSEs utilized in the IMPACT model aresynthesized from a variety of sources16 and aresummarized in Appendix B, Tables B.11–B14.

PRODUCTIVITY AND AREA GROWTH

One of the main assumptions for nonprice sup-ply (area and yield) projections is that they can-not be constrained to be constant over the pro-jections period, so growth rate estimates areindividually specified over five-year periods:1997–2000, 2001–05, 2006–10, 2011–15, and2016–20. Projections are estimated to be con-sistent with historical experience in terms of

BOX 3.1 PSEs and CSEs

According to OECD (1999):

• PSEs are “an indicator of the annual monetary value of gross transfers from consumers andtaxpayers to support agricultural producers, arising from policy measures which supportagriculture.”

• CSEs are “an indicator of the annual monetary value of gross transfers to consumers ofagricultural commodities, arising from policy measures which support agriculture, regardlessof their nature.”

The IMPACT model operationalizes these measures as proportianal price wedges between world anddomestic prices. Thus, a PSE of 0.10 indicates that prices for domestic producers are 10 percent higherthan world prices (hence, a subsidy to producers), while a CSE of –0.10 indicates that prices for domesticconsumers are 10 percent higher than world prices (hence, a tax on consumers).

THE IMPACT MODEL 57

slowing rates of public investment in agricul-tural research and rural infrastructure, includ-ing roads. Thus, projected future yield trendsexplicitly account for the slowdown in yieldgrowth that has occurred across most com-modities in most regions over recent decades.The growth contribution of modern inputs suchas fertilizers is accounted for in price effects inthe yield response function and as a comple-mentary input with irrigation and modern vari-

eties generated by research. The methodologymakes use of ex-post and ex-ante studies of agri-cultural research priority setting, syntheses ofsources of agricultural productivity growthstudies, examinations of the role of industrial-ization on growth, and “expert opinion” to gen-erate the projected time path of yield growth.Yield growth projections also account forexpected effects on yields of environmentaldegradation.

BASELINE PROJECTIONS TO 2020

Cereals

Demand. Total cereal demand will increase by1.3 percent per year between 1997 and 2020,particularly in developing countries, but it willdecrease from historic rates because popula-tion growth rates are slowing and income elas-ticities of food demand for cereals are gradu-ally declining in many countries. Nevertheless,

the amount of additional cereal needed tomeet effective demand by 2020 is as large as theincrease during the previous 23 years. Between1974 and 1997, global cereal demand grew bynearly 636 million tons, developing-countrycereal demand by 553 million tons, and devel-oped-country cereal demand by 83 milliontons. By comparison, cereal demand in devel-oping countries is projected to grow by 49 per-cent (557 million tons) between 1997 and 2020,

4

PROJECTIONS OF GLOBAL

FOOD SUPPLY AND DEMAND

AND CHILD MALNUTRITION

58

FIGURE 4.1 Total cereal demand by region, 1997 and 2020

Source: IMPACT projections, June 2001.

while cereal demand in the developed world isprojected to increase by 13 percent (97 milliontons) (Figure 4.1). Developing Asia will accountfor 52 percent (344 million tons) of the globalincrease in cereal demand between 1997 and2020, with China alone accounting for 26 per-cent (173 million tons). India will account for

an additional 12 percent (78 million tons) ofworldwide cereal demand growth, and otherdeveloping Asian countries will account for 14percent (92 million tons) (Figure 4.2).

Cereal Demand Composition. The compositionof cereal demand will change significantlybetween 1997 and 2020 as incomes and urban-ization rise, especially in Asia. In terms of totalcereal demand, rice and wheat were the majorcereal crops in developing countries in 1997,accounting for 33 and 30 percent, respectively,of total cereal demand (Figure 4.3). Maize willovertake these two cereals by 2020, rising from26 percent of total cereal demand in 1997 to 30percent in 2020, while the shares of both riceand wheat will decline to 29 percent. In Asia,rice accounted for 43 percent of total cerealdemand in 1997 but will only account for 39percent in 2020. Maize will be the beneficiaryof this decline in relative rice demand, with itsshare rising from 22 percent in 1997 to 28 per-cent in 2020.

Per capita food demand for the various cere-als reveals a slightly different story than aggre-

PROJECTIONS OF GLOBAL FOOD SUPPLY AND DEMAND 59

FIGURE 4.2 Share of regions in cereal

demand increase, 1997–2020

Source: IMPACT projections, June2001.

FIGURE 4.3 Cereal demand composition by crop, 1997

and 2020


gate demand, as per capita food demand formaize in the developing world remains con-stant between 1997 and 2020, while per capitafood demand for wheat increases 6 percent (4kilograms per capita) (Table 4.1). Per capitafood demand for rice declines 1 percent anddemand for other coarse grains increases 17percent. In fact, while per capita food demandfor other coarse grains declines in all regionsexcept Sub-Saharan Africa, where it willincrease 11 percent above 1997 levels, the largeincrease in per capita demand worldwide is theresult of the sharp increase in consumption inSub-Saharan Africa (49 kilograms per capita by2020).

Per capita food demand for rice in Asia willstay relatively constant, increasing by only 1.8percent, but per capita food demand for maizewill decline by a significant 16 percent from1997 levels. Much of this decline comes froma shift in consumption to wheat, which isexpected to experience demand growth of 9percent between 1997 and 2020. As noted ear-lier, rapidly growing incomes and urbanizationwill drive this consumption shift.

Feed Demand for Cereals. Animal feed use, drivenby rising meat demand, will be responsible formost of the increasing demand for maize indeveloping countries as a whole and Asia in par-ticular between 1997 and 2020. Whereas feedaccounted for 21 percent of total worldwidecereal demand in 1997, it will account for 35 per-cent of the cereal demand increase between1997 and 2020 (Figure 4.4). Feed demand forcereals is projected to grow by 2.5 percent peryear in developing countries and 0.6 percent peryear in developed countries, accounting for totalfeed demand in-creases of 84 percent in thedeveloping world and 16 percent in the devel-oped world. That will amount to an increase of198 million tons in the developing world and 68million tons in the developed world (Table 4.2).Sub-Saharan Africa will have particularly rapidgrowth in demand for cereal feed at 2.8 percentannually, resulting in a total increase of 98 per-cent. Feed demand for maize in the developingworld will lead all other cereal crops with a rateof increase of 2.9 percent annually, for a totalincrease of 92 percent, and it will grow at aneven faster 3.1 percent per year in Asia, for anoverall increase of 111 percent.

60 CHAPTER 4

TABLE 4.1 Per capita food demand for cereals by crop and region,

1997 and 2020

Developing Developed

world world Asia World

Crop 1997 2020 1997 2020 1997 2020 1997 2020

(kilograms/capita)

Wheat 64.2 67.7 99.7 103.8 63.8 70.0 72.1 74.3

Maize 18.9 18.6 11.9 11.4 11.1 9.3 17.3 17.3

Rice 72.0 70.9 11.0 11.4 94.7 96.4 58.3 60.0

Other coarse 12.0 13.5 8.8 7.4 8.0 7.5 11.3 12.4

grains

––––Source: IMPACT projections, June 2001.

Cereal Production. Global cereal production isprojected to grow at a rate of 1.26 percent peryear during 1997–2020, slightly lower than the1.32 percent per year increase in demand. Thedifference is caused by the slightly higher levelof cereal production, compared with demandin the 1997 base year, due to the net accumula-tion of global cereal stocks in that year. The netincrease in stocks in the base year is drawn downduring the first three years of the projections

period to achieve long-run equilibrium, that is,a supply-demand balance with no annualchange in stocks. Regional production increaseswill not satisfy rising Asian cereal demand, andEast Asian demand in particular will seriouslyoutstrip production (Figure 4.5). The only devel-oping region projected to have surplus cerealgrowth between 1997 and 2020 is Latin America,with a surplus of 10 million tons. The developedworld, with surplus cereal growth of 73 milliontons between 1997 and 2020, will provide thebulk of excess cereal production. Asia’s share ofworldwide cereal production will increase 2 per-cent to reach 41 percent by 2020. Meanwhile,Sub-Saharan Africa’s share will increase from 4to 5 percent and Latin America’s from 7 to 8 per-cent. The share of the developed countries intotal world cereal production will declinethrough 2020.

Cereal Area. Sub-Saharan Africa and LatinAmerica will be able to increase their share ofworldwide cereal production because they willexpand cereal area significantly between 1997and 2020, and will therefore not be as depen-dent as the rest of the world on cereal yield


FIGURE 4.4 Share of food, feed, and other uses in total cereal demand

of developing countries, 1997 and 1997–2020 increase


TABLE 4.2 Feed demand for cereals by

region, 1997 and 2020

Region 1997 2020


Latin America 57.7 98.1West Asia/North Africa 35.9 59.4Sub-Saharan Africa 3.9 7.7South Asia 2.9 6.4East Asia 118.9 233.2Southeast Asia 15.1 27.1Developed world 425.0 492.6Developing world 234.5 432.0––––Source: IMPACT projections, June 2001.

increases to drive production growth (Figure4.6). Area under cereal production in Sub-Saharan Africa is projected to expand 27 per-cent, with area expansion responsible for 32percent of total production growth. Area undercereal production will also expand by a signifi-cant 15 percent in Latin America, where it willbe responsible for 23 percent of production

growth during 1997–2020. These two regionswill be exceptions, as cereal area is projecteto expand only 9 percent between 1997 and

2020 in the developing world as a whole, pro-viding for only 15 percent of cereal productiongrowth during this period. Asia, in particular,possesses little arable land not already undercultivation.

62 CHAPTER 4

FIGURE 4.5 Cereal production and demand increases by region, 1997–2020


FIGURE 4.6 Share of area and yield increase in regional cereal production growth, 1997–2020


Cereal Yields. Despite the importance of yieldsto overall growth in cereal production, yieldgrowth rates will continue to slow across allcereals and all regions, with the notable excep-tion of Sub-Saharan Africa, where yields areprojected to recover from the 1982–97 periodof stagnation (Figure 4.7). The pattern ofgrowth of cereal yields began to slow signifi-cantly after 1982. In the developed world, theslowdown in crop area, yield, and productiongrowth was primarily policy induced, as NorthAmerican and European governments drewdown cereal stocks and scaled back farm price-support programs in favor of direct paymentsto farmers. The economic collapse and subse-quent economic reforms in the former cen-trally planned economies in Eastern Europeand the FSU further depressed crop produc-tion for developed countries as a whole. Anumber of factors have contributed to theslowing of cereal productivity growth in devel-oping countries, particularly in Asia, since theearly 1980s. Increased intensity of land use—and the high levels of input use alreadyachieved in much of Asia—has led to higher

and higher input requirements in order to sus-tain yield gains. Public investment in cropresearch and irrigation infrastructure has alsoslowed considerably, with consequent effectson yield growth.

These forces are expected to continue toslow cereal yield growth rates from 1.6 percent annually in 1982–97 to 0.9 percentannually in 1997–2020, with cereal yieldsincreasing 25 percent (0.7 tons per hectare)worldwide during this period. In the develop-ing world, cereal yield growth rates areexpected to decline from 1.9 percent annuallyin 1982–97 to 1.1 percent annually in 1997–2020, with average cereal yields increasing atotal of 32 percent (0.7 tons per hectare) dur-ing this period. Because high-yielding varietiesof maize were introduced later in developingcountries, maize will have the fastest yieldgrowth of any cereal crop in 1997–2020 bya total of 43 percent (1.2 tons per hectare)(Figure 4.8). Nevertheless, even maize yieldgrowth rates in the developing world willdecline from 2.2 to 1.6 percent annually.


FIGURE 4.7 Yield growth rates by region, all cereals, 1967–82, 1982–97, and 1997–2020


The one region with improving cereal yieldgrowth rates, Sub-Saharan Africa, will experi-ence yield growth of 1.5 percent annuallybetween 1997 and 2020 for a total yield increaseof 46 percent, thus substantially surpassing theyield growth rate of 0.1 percent achievedbetween 1990 and 1997. Cereal yields in thedeveloped world are expected to grow at 0.7percent per year for a total increase of 18 per-cent between 1997 and 2020, with growth ratesdeclining from 1.1 percent per year between1990 and 1997.

Cereal Prices. International cereal prices areprojected to decline between 1997 and 2020,although these declines will slow significantlyfrom those achieved in the recent past.Between 1982 and 1997, real world wheatprices declined by 28 percent, rice prices by 29percent, and maize prices by 30 percent.

During 1997–2020, wheat prices are pro-jected to decline 8 percent; rice prices, 13 per-cent; other grain prices, 11 percent; and maizeprices to remain fairly constant with a percentdrop (Table 4.3). These price declines willreally only begin to take effect after 2010. Infact, between 1997 and 2010, wheat prices are

projected to decline by 3 percent, maize andrice prices to stay constant, and other grainprices to decline by 4 percent. This tighter pre-dicted price scenario indicates that additionalshocks to the agricultural sector, particularlyfailure to meet demands for agricultural waterand other inputs, could put serious upwardpressure on food prices.

Cereal Trade. Rapid growth in world cerealtrade will accompany slow declines in cerealprices, as the United States and the EuropeanUnion increase their cereal exports to keeppace with rising world demand. U.S. cereal

64 CHAPTER 4

FIGURE 4.8 Cereal yields by crop in developing countries, 1997 and 2020


TABLE 4.3 World prices by cereal

crop, 1997 and 2020

Crop 1997 2020

(US$/metric ton)

Wheat 133 123

Maize 103 102

Rice 285 250

Other coarse 97 86

grains


exports will increase from 77 million tons in1997 to 120 million tons in 2020, while EU15cereal exports will increase from 20 to 29 mil-lion tons (Figure 4.9). Southeast Asia’s largenet exports of relatively high-value rice maskits rising volume of net imports of total cere-als, which are expected to increase from 3 mil-lion to 9 million tons of cereal imports by 2020,whereas the value of cereal trade will shiftfrom minor net imports in 1997 to net exportsvalued at US$0.8 billion in 2020 (Table 4.4).

Chinese import markets will absorb muchof the increase in cereal exports from theUnited States and the EU15, consuming a netof 48 million tons of imported cereal in 2020,or 40 million tons more than in 1997. Netcereal imports will undergo the most dramaticshift in South Asia, increasing 18 million tonsabove a slight trade deficit in both volumetricand value terms in 1997. WANA, the largestcereal importer in the world in both 1997 and2020, will require 73 million tons of imported

grain in 2020. Net cereal imports into Sub-Saharan Africa will rise a hefty 15 million tonsbetween 1997 and 2020, and the region willrequire 27 million tons of cereal imports to sat-isfy demand in 2020.

Meat

Meat Demand. Rising meat demand will driveincreased cereal feed demand. Global meatdemand, 208 million tons in 1997, is projectedto grow 57 percent (118 million tons) by 2020(Table 4.5). Meat demand in the developingcountries, rising a projected 92 percent (102million tons) between 1997 and 2020, willaccount for the lion’s share of the increase inglobal demand, and Asia, led by China, will inturn account for the major share of theincrease in meat demand in the developingcountries. In fact, China alone will account for43 percent of additional meat demand world-wide between 1997 and 2020. India, followingits traditional pattern of low meat consump-


FIGURE 4.9 Net cereal trade by region, 1997 and 2020

Source: IMPACT projections, June 2001.Note: Positive figures indicate net exports; negative figures indicate net imports.

tion, will only account for 4 percent of addi-tional meat demand. The rest of Asia, how-ever, will exhibit strong growth, accounting for13 percent of the global increase. Meat priceswill remain relatively stable throughout theprojection period.

While per capita meat demand in the devel-oping world will increase substantially between1997 and 2020, per capita meat consumption inthe developing countries in 2020 will still be farbelow that in the developed countries (Figure4.10). While per capita livestock demand in thedeveloped world will be 84 kilograms in 2020,livestock demand in the two highest meat-con-suming developing regions, Latin America andEast Asia, will only reach 69 and 70 kilogramsper capita, respectively. Other developing regionswill have much lower per capita meat demand.

Despite very rapid growth, Chinese meatdemand will only surpass the per capita meatdemand of one developed country, Japan, by2020, reaching 71 kilograms per capita toJapan’s 52 kilograms per capita. Japan has thelowest meat consumption of any developedcountry, largely because its consumption offish is exceptionally high.

While demand for all meat products willincrease substantially over the IMPACT pro-jections period, the demand for poultry willincrease the most, rising 131 percent in thedeveloping world and 34 percent in the devel-oped world. Poultry demand is projected torise from 26 to 32 percent in the developingworld and from 29 to 33 percent in the devel-oped world.

Per capita poultry demand in the developedworld will rise from 22 kilograms per capita in1997 to 28 kilograms per capita in 2020, andfrom 7 to 11 kilograms in the developingworld. Poultry will represent the only meatproduct for which per capita demand in thedeveloping world rises less than per capitademand in the developed world, thus widen-

66 CHAPTER 4

TABLE 4.4 Net cereal trade value

by region, 1997 and 2020


(US$ billion)

United States 10.1 13.3

EU15 4.8 3.1

Former Soviet Union –0.3 0.9

Eastern Europe –0.1 1.6

Latin America –3.1 –0.4

Sub-Saharan Africa –2.3 –3.9

West Asia/ –7.8 –8.8

North Africa

China –1.9 –4.9

Southeast Asia –0.0 0.8

South Asia –0.2 –2.7

––––Source: IMPACT projections, June 2001.Note: Positive figures indicate net exports;negative figures indicate net imports.

TABLE 4.5 Meat demand by region,

1997 and 2020

Region 1997 2020





Eastern Europe 8.0 9.5

Former Soviet Union 12.0 13.5

East Asia 55.5 108.7


South Asia 7.3 15.8



World 208.1 326.5


ing the deficit between developing and devel-oped world consumers even further. Poultry’srelative share of meat demand will expand pri-marily at the expense of pork, which in 1997accounted for 37 percent of total meat demandin the developed world and 42 percent in thedeveloping world. The developed and devel-oping worlds’ increase in pork demand willonly account for 18 and 31 percent, respec-tively, of their increase in meat demand in2020.

Empirical observations of changing meatconsumption patterns in both the developingand developed worlds bear out these results.Meat demand in China will have a particularlyimportant effect on worldwide consumptiontrends. A number of studies (including Gao,Wailes, and Cramer 1996; Wang et al. 1998)have noted that pork consumption, while veryhigh in China because pork prices are low, risesslowly with income, and that higher-incomegroups tend to substitute poultry for pork, con-sidering poultry a luxury good. Wang et al.(1998) find that urban Chinese consumers

gradually shift away from pork to beef andpoultry, mainly because they desire greatervariety and beef and poultry products aremore available in urban than rural markets.They find that the demand for poultry is highlyelastic, compared with other meats, thus indi-cating that marginal expenditures on poultryare likely to increase in the future. The greateravailability of poultry and beef will depend inlarge part on ongoing structural changes tolivestock production within China, and thepoultry sector is industrializing rapidly, notonly in China but throughout Asia (Hoffman1999). We expect replication of these trends toa greater or lesser degree throughout thedeveloping world, as shifting demand towardmeat focuses heavily on poultry as a low-costsource of animal protein.

A number of studies claim that higher poul-try demand in the developed world is the resultof shifting taste patterns driven by concernsabout the health risks of consuming largequantities of red meat (Ward and Lambert1993; Capps and Schmitz 1991; McGuirk and


FIGURE 4.10 Per capita meat demand by region, 1997 and 2020


Mundlak 1991; and Kinnucan et al. 1997). Thedebate over the extent of this demand shift hasnot been resolved in the literature, with Ealesand Unnevehr (1993) and Davis (1997), amongothers, positing that the shift to poultry is actu-ally the result of productivity-driven pricedeclines in the poultry sector. But even if theeffects of health concerns have not been over-stated in the literature, Eales et al. (1998) pointout that structural shifts on the supply side areplaying a powerful role in driving poultry con-sumption higher. Whatever the reason, percapita poultry demand in the United Statesrose 10 percent between 1996 and 2000, whilered meat consumption only increased 2 per-cent (USDA 2000).

Meat Trade. As was the case for cereals, projectedinternational meat trade will expand tremen-dously between 1997 and 2020 because regionalsupply and demand will be out of balance (Fig-ure 4.11). The three main meat exporters, the

United States, Latin America, and the EU15, allwill experience significant increases in the valueof their meat exports by 2020. Meat exports willrise sharply in the United States, where growingpoultry exports (79 percent of all meat exports)will drive overall meat exports. The UnitedStates will also see a shift from a trade balancein pork in 1997 to exports of 1 million tons in2020. Latin American meat exports are projectedto expand between 1997 and 2020, but they will decline in value because poultry exports will replace high-value pork exports. Meatexports in the EU15 will also increase.

Among the meat-importing countries,China will experience the largest projected risein meat imports, from near trade balance in1997 to 4 million tons of imports in 2020.Poultry and pork products will drive thisincrease, with poultry imports rising from vir-tually zero in 1997 to 2 million tons in 2020,and pork imports rising 1 million tons from atrade surplus in 1997 to a trade deficit of 1 mil-

68 CHAPTER 4

FIGURE 4.11 Net meat trade by region, 1997 and 2020


lion tons in 2020. Meat imports into WANAand Southeast Asia will each increase by about1 million tons, while South Asia will changefrom a meat exporter in 1997 to a meatimporter in 2020. Sub-Saharan Africa will haveonly a minimal increase in meat imports.

Roots and Tubers

Aggregate roots and tubers demand in thedeveloping world will increase by 55 percent(248 million tons) between 1997 and 2020 (Fig-ure 4.12). Sub-Saharan Africa accounts for 44percent of this increase, indicating that rootsand tubers will continue to be an importantpart of the diet in that region. Asia will alsoaccount for a significant amount of the totalincrease, with East Asia accounting for 21 per-cent and South Asia, 14 percent (Figure 4.13).

Cassava and other roots and tubers willremain the dominant roots and tuber categoryin Sub-Saharan Africa, representing 68 percentof the total increase in roots and tubersdemand between 1997 and 2020 (Figure 4.14).Cassava will also drive roots and tubersdemand increases in Southeast Asia, account-ing for 80 percent of the total. Potato demandrepresents 93 percent of the total increase in

roots and tubers demand in South Asia, andpotato and sweet potato demand represents71 percent and 27 percent, respectively, ofChina’s total increase.

In the developing world, the supply of rootsand tubers is expected to increase only 51 per-cent between 1997 and 2020, thus slightly lag-ging demand growth. In Asia, roots and tubersdemand will increase 43 percent and produc-tion only 35 percent. In Southeast Asia, how-ever, demand will increase 60 percent, whileproduction will only expand 10 percent, thusbringing regional supply and demand intogreater overall balance. Supply and demandgrowth will be identical in Sub-Saharan Africaand in Southeast Asia. In the developed world,roots and tubers demand will increase by only3 percent and production by 9 percent, reflect-ing the relative inferiority of roots and tubersas a food commodity.

Rapidly improving yields will be necessaryto drive roots and tubers production increasesthroughout the developing world, and thearea planted to roots and tubers will actuallyshrink significantly in the developed world.Sub-Saharan Africa is the exception: areaexpansion will be responsible for a significant


FIGURE 4.12 Roots and tubers demand by region, 1997 and 2020


27 percent of additional roots and tubers pro-duction there.

Trade in roots and tubers will change sub-stantially between 1997 and 2020 (Figure 4.15).A decline of 10 million tons in roots and tubersexports out of Southeast Asia and a correspond-ing decline in the EU15’s net imports from 8 mil-lion tons in 1997 to 2 million in 2020 will be pri-marily responsible for the projected 3 million

tons decline in worldwide roots and tubers tradebetween 1997 and 2020. With the exception ofChina, no developing country will import morethan 1 million tons of roots and tubers in 2020.

Soybeans

Latin America will retain its dominant positionas the top regional consumer of soybeans in thedeveloping world, with demand increasing 78percent between 1997 and 2020. Production willmore than meet demand with an increase of 81percent. Among the developing countries, Chinawill challenge Brazil for the position of top soybean-consuming nation in the developingworld, as China’s demand increases by 96 per-cent, compared with a 63 percent increase inBrazil (Table 4.6). Brazilian production willincrease by 78 percent and Argentine produc-tion by 91 percent. China, on the other hand,will not be able to supply domestic demandincreases from domestic production, as soybeanproduction there will only increase by 78 per-cent. As a result, Chinese soybean demand willoutstrip domestic production by 12 million tons

70 CHAPTER 4

FIGURE 4.13 Share of roots and tubers

demand increase by region,

1997–2020

Source: IMPACT projections, June2001.

FIGURE 4.14 Increase in roots and tubers demand by crop and region, 1997–2020


in 2020, while Latin American soybean produc-tion will exceed domestic demand by 14 milliontons. This surplus should enable Latin Americato challenge the United States for a substantialshare of the soybean export market in years to

come. The European Union, however, willremain the main market for soybean imports, asthe deficit between domestic supply anddemand will be a tremendous 19 million tons.

Supply and demand changes will result in asignificant increase in international soybeantrade from 32 million tons in 1997 to 52 mil-lion tons in 2020. The United States willstrengthen its position as the world’s dominantsoybean exporter, increasing net exports by 8million tons in 2020 (Figure 4.16). Together,the United States and Latin America will sup-ply the majority of world import needs. TheEuropean and Chinese markets will dominateimports of soybeans, with imports of 19 mil-lion and 12 million tons, respectively, in 2020.

Soybean prices are projected to remain fairlyconstant between 1997 and 2020, actually ris-ing from $247 per ton in 1997 to $250 per tonon the strength of rising feed demand as pro-duction of livestock increases worldwide.


FIGURE 4.15 Net roots and tubers trade by region, 1997 and 2020


TABLE 4.6 Soybean supply and demand,

selected countries, 1997 and 2020

Production Demand

Region/Country 1997 2020 1997 2020


Argentina 14.1 26.8 13.0 22.2

Brazil 27.1 48.1 21.8 35.6

United States 70.9 94.9 46.0 62.9

EU15 1.4 1.9 16.6 21.3

China 14.3 25.5 19.2 37.6



Edible Oils

Southeast Asia, the largest producer of edibleoils, will increase its already sizeable produc-tion surplus, with production growth exceed-ing demand growth by 7 million tons between1997 and 2020. Much of the excess will beexported to East Asia, which will increase itsedible oil imports from 4 million tons in 1997to 10 million tons in 2020 (Figure 4.17). Sub-Saharan Africa, although it will remain thesmallest producer and consumer of edible oilsin the developing world in 2020, will require 2million tons of imports to satisfy domesticdemand in 2020. Latin America will increaseits edible oil exports by 3 million tons. Edibleoilseed prices will decline slightly between1997 and 2020, from $539 to $490 per ton.

Eggs and Milk

Eggs have limited tradability, so trends in pro-duction and demand will stay close together dur-ing 1997–2020. Egg demand will grow fastest inSouth Asia (113 percent) from admittedly lowlevels. Although the increase in demand for eggs

in China trails all other developing regions,China will continue to dominate egg consump-tion in the developing world, at a per capita con-sumption level of 19 kilograms per capita (Fig-ure 4.18). Per capita egg demand will actuallyfall slightly in the developed world by 2020.

Milk demand in Sub-Saharan Africa is pro-jected to increase by 110 percent and productionby 111 percent between 1997 and 2020. Never-theless, per capita milk consumption will onlyincrease by 1 percent per year (Figure 4.19). SouthAsia will continue to dominate milk consump-tion and production in the developing world,with demand increasing from 43 to 47 percent ofthe developing world total, and productionincreasing from 46 to 50 percent. Per capitademand in South Asia will grow at a rate of 2 per-cent annually. Per capita milk demand will alsoremain quite low in both East Asia and SoutheastAsia at only 19 kilograms per capita in 2020.

Agricultural Trade

The United States, Latin America, the EU15, andSoutheast Asia will be the four main net

72 CHAPTER 4

FIGURE 4.16 Net soybean trade by selected countries, 1997 and 2020


exporters of agricultural goods in value terms in2020, while all other developing countries will benet importers of agricultural goods (Table 4.7).17

The United States’ agricultural trade revenueswill increase by $13 billion, for net exports of $31

billion in agricultural goods by 2020. Net agri-cultural exports as a percentage of agriculturalproduction will increase from 11 percent in 1997to 16 percent in 2020 (Table 4.8). Latin America


FIGURE 4.17 Net trade in edible oils by region, 1997 and 2020


TABLE 4.7 Value of regional

agricultural trade, 1997 and 2020


(US$ billion)


EU15 0.8 5.1

Former Soviet Union –3.5 –1.6



West Asia/North Africa –12.2 –18.7

China –5.1 –21.5



––––Source: IMPACT projections, June 2001. Note: Positive figures indicate net exports;negative figures indicate net imports.

TABLE 4.8 Value of agricultural trade

as a percentage of total agricultural

production, 1997 and 2020


(percent)



China –2.1 –6.1

West Asia/North Africa –32.6 –33.5



Former Soviet Union –4.6 –2.1

EU15 0.5 3.2



will retain its position as the second largest agri-cultural exporter. Exports as a percentage of agri-cultural production will rise from 5 to 8 percent.The European Union’s export status will switchfrom minor to major between 1997 and 2020. Netagricultural exports will represent 3 percent oftotal production in 2020. Finally, Southeast Asia,a major player in worldwide agricultural marketsin 1997 with $6 billion in exports, will experiencea decline of $1 billion. Thus the importance ofits agricultural exports as a percentage of totalagricultural production will decrease from 9 to5 percent.

Among the agricultural importing regions,China will move past WANA to become thelargest importer in the world of agriculturalcommodities in value terms by 2020. Chineseagricultural imports will increase from $5 bil-lion to $22 billion by 2020. Chinese net agri-cultural imports as a percentage of total agri-cultural production will only increase from 2

to 6 percent, however, indicating the enor-mous size and continued rapid growth ofChina’s agricultural sector. Growth in demandfor oil crops, particularly for feed, will helpboost its rapidly growing import bill.

WANA had by far the largest net agricul-tural import bill as a percentage of total agri-cultural production in the world in 1997 at 33percent, but despite its large absolute increasein net agricultural imports, the share of totalproduction will only rise to 34 percent by 2020.

South Asia will experience a drastic shift frombeing a small importer of agricultural productsin 1997 to a major agricultural importer in 2020,translating into a substantial rise in net agricul-tural imports as a percentage of total agricul-tural production from 2 to 4 percent.

Sub-Saharan Africa’s net agricultural importbill will rise a modest $3 billion between 1997and 2020. Nevertheless, net agriculturalimports will represent a substantial 9 percent

74 CHAPTER 4

FIGURE 4.18 Per capita egg demand by region, 1997 and 2020



FIGURE 4.19 Per capita milk demand by region, 1997 and 2020


of total Sub-Saharan African agricultural pro-duction in 2020, a slight increase.

Per Capita Calorie Availability

At 3,536 calories per day, an increase of 392calories a day, China’s per capita calorie avail-ability will be higher than the average for thedeveloped world by 2020 (Figure 4.20). Underthe baseline scenario, per capita calorie avail-ability will increase in all developing regions,from an average of 2,667 calories per capita in1997 to 3,015 calories per capita in 2020. WANAand Sub-Saharan Africa, the two most import-dependent countries in 2020, will experiencethe smallest increases in per capita calorie avail-ability at 156 and 211 kilocalories, respectively.But Sub-Saharan Africa will be in far worseshape: WANA will average 3,208 calories perperson per day, compared with 2,442 calories inSub-Saharan Africa. South Asia, primarily India,

will also see large increases in kilocalorie avail-ability at 610 additional calories.

Projections of the Number of Malnourished

Children: Slow Progress

The number of malnourished children underthe age of five in the developing world is pro-jected to decline by only 21 percent from 166million in 1997 to 132 million in 2020. Anincrease of child malnutrition in Sub-SaharanAfrica of 34 percent, or 3 million children, is anindicator of a disturbing trend, particularly inthe northern part of the region (Figure 4.21).All of the other regions will see declines in thenumber of children who are malnourished.China, at 54 percent, will have the largestdecline, followed closely by Latin America, witha 52 percent decline. Although child malnutri-tion is expected to decline by 31 percent inSouth Asia, India will still be home to 44 mil-

lion malnourished children in 2020, represent-ing 34 percent of the total in the developingworld.

The regions with the highest prevalence ofchildhood malnutrition in the world, SouthAsia and Sub-Saharan Africa, will both see

declines in the share of children under the ageof five who are malnourished—South Asia of10 percent and Sub-Saharan Africa, 4 percent(Figure 4.22). Although per capita kilocalorieavailability is projected to be 426 calories higherin India than in Sub-Saharan Africa in 2020, the

76 CHAPTER 4

FIGURE 4.21 Number of malnourished children by region, 1997 and

2020


FIGURE 4.20 Daily per capita calorie consumption by region, 1997 and 2020


percentage of children who are malnourishedwill be 13 percent higher in India than in Africa.These figures indicate that aggregate food con-sumption is only one of many factors that deter-mine rates of childhood malnutrition. The lowstatus of women and limited female educationin India relative to Africa are important con-tributors to the high relative levels of childhoodmalnutrition in India.

LAND AND WATER: LIMITING FACTORS

TO GLOBAL FOOD SUPPLY?

The IMPACT baseline results show a future inwhich effective demand will be met in the con-text of constant or slowly declining prices, butalso a future in which progress in reducingchildhood malnutrition is very slow. In this sec-tion we assess whether land and water con-straints will pose serious threats to long-termcereal production growth. We will also look atthe effects of land degradation and conversionof land to urban uses on agricultural produc-tion and the effects of increasing water scarcityon future global food supply. This assessment

not only demonstrates the plausibility of thebaseline results with respect to land and wateravailability, it also indicates the risks of reducedinvestments and policy efforts, which couldlead to heightened resource constraints (espe-cially water supply), slow production growth,and significantly worse than projected out-comes in child malnutrition.

Cropland Potential and Land Loss

to Urbanization

Crop area harvested totaled 1,500 millionhectares in 1997, of which about 1,000 millionhectares were in the developing world and 500million hectares in developed countries (FAO2000a). Cereal crop area harvested totaled 738million hectares in 1997, with 258 millionhectares in the developed world and 480 mil-lion hectares in the developing world. Underthe baseline projections, cereal crop area is pro-jected to expand by 46 million hectaresbetween 1997 and 2020, almost all accountedfor by developing countries. Roots and tubersand soybean area is projected to increase byanother 16 million hectares. Can the existing


FIGURE 4.22 Malnourished children as a percentage of total children

under five years by region, 1997 and 2020


land base support this projected increase incereal crop area harvested?

In order to estimate cropland potential, theentire land area potentially convertible to agri-cultural uses must be taken into account.According to FAO (2000a), in 1994, total landresources were 13,044 million hectares, ofwhich 1,353 million hectares were classified asarable land, 114 million as having permanentcrops, 3,399 million as pasture, 4,172 millionas forest and woodland, and 4,003 millionhectares as other land, including built-on areas,roads, and barren land. Out of this area, Bur-ingh and Dudal (1987) identified 700 millionhectares as prime agricultural land and 2,600million hectares with low or medium capabil-ity for crop production. This would yield apotential land area suitable for crop productionof at least 3,300 million hectares, or a crop areapotential about 1,800 million hectares aboveexisting crop area.

As most of the currently cultivated landconstitutes relatively good agricultural land,the productivity of other landforms convert-ible into cropland should be lower than theexisting stock of land. Conversion may alsoeliminate forest and rangelands that now serveimportant functions. According to Kendall andPimentel (1994), the world’s arable land couldexpand at most by 500 million hectares, withproductivity below present levels. About 87percent of potential cropland is located indeveloping countries, mainly in Sub-SaharanAfrica and Latin America. In Asia, on the otherhand, nearly 80 percent of the potentiallyarable land is already under cultivation, andland for agricultural production is scarce inparts of China, Indonesia, and elsewhere(Plucknett 1995). Although global per capitaarable land has been decreasing steadily—from0.35 hectares in 1970 to 0.24 hectares in 1994—per capita area harvested has declined muchmore slowly—from 0.23 to 0.20 hectares in the

same period. The ratio of crop area harvestedto arable land, representing an aggregate crop-ping intensity index, has improved steadilyover the past three decades worldwide, from1.05 in 1970 to 1.20 in 1994, and from 1.28 to1.56 for developing countries during the sameperiod, making it less necessary to bring newland under cultivation (computed from FAO2000a).

The world’s urban population is expected tobe 4.3 billion by 2020, implying an overall urbangrowth rate of 2 percent between 1995 and2020, and 57 percent of the worldwide popula-tion will reside in urban areas, up from 45 per-cent in 1995. With urban populations expectedto be nearly stable in Europe and North Amer-ica during this period, approximately 90 percentof urban population growth will occur in devel-oping countries. Roughly 185,000 people willbe added to the urban population every daybetween 1995 and 2020. In China and much ofthe rest of Asia, the urban population’s share oftotal population is expected to double over thattime. Sub-Saharan Africa is expected to havealmost half of its population living in urbanareas by 2020; Latin America, 81 percent; andWANA, 68 percent (FAO 2000a).

There is no doubt that this rapid urbaniza-tion will remove some agricultural land fromproduction. Indeed, the conversion of landfrom agricultural to higher-valued uses onthe fringes of urban areas is part of theprocess of economic development, generat-ing in most cases significant economic bene-fits (Crosson 1986). Strategies biased towardurban and industrial growth, together withthe neglect of the agricultural sector, havealso led to significant damage to prime agri-cultural land (Bhadra and Brandão 1993).However, there is little evidence that theprocess of converting land to urban uses posesa serious threat to future global food produc-tion. For developing countries, urbanization

78 CHAPTER 4

is expected to lead to the conversion of476,000 hectares of arable land annually, a losstotaling 14 million hectares between 1990 and2020 (USAID 1988). Meanwhile, the baselineprojected increase in crop area harvested of62 million hectares necessary to meet effec-tive global food demand by 2020 is muchlower than both the theoretical maximumadditional potential crop area of 1,833 millionhectares and the more realistic potential foreconomically feasible conversion of landresources to agricultural uses of 500 millionhectares. A possible loss of 14 million hectaresof agricultural land to urban uses in the devel-oping countries appears small compared withpotential expansions in crop area, but couldeliminate highly productive land. The primaryconstraint to further crop area expansion isnot a physical limit: rather, it is the projectedflat or slowly declining real cereal prices thatrender expansion of cropland unprofitable inmany cases.

Physical Limits to Crop Productivity

Global food production can rise either throughexpansion of cropping area and greater croppingintensity or through increases in agriculturalproductivity. Crop area harvested is expected togrow only slowly between 1997 and 2020, thusplacing the burden for increases in agriculturalproductivity on higher yields. Are the projected1997–2020 yield growth rates biologicallyachievable? Will agricultural productivity be ableto keep up with global food requirements? Orare biophysical yield limits looming as a majorconstraint in the near future?

The earth’s food production systems wouldreach biophysical limits when all agriculturalland is being cultivated and irrigated at maxi-mum potential yields, with remaining landsuitable for grazing fully used. Maximum the-oretical yields are calculated for specific cropsas the highest limit of biological potential for

a given location on the basis of photosyntheticpotential, land quality, length of the growingseason, and water availability. Maximum the-oretical yields in grain equivalents calculatedby Linneman et al. (1979) and Luyten (1995)range from about 7.6 tons per hectare per sea-son in FSU to just over 8 tons per hectare perseason in China, India, and the rest of SouthAsia, to in excess of 9 tons per hectare per season in Southeast Asia, Sub-Saharan Africa,North America, and Western Europe. Base-line yield levels simulated by IMPACT for 2020are below maximum theoretical yields. Nev-ertheless, Cassman (1999) points out thatthese country average yields imply that themost productive cereal areas in northernIndia, southern China, and the North Ameri-can plains will be approaching biophysical lim-its. Achieving consistent production at thesehigh levels without environmental damagewill require improvements in soil quality andfarm management driven by continuing agricultural research investment (Cassman1999).

Land Degradation

The most comprehensive assessment of globalland degradation, by Oldeman, Hakkeling, andSombroek (1991), classifies the main types ofland degradation as soil erosion from wind andwater, chemical degradation (loss of nutrients,soil salinization, urban-industrial pollution, andacidification), and physical degradation (com-paction, waterlogging, and subsidence oforganic soils). Out of the total land resourcebase, Oldeman, Hakkeling, and Sombroek esti-mate that 1,964 million hectares have sufferedsome degree of degradation. Water erosionaccounted for 56 percent, wind erosion for 28percent, chemical degradation for 12 percent,and physical degradation for 4 percent. How-ever, chemical degradation was the prime cul-prit, accounting for 40 percent (an estimated 562


million hectares) of degraded agricultural land.Land degradation leads to reductions in cropyields, may reduce total factor productivity byrequiring the use of higher input levels to main-tain yields, may lead to the conversion of landto lower-valued uses, and may cause temporaryor permanent abandonment of plots.

Crosson (1995), based on the analysis byOldeman, Hakkeling, and Sombroek (1991),estimates the 1945–90 cumulative crop pro-ductivity loss from land degradation world-wide at approximately 5 percent, which isequivalent to a yield decline of 0.11 percent peryear. While this loss is not insignificant, theimpact of degradation was dwarfed by cropyield growth of 1.9 percent annually between1967 and 1997. Nevertheless, crop yield lossesdue to past erosion show cumulative crop yieldreductions that range from 2 to 40 percentacross all African countries, with a mean of 8.2percent for the entire continent and 6.2 percentfor Sub-Saharan Africa (Scherr and Yadav1996). These national-level estimates confirmthat land degradation can be devastating insome countries, especially in fragile environ-ments within country subregions. Moreover,while the estimated aggregate rates of yieldloss from land degradation are not huge,increases in these rates as a result of poor pol-icy or reduced investments could be a signifi-cant drag in the future, given the relatively lowbaseline projections for crop yield growth dur-ing 1997–2020.

Water and Irrigation

Between the 1950s and 1980s, irrigationexpanded rapidly. It currently accounts for 72percent of global water withdrawals and 90percent of withdrawals in low-income devel-oping countries. Dramatic yield increases dur-ing and after the Green Revolution wereachieved, in large part, through the adoptionof high-yielding varieties of wheat and rice,

which depend on timely nutrient and pest con-trol management as well as irrigation applica-tions to secure and control soil moisture. Thus,irrigated agriculture was a major factor inachieving rapid growth in cereal yields duringthe peak and post–Green Revolution periods.

By the mid-1990s, irrigated agriculture sup-plied nearly 40 percent of world food produc-tion on 17 percent of total cultivated land. InIndia, for example, irrigated areas (one-thirdof total cropped area) account for more than60 percent of total production. Irrigation alsofurthers stability through greater productioncontrol and wider scope for crop diversifica-tion. Moreover, in many developing countries,supplementary irrigation constitutes animportant element of rural development poli-cies, raising rural incomes and employmentand permitting increased agricultural and ruraldiversification through secondary economicactivities derived from extended and more var-ied agricultural production (as compared withrainfed agriculture) (Wolter and Burt 1997).

Thus, irrigation plays a vital role in achiev-ing food security and sustainable livelihoods indeveloping countries, both locally, throughincreased income and improved health andnutrition, and nationally, by bridging the gapbetween production and demand. However,new irrigation development has slowed sincethe late 1970s due to escalating constructioncosts for dams and related infrastructure, lowand declining prices of staple cereals, declin-ing quality of land available for new irrigation,and increasing concerns over the environ-mental and negative social impacts of large-scale irrigation projects. Declining expendi-tures are reflected in the declining growth incrop area equipped for irrigation. Accordingto FAO’s FAOSTAT database (2000a), theannual growth rate in global irrigated areadeclined from 2.2 percent during 1967–82 to1.5 percent during 1982–95. The decline was

80 CHAPTER 4

slower in developing countries, falling from 2.0to 1.7 percent annually during the same period,with some recovery during the early 1990s.

In 1997, cereal harvested irrigated area was218 million hectares, of which developed coun-tries accounted for 42 million hectares anddeveloping countries for 176 million hectares.18

Reflecting both the slowdown in the expansionof irrigated area and the rapid growth of non-agricultural water use demand, cereal irrigatedarea is projected to grow under the baselinescenario to 248 million hectares, with anincrease of only 1 million hectares in developedcountries and 29 million hectares in develop-ing countries. Rosegrant and Cai (2000), usinga prototype model linking IMPACT to a globalwater simulation model, show that long-termfood production growth is highly dependenton rates of growth in investment in irrigationand water infrastructure and improvements inwater use efficiency. Water will likely be amajor constraint to the achievement of foodsecurity in many developing countries in thefuture. This is especially true of the countriesof Central and Western Asia, North Africa, and

much of Sub-Saharan Africa, where popula-tion growth is expected to continue to be highand exploitable per capita water resourcesquite low. Water is important for food pro-duction not only because of direct effects onyields and cultivated area, but also because reli-able water supplies induce farmers to invest inother essential crop inputs, such as improvedgermplasm, fertilizers, and capacity buildingfor better resource management.

As this analysis indicates, the area and yieldexpansions projected under the IMPACT base-line fall within the realm of feasibility,although they may not be easily achievable.Since area under cultivation already consti-tutes a relatively high proportion of the mostproductive land, crop yields in these regionsmay suffer as agriculture expands ontoincreasingly marginal areas, simultaneouslyraising the risk of severe environmental dam-age. Shortfalls in investment in yield-enhanc-ing technologies and in research into theoptimal use of marginal agricultural landcould lead to slower than projected growthrates of area and yields.


As we approach 2020, the developing worldwill increasingly define the global food situa-tion. The IMPACT projections confirm thatdemand for cereals, meat, and dairy productswill increase rapidly in many developingregions. The fast growth of meat demand inAsia, for example, will pull cereal feed demandupward as well. At the same time, however, sig-nificant malnutrition will persist throughoutthe developing world, and trends will be par-ticularly worrying in some regions. As earlierdiscussions on area expansion in Latin Amer-ica and Sub-Saharan Africa and malnutrition inSouth Asia and Sub-Saharan Africa suggest,some trends may be inherently less predictablethan others, hinging critically on fundamentalassumptions about technology developmentand policy action. In South Asia, poverty is particularly intractable because populationgrowth rates are so high. In Sub-SaharanAfrica, the future is even less predictablebecause many African countries face majoreconomic crises and political instability, as wellas the HIV/AIDS epidemic. More concertedinternational and regional efforts must be

made to improve the outlook for South Asiaand to place Sub-Saharan Africa on a trajectoryof sustainable growth.

Our baseline results reflect our best assess-ments of a wide range of underlying policy,technological, and behavioral assumptions.But baseline outcomes may change signifi-cantly in the face of a range of realistic butwidely varying policy strategies and develop-ment paths for key drivers of the global foodeconomy. Investments in agriculture, waterresources, and social sectors may decline, andslow progress on economic policy reform maydampen income growth. Alternatively, policy-makers may take more aggressive steps towardimproving agriculture and other rural eco-nomic sectors, boosting investment, and accel-erating the pace of policy reform.

Real causes for concern over the future ofcrop yield growth rates, outlined more fullybelow, lead to the almost unavoidable conclu-sion that yield growth will be slower in thefuture than it has been in the past in mostregions—an assumption already incorporatedin the baseline projections. Nevertheless, the

5

ALTERNATIVE GLOBAL

SCENARIOS FOR FOOD

SUPPLY, DEMAND, TRADE,

AND SECURITY

82

prospects for future yield performance remainunclear. They will depend significantly on (1)trends in public and private sector agriculturalresearch investment, (2) the capacity for ongo-ing agricultural research to push in controver-sial new directions (such as biotechnology) inan increasingly watchful international envi-ronment, (3) continued investment in irriga-tion infrastructure, and (4) the rate of yield lossfrom environmental degradation. Worse thanexpected performance in any of these areascould cause even larger yield growth declinethan projected under the baseline scenario.

From among many potential scenariosassessing sources of variability in the worldfood situation, we focus on two sets of scenar-ios: one set exploring changes at the globallevel, the other set examining changes in a num-ber of variables specific to Asia and Sub-Saharan Africa. Among the global trends, thescenarios that follow will assess the potentialeffects of (1) slower population growth rates,(2) varying rates of yield growth, (3) higherlivestock productivity, and (4) full trade liber-alization in 2005–06. Among our regional scenarios, two will focus on Asia, with one sce-nario assessing the impacts of slower agricul-tural growth in India and China and anotherscenario simulating the effects of higher feedratios due to rapid industrialization of the live-stock sector. Two other scenarios will focusspecifically on Sub-Saharan Africa, presentingan optimistic and pessimistic future for thatregion, while a final regional scenario will sim-ulate the effects of dramatically higher meatdemand growth in India. Last, we will presenttwo alternative growth and investment sce-narios for the world in 2020: one containing aseries of investment declines, combined withslower policy reform; the other consisting ofincreased policy efforts and investments in agri-culture, irrigation and water, and social sectors.Strikingly, policies and investments that mod-

erately disfavor agriculture and rural develop-ment lead to much worsened food security vis-à-vis the baseline, while policies that moveaggressively to strengthen agricultural and eco-nomic policy reform yield much improved—though far from utopian—outcomes. Together,the alternative scenarios and the baseline estab-lish a range of possible pictures of the worldfood situation in 2020 that are vastly differentin terms of human suffering and that are highlydependent on decisions made by policymakers.

OVERVIEW OF GLOBAL SCENARIOS

What if population growth rates, often toutedas the chief underlying cause of food insecu-rity, were to slow more dramatically thanexpected? Would world food problems vanish?Will a technology revolution succeed inachieving higher levels of yield growth, or willenvironmental damage and other factors meanmuch slower yield growth over the projectionsperiod? With meat consumption on the rise,are prospects for higher production viaincreased efficiency of feeding ratios promis-ing? In short, how different would the worldfood picture in 2020 be under alternative sce-narios for crop and meat productivity growth?IMPACT results for these global scenariospoint to areas where policy action must focusto enhance future world food prospects.

Low Population Growth

Introduction. Many who paint a dire scenarioof human existence in the year 2020 point tohigh population growth rates in the develop-ing world as the main force behind continuedmalnutrition and environmental degradationin many regions. Were population growthrates to slow more quickly than baseline pro-jections indicate and instead follow a slowergrowth path, would food security problemsvanish? In this chapter, under the IMPACT low

ALTERNATIVE GLOBAL SCENARIOS FOR FOOD SUPPLY, DEMAND, TRADE, AND SECURITY 83

population growth scenario, we assess the effectsof replacing the United Nations’ (UN) mediumpopulation growth assumptions with their low population growth assumptions in aneffort to determine the effects of low popula-tion growth on food security.

Alternative UN population projections to2020 tend to cluster fairly close togetherbecause the momentum inherent in the cur-rent young age structure will be the primaryfactor behind future population growth every-where except Europe, where the populationtends to be older, and Africa, where high fer-tility is the dominant factor. Within the almostassured range, however, population growthrates may vary as the ongoing demographictransition in the developing world toward rel-atively low fertility and mortality takes unex-pected turns (Bongaarts and Bulatao 1999).Uncertainty about the pace of the demo-graphic transition in developing countriesremains a caveat behind all population projec-tions, with the large fertility decline since the1960s representing one surprising past trend.The potential future impact of the AIDS epi-demic injects a particularly important sourceof uncertainty into population projections.Caldwell (2000), for example, states that “AIDShas probably reduced the world’s currentannual population growth rate from 1.5 to 1.4percent.” The effects of AIDS on populationgrowth in certain regions, such as Sub-SaharanAfrica, will be considerably more substantial.(See the subsequent sections on Sub-SaharanAfrica.)

Population growth affects both future fooddemand and supply. The effects of populationgrowth on the demand side are fairly clear, withan exogenous increase in population growth atconstant aggregate income levels implyinglower per capita income and food consump-tion. The effects of population growth on foodproduction are generally less clear and highly

dependent on context. The size and composi-tion of the labor force, speed of technical inno-vation, and extent of environmental degrada-tion all represent potential avenues throughwhich population growth can influence foodproduction (Srinivasan 1988). Pessimists(Brown and Kane 1994; Ehrlich, Ehrlich, andDaily 1993), guided by the belief that many cur-rent agricultural practices are unsustainableand that the limits of further expansion of agri-culture have already been reached or exceeded,generally focus on the potentially negativeeffects of exogenous population growth on theenvironment—hence crop yields—through soildegradation and erosion (Srinivasan 1988; Bon-gaarts 1996). Ruttan (1994), surveying the linkbetween population and food in a number ofcountries, noted several factors frequently lim-iting the flexibility of farmers to respond to ris-ing population densities: (1) the resistance ofyield ceilings to further increases, (2) the fallingincremental response of yields to higher fertil-izer use, (3) the increasing cost of irrigationexpansion, and (4) the limited capacity ofresearch and extension institutions (cited inAhlburg 1998). Where these conditions holdtrue, slower population growth could relievethe pressure on the natural environment andreduce the strain on institutions and copingmechanisms in regions threatened by high pop-ulation densities.

Optimists such as Simon (1981), on theother hand, stress the role of populationgrowth in encouraging Boserupian-style tech-nological innovation, countering the staticanalysis of the pessimists with an emphasis onthe dynamism of individual economic actorsunder changing circumstances. The history ofagricultural development shows that the rela-tive proportions of land and labor endow-ments, nonagricultural demand for labor, andthe demand for final agricultural productsinfluence technical change in agriculture. As

84 CHAPTER 5

Boserup (1981) noted, farmers will adjust togrowing land scarcity by intensifying land usethrough reductions in fallow periods, increasedland investments, and implementation ofmore advanced farming techniques. Histori-cally, countries in Asia with high populationdensity have responded to land scarcitythrough the development and implementationof biological and chemical technologies as wellas substantial expansion of irrigation infra-structure (Pingali and Binswanger 1988). Abody of work (Ruttan and Hayami 1984;Hayami and Kikuchi 1983) provides empiricalsupport for the claim that population growthtends to spur innovation through the responseof technical and institutional change to emerg-ing scarcities, and that, conversely, slower pop-ulation growth tends to slow the pace of inno-vation (Srinivasan 1988). More recently,Bongaarts (1996) found strong positive effectsfrom higher population density on crop yieldsin both longitudinal and cross-sectional data,indicating that farmers do respond to highpopulation densities in a Boserupian fashion.In a study of slowing population growth inNorthern India, Evenson (1988) found that theloss of Boserupian-induced structural changereduced the real income gains from slowerpopulation growth by 36 percent.

The evidence for the impact of populationgrowth on agricultural growth and crop yieldsis thus mixed. Empirical evidence measuringthe impact of population growth on generaleconomic growth is also mixed. Pritchett’s(1996) study of the impact of populationgrowth on overall economic developmentconcludes that the capital stock is dynamicallyrelated to population growth, but it also findsa slight negative relationship between theresidual of output not accounted for by factoraccumulation and population growth. In theface of such contradictory evidence, and giventhe relatively small changes in population

growth between the baseline and low popu-lation scenario, we take a moderate position.As a result, our low population growth sce-nario assumes no second-order effects onaggregate income and crop yields from slow-ing population growth other than those thatoperate through international prices. Thus,aggregate income growth is maintained at thebaseline rates, resulting in an increase in percapita income growth under the low popula-tion scenario.

Projections. The low population growth sce-nario reduces the world population in 2020from an estimated 7,456 million people to7,068 million people, a decline of 388 millionpeople. Southeast Asia has the highest per-centage change—a 7 percent decline—fromthe total under the UN’s medium scenario(Table 5.1). The decline of 109 million peoplein South Asia represents 28 percent of the totalworldwide decline in projected population.Under the UN’s low growth scenario, the pro-jected population of the developing world as


TABLE 5.1 UN medium and low population

projections in 2020, by region

Medium UN Low UN

Region projection projection Change

(millions of people)

South Asia 1,780 1,671 109

Southeast Asia 649 604 45

East Asia 959 914 45

Sub-Saharan Africa 1,545 1,477 68

Latin America 652 613 39

West Asia/

North Africa 505 483 32

Developing world 6,096 5,757 339

Developed world 1,361 1,311 50


a whole will decline 6 percent, by 339 millionpeople.

Slower population growth would have amodest impact on per capita cereal and meatconsumption (Table 5.2). Per capita cereal con-sumption would increase 3 percent above thebaseline level in both the developing and thedeveloped world, with the highest percentageincreases occurring in Latin America and Sub-Saharan Africa. Per capita meat consumptionwould increase by a similarly modest 5 percentin the developing world and 2 percent in the

developed world. All developing regions would

have per capita meat production increases of

5 to 6 percent, except for East Asia, where the

increase in consumption would be slightly

lower, and Sub-Saharan Africa where meat

consumption would increase by 7 percent. East

Asia’s baseline projections are slightly more

robust in the low population growth scenario

because China’s one-child policy has already

slowed population growth rates significantly

below levels in other developing regions.

86 CHAPTER 5

TABLE 5.2 Per capita cereal and meat demand under various UN

population projections in 2020, by region

Cereal demand

Medium UN Low UN Percentage

Region projection projection Change change

(kilograms/capita) (percent)

South Asia 198 204 6 3.0

Southeast Asia 254 261 8 2.8

East Asia 384 394 11 2.6

Sub-Saharan Africa 163 171 7 4.9

Latin America 323 339 16 5.0

West Asia/North Africa 387 399 12 3.1

Developing world 275 284 9 3.3

Developed world 604 620 15 2.6

Meat demand


Region projection projection Change Change

(kilograms/capita)

South Asia 8.9 9.4 0.5 5.6


East Asia 70.3 73.3 3.0 4.3

Sub-Saharan Africa 11.7 12.5 0.8 6.8

Latin America 68.9 73.0 4.1 6.0

West Asia/North Africa 25.7 27.0 1.3 5.1

Developing world 34.9 36.7 1.8 5.2

Developed world 84.0 86.0 2.0 2.4


Projected cereal prices under the low popu-lation growth scenario shift significantly morefrom their baseline levels than does per capitaconsumption (Table 5.3). Rice prices are the mostelastic, declining 15 percent in 2020, while maizeis the least elastic with a price decline of 9 per-cent from the baseline level. These price declinesaffect yields in turn by reducing incentives tofarmers to invest in production, with cereal yields

falling by 0.1 ton per hectare in Latin America,WANA, South Asia, and East Asia (Table 5.4).Cereal yields in Southeast Asia and Sub-SaharanAfrica drop by less than 0.1 tons per hectare.

Dramatic declines in the number of mal-nourished children under the age of five in thedeveloping world will occur under the lowpopulation growth scenario. Per capita kilo-calorie consumption under the low populationgrowth scenario will increase by 88 kilocalo-ries above baseline levels, with Latin Americaleading the way with an increase in per capitaconsumption of 109 kilocalories (Table 5.5).Not only is the population under five smaller,but consumption of calories per capita ishigher due to lower food prices and higherincomes resulting from lower populationgrowth. The number of malnourished chil-dren in the developing world is projected todrop by an additional 29 million under the lowpopulation growth scenario, to 102 millionmalnourished children in 2020 (Table 5.6).While this number is still unacceptably high, a28 percent decline in child malnourishment istruly remarkable; it reveals the extent to whichhigh population growth in impoverishedregions adds millions of children to popula-tions that are already highly food stressed andunable to cope with the additional burden.19

South Asia stands out in this analysis: underthe low population growth scenario, child mal-nourishment in that region drops 25 percent(16 million children) below baseline levels.Unfortunately, even low population growthrates cannot reverse the tragic trend of increas-ing absolute numbers of malnourished chil-dren in Sub-Saharan Africa, where despite a 15percent decline (6 million malnourished chil-dren) from baseline levels, the number of mal-nourished children will still increase from 33to 34 million children by 2020.20

The results show that per capita consump-tion of cereal and meat products and cereal


TABLE 5.3 Cereal prices under various

UN population projections in 2020


Crop projection projection change

(US$/metric ton) (percent)

Wheat 123 109 11

Maize 102 93 9

Rice 250 212 15

Other coarse 86 77 10

grains


TABLE 5.4 Cereal yields in various

regions under medium and low UN

population projections in 2020

Medium UN Low UN

Region projection projection

(metric tons/hectare)

South Asia 2.5 2.4


East Asia 5.5 5.4







prices certainly react to changes in populationgrowth projections, but these adjustments donot dramatically modify the broad food supplyand demand situation in 2020 painted by thebaseline analysis. South Asia and particularlySub-Saharan Africa will still lag behind other

developing regions in per capita food con-sumption. The developed world benefits morethan the developing world from low popula-tion growth because levels of per capita con-sumption are so much higher in the developedworld that much lower percentage increases in

88 CHAPTER 5

TABLE 5.5 Total per capita kilocalorie

availability under UN population projections in

2020, by region

Medium UN Low UN


(kilocalories/capita)

South Asia 2,755 2,838 83

Southeast Asia 3,036 3,122 86

East Asia 3,500 3,587 87

Sub-Saharan Africa 2,442 2,530 88

Latin America 3,274 3,383 109

West Asia/North

Africa 3,208 3,283 75

Developing world 3,015 3,103 88

Developed world 3,462 3,523 61


TABLE 5.6 Number of malnourished children

under UN population projections in 2020, by

region

Medium UN Low UN


(millions of children)

South Asia 63.3 47.6 15.7


East Asia 8.5 6.2 2.3

Sub-Saharan Africa 39.3 33.7 5.6


West Asia/North

Africa 4.0 3.0 1.0



per capita consumption still lead to higherabsolute increases than in the developingworld.21 Nevertheless, while population growthrates may not dramatically affect the food secu-rity of populations in the aggregate, lower pop-ulation growth rates significantly reduce child-hood malnutrition. Dramatic declines inchildhood malnutrition across all developingregions—but particularly South Asia—underthe low population growth scenario arguestrongly for continued efforts to encourage vol-untary use of contraception and family plan-ning services in developing countries. Beyondthe narrow focus on population growth, how-ever, these results also point to the importanceof institutions that both lower the environ-mental costs of high population density andencourage technology uptake in farming sys-tems undergoing the process of population-induced innovation in farming techniques.

Low-Yield and High-Yield Growth Rate

Scenarios

Introduction. Declining yield growth rates overthe last two decades have generated much con-cern about the future of agricultural produc-tion. For a number of reasons, however, direpredictions regarding the imminent collapse ofagricultural yield growth rates should bereceived skeptically. First, declining yield growthrates can be consistent with long-term linearincreases in cereal yields and may simply reflectconstant unitary increases over an increasingbase level. Second, as the share of lower-yieldregions in world agricultural production grows,the aggregate worldwide yield growth rate maydecline even if yield growth rates in each indi-vidual region do not (Dyson 1996).22 However,despite these important structural elementsbehind declining yield growth rates, severalworrisome trends are evident, particularly thefact that yield growth rates have clearly slowedsomewhat at the regional level. It appears that

many yield gains in recent decades can be attrib-uted to advances that may not be replicable,including higher crop planting density throughchanges to plant architecture, higher usablefood product weight as a fraction of total plantweight, multiple harvesting, introduction ofstrains with greater fertilizer responsiveness,and better management practices. Crop yieldsmay be approaching their physical limits inmany high-yield systems, primarily in devel-oped countries. For example, the maximumyield potentials of rice and maize have changedlittle over the past three decades (Fedoroff andCohen 1999)

These real causes for concern lead to thealmost unavoidable conclusion that crop yieldgrowth rates will be slower over the projec-tions period than they have been in the past.The vast number of factors that will determineyield growth to 2020 include (1) rates of fer-tilizer and pesticide application, (2) the pace ofresearch investment and advances in biotech-nology, (3) physical and human capital devel-opment, and (4) the degradation and misuseof the natural resource base, with soil erosionand water shortages particularly relevant.

Rapid factor accumulation is often down-played in discussions about the new seed tech-nologies that represent the most visible cat-alyzing agent for the Green Revolution. AsMurgai (1999) reports, however, fertilizer andcapital input accumulation—rather thangrowth in total factor productivity (TFP)—wereresponsible for the preponderance of yieldgrowth in the Punjab region during the actualperiod of adoption of high-yielding varietiesbetween 1965 and 1973; TFP growth only accel-erated rapidly between 1974 and 1984. Mund-lak, Larson, and Butzer (1997) also find in across-country analysis of the determinants ofagricultural production that capital accounts forabout 40 percent of total output in the coretechnology. They conclude that “agricultural


technology is cost-capital intensive comparedto nonagriculture.” These results reinforce theoft-made observation that the key to sustainableyield growth is not a “silver bullet” technologi-cal breakthrough, but rather dynamic interac-tions among a number of individual factors—acaveat that must be kept in mind when assessingthe potential benefits of biotechnology (Pin-strup-Andersen and Cohen 2000). Dynamicinteractions are also not necessarily benign;while agricultural research and high input appli-cations hold the key to boosting—or even stabi-lizing in some particularly high-stress regions—crop yields in environmentally fragile areas, thenegative repercussions on the environment ofhigh rates of input application are already appar-ent in many parts of Asia. Only time will tell, butthe unavoidable conclusion is that worse thanexpected performance in any of the key areascould cause even larger yield growth declinesthan under the baseline scenario.

Projections. In order to assess the sensitivity ofcereal prices to yield growth rates, we modeledfour alternative yield scenarios and charted theprojected price trends for the four cereals overthe 23-year period between 1997 and 2020. Onepossible general trend over the coming yearsinvolves significant declines in the resourcesavailable for agricultural research and irriga-tion development. We assess this trend in theIMPACT model for two separate scenarios,both of which assume no growth in irrigatedarea during the projections period. Addition-ally, the first low-yield scenario assumes adecline in specified yield growth rates formeats, milk, and all crops in the developedworld of 40 percent from the baseline level, anda decline of 20 percent in all developingregions. The second low-yield scenarioassumes a decline in specified yield growthrates of 50 percent from baseline assumptionsin the developed world and 40 percent in the

developing world. Table 5.7 summarizes theseassumptions.

But it is also possible that potential threatsto world agricultural production will galvanizegovernments, multilateral institutions, and pri-vate firms into increasing their investmentsinto agricultural research and irrigation, thusleading to significant yield increases and expan-sion of irrigation infrastructure. IMPACThigh-yield scenarios assume an increase in theexpansion of irrigated area of 1 percent peryear greater than the baseline growth rate.Additionally, the first high-yield scenarioassumes an increase in specified yield growthrates for livestock, milk, and all crops of 10 per-cent above baseline assumptions in the devel-oped world and 20 percent in the developingworld. The second high-yield scenario assumesan increase in specified yield growth rates of20 percent above baseline growth rates in thedeveloped world and 40 percent in the devel-oping world.

90 CHAPTER 5

TABLE 5.7 Percentage of baseline yield

growth rates achieved and cereal yields

realized under high- and low-yield

scenarios, 2020

Developed Developing

Scenario world world

Share of baseline growth rate (percent)

Low-yield 1 60 80

Low-yield 2 50 60

High-yield 1 110 120

High-yield 2 120 140

Cereal yields (metric tons/hectare)

Low-yield 1 3.8 2.2

Low-yield 2 3.8 2.1

High-yield 1 4.0 3.1

High-yield 2 3.9 3.4


Our low-yield scenarios thus envision agreater downside for yield growth rates in thedeveloped world than in the developing world.This assumption meshes with empirical obser-vations showing that recent worldwide yieldslowdowns are largely the result of trends inthe developed world, where crop yields arehighly sensitive to price incentives and theongoing removal of agricultural protection(Dyson 1996). Overall, crop yields in the devel-oped world are also closer to ceiling levels thanyields in the developing world. Similar consid-erations underlie our assumption that theupside for yields in the developing world isgreater than the upside for yields in the devel-oped world. In much of the developing worldthe gap between current crop yields and yieldceilings offers greater room for substantialincreases given appropriate incentives andinvestments.

It should be noted that IMPACT simulationslead to changes in realized yields different fromthose that would result from straight-line cal-culations made from the initially specifiedyield growth rates, because the model capturesthe feedback effects between changes in yieldsand output prices.23 For example, if cereal yieldgrowth assumptions are lower than the base-line, cereal prices increase relative to baselineprices, which subsequently leads to partiallycountervailing increases in cereal yields (and

area) in response to higher price incentives.Thus, developing-country cereal yields in 2020under the first low-yield scenario will onlydecline to 2.2 tons per hectare from the base-line 2020 projection of 3.1 tons per hectare,and developed-country cereal yields will de-cline from 4.0 to 3.8 tons per hectare. Underthe second low-yield scenario, realized devel-oping-country cereal yields will decline to 2.1tons per hectare, while realized developed-country yields will decline to 3.8 tons perhectare. As for the high-yield scenarios, real-ized developing-country cereal yields underthe first high-yield scenario will improve to 3.1tons per hectare, while developed-countryyields will remain at 4.0 tons per hectare.Under the second high-yield scenario, realizeddeveloping-country yields will improve to 3.4tons per hectare, while developed-countryyields will actually decline to 3.9 tons perhectare due to countervailing price declines(Table 5.7).

Changes in the rate of growth of crop yieldshave huge impacts on projected cereal prices.Wheat prices increase by 20 percent under thefirst low-yield scenario. Rice prices increase by34 percent and maize prices by 22 percentabove the baseline 2020 price under this samescenario (Table 5.8).

The second low-yield scenario presents anadditional decline in yield growth rates from


TABLE 5.8 Cereal prices under baseline and various yield scenarios in 2020

Crop Low-yield 1 Low-yield 2 Baseline High-yield 1 High-yield 2

(US$/metric ton)

Wheat 148 164 123 106 92

Maize 124 140 102 87 75

Other coarse grains 107 122 86 72 62

Rice 334 392 250 193 156


those predicted in the first low-yield scenario,and prices react accordingly. Wheat pricesincrease by 33 percent above the projected 2020price under the baseline scenario. Other grainprices react similarly. Rice and maize prices arethe hardest hit, increasing 57 and 37 percentabove 2020 projected baseline prices to $392per ton for rice and $140 per ton for maize.

Prices for all crops decline under the high-yield scenarios, although the individual cere-als have significantly varying sensitivity toyield growth rate changes. Under the firsthigh-yield scenario, cereal prices decline fromprojected prices in 2020 under the baseline sce-nario by 15 to 24 percent. Price trends underthe second high-yield scenario are higher inmagnitude but similar in form to those underthe first high-yield scenario. Prices under thesecond high-yield scenario decline relative toprojected 2020 prices under the baseline scenario by 26 percent ($27 per ton) for maize,25 percent ($31 per ton) for wheat, 28 percent($24 per ton) for other grains, and 38 per-cent ($94 per ton) for rice.

As these projections indicate, yield growthwill play an important role in ensuring even themild declines in cereal prices projected by theIMPACT baseline, with changes in yieldgrowth rates at the margins exercising a sig-nificant influence on international cerealprices. Rice prices are particularly sensitive tothe low-yield growth scenario because a highproportion of rice is produced in developingcountries, which are affected the most underthis scenario. Clearly, the rates of crop yieldgrowth achieved over the next few decades—and therefore rates of investment growth foragricultural research and infrastructure devel-opment—will fundamentally determine theprice of food for the poor.

One particular point worth noting is thatthe developing countries cannot rely on thedeveloped world to supply the preponderance

of their future food needs. Trade will be animportant component of food security in thedeveloping world into the next century, buttrade cannot substitute for well-targeted andwell-funded investments in domestic food pro-duction as well as proper policies that provideincentives to local farmers. Production short-falls due to changing incentives and policies inthe developed world could have significantprice implications and a devastating impact onthe poor if the degree of food dependency inthe developing world is too high. Additionally,the potential upside of investments in agricul-tural production are higher in the developingworld, with the empowerment of localresearch capacity in developing countries nec-essary to ensure the widespread diffusion oftechnologies from the developed to develop-ing worlds. Historically, rates of return to agri-cultural research in the developing world havebeen tremendous; in a meta-analysis, Alston etal. (2000) found median rates of return rang-ing from a low of 34.3 percent in Sub-SaharanAfrica to a high of 49.5 percent in Asia and thePacific. Even if these high rates of return arediscounted significantly, they still indicate con-siderable underinvestment in crop research inthe developing world. Local research can beparticularly relevant to the pace at which localfarmers appropriate available technologies.For example, Murgai (1999) points out that thedifference between early and late adopters ofGreen Revolution technologies in the Punjabregion of India was probably due to their prox-imity to Punjab Agricultural University in Lud-hiana and access to extension programs in thecentral districts.

Higher Livestock Productivity via Lower

Feeding Ratios

Introduction. Rapid meat demand growth isalready prompting greater production effi-ciency, but what if feed ratios were reduced

92 CHAPTER 5

through improved technology to allow forgreater feedgrain efficiency in meat produc-tion? Would falling feed ratios dramaticallyreduce international cereal prices, and if so,what impact would lower cereal prices have onfood security and the incidence of childhoodmalnutrition in the developing world? The lowfeed ratio alternative scenario will explore thepotential ramifications for livestock productionand prices, grain prices, and cereal food con-sumption of a decline in feed ratios of 5 per-cent in the developing world and 10 percent inthe developed world, compared with the pro-jected levels in 2020 under the baseline.

As was shown in the historical section, inrecent years feed usage has declined in impor-tance as a driving factor behind overall cerealdemand growth. While overall food demandgrew 21 percent between 1984/86 and 1995/97,feed demand for cereals only rose 8 percent dur-ing this period. Meanwhile, worldwide meatproduction rose 34 percent; milk production, 5percent; and egg production, 54 percent (FAO2000b). Lagging growth in cereal feed demandrepresented a departure from the previous 10-year period, when feed demand increasestrended much more closely to food demandincreases. A number of factors—outlined fullyin the historical section of this volume—haveplayed a role in slowing feed demand, includ-ing price trends in cereal markets, substitutionof oil crops for cereals, higher productivity inthe meat production sector itself, and the sharpdownturn in the livestock sector in the transi-tion economies of Eastern Europe and theFSU.

The projected annual growth in feeddemand under the baseline scenario of 1.5 per-cent worldwide between 1997 and 2020 isslightly lower than the projected annualgrowth in meat production of 1.8 percent overthe same period, but it still represents sub-stantial recovery from the slow rates of growth

achieved between 1984/86 and 1995/97 (FAO2000b). The higher growth rate in feed demandis largely because most additional meat pro-duction between 1997 and 2020 will occur inthe developing world, where annual growthrates in meat production will average 2.8 per-cent, compared with 0.7 percent in the devel-oped world. It is thus not surprising that feeddemand growth is projected to average 2.7 per-cent a year in the developing world and 0.7 per-cent in the developed world.

Based on a number of factors, we concludethat feed demand growth will be rapid over thisperiod. First, livestock production in the devel-oping world, especially Asia, will increasinglyshift toward higher cereal intensity and indus-trialized systems. Second, the shift towardpoultry production in the developed worldover the last 20 years has already gone throughits most rapid growth phase, with poultry pro-duction already surpassing beef and veal out-put. It will be muted by the increasing domi-nance of the developing world in worldwidemeat production (Smil 2000). Thus, whereasworldwide annual growth in poultry produc-tion averaged 5.0 percent between 1982 and1997, it will only average 2.6 percent between1997 and 2020. Third, consumer preference forleaner animals is increasingly driving meatproduction in the developed world, and leaneranimals are inherently less efficient to produce.Historically, this trend has been most evidentin the pork sector, where the feed-to-meat pro-duction ratio actually increased between 1930and 1997, but it will become increasingly evi-dent in other livestock sectors as well (thegrowing popularity of free-range chickens, forexample) (Smil 2000).24

Nevertheless, it is possible that advances infeeding efficiency could lead to significantdeclines in meat production feed ratios, espe-cially if the developing world experiences ashift in meat preference toward poultry simi-


lar to what occurred in the developed worldover the last two decades of the twentieth cen-tury. Techniques to improve feeding efficiencyinclude better processing of concentrates androughage feeds and the use of additives likesupplementary amino acids. Other factors thatcould contribute to lower cereal feed ratiosinclude continued strong growth in the use ofoilcrops for feed and use of other alternativessuch as bananas, as well as the possibility of aslowdown in the commercialization of Asianlivestock production (Smil 2000; Hoffman1999). The pace of income growth and overalleconomic development in Asia over the next20 years will play an essential role in deter-mining the extent of industrialization and com-mercialization of livestock production (Stein-feld and Kamakawa 1999).

Projections. Lower feed ratios under this sce-nario result in lower livestock prices than thebaseline (Table 5.9). Poultry prices are mostaffected by the low feed ratio scenario, withprices falling 8 percent from the baseline levelin 2020. Pork and beef prices will decline 4 per-

cent and sheep and goat prices by only 2 per-cent. Dairy products are similarly affected,with egg prices falling 7 percent and milk prices3 percent. These price declines will boost meatdemand worldwide: poultry demand willincrease 4 percent; pork, 3 percent; and beef,2 percent (Table 5.10).

More significant price responses take placein the grains sector, where the decline in

94 CHAPTER 5

TABLE 5.9 Meat prices under the

baseline and low feed ratio scenarios, 1997

and 2020

1997 2020

Low feed

Commodity Base year Baseline ratio

(US$/metric ton)

Beef 1,808 1,740 1,670

Pork 2,304 2,239 2,123

Sheep and goat 2,918 2,832 2,764

Poultry 735 703 644

Eggs 1,231 1,191 1,111

Milk 318 289 281


TABLE 5.10 Meat and dairy demand

under the baseline and low feed ratio

scenarios in 1997 and 2020

1997 2020

Low feed



Beef 57 85 87

Pork 83 119 122

Sheep and goat 11 17 17

Poultry 57 105 109

Eggs 51 67 69

Milk 545 768 777


TABLE 5.11 Cereal prices under the

baseline and low feed ratio scenarios in 1997

and 2020

1997 2020

Low feed


(US$/metric ton)

Wheat 133 123 106

Maize 103 102 73

Other coarse

grains 97 86 61

Rice 285 250 233


derived demand for maize leads to a steep 28percent decline in the maize price (Table 5.11).Reverberations of this price decline reach thefood sector, particularly in the developingworld, where food demand for maize climbs10 percent (13 million tons) above the baselinelevel, thus partially counteracting the effectson total demand of sharply lower feeddemand. Food demand increases are particu-larly high in Sub-Saharan Africa at 11 percent,Latin America at 7 percent, and South Asia at3 percent. Although prices of all cereals decline,total demand for cereal crops decreases by 4percent worldwide, with demand falling formaize and other coarse grains but not forwheat and rice (Table 5.12).

The low feed ratio scenario has a notableimpact on regional cereal trade patterns (Table5.13). Cereal imports increase sharply in SouthAsia and Sub-Saharan Africa as low food pricesstimulate food demand, and the relative unim-portance of the livestock industry in these

regions mutes the countervailing influence oflower feed ratios. Compared with the 2020baseline, net imports into Sub-Saharan Africaincrease by 89 percent, and those into SouthAsia rise by 81 percent. Those regions withlarger livestock sectors see their net importsdecline, in some cases significantly. Reducedfeed demand has its largest impact on cerealimports into East and Southeast Asia. Netimports into Latin America fall 100 percent toachieve trade balance, albeit from a very lowbase. WANA, the largest cereal importer in theworld in 2020, will see net cereal imports drop8 percent. Interestingly, cereal trade in thedeveloped world responds only minimally tothe low feed ratio scenario, with net exportsrising just 1 percent.

These results indicate that significantadvances in the efficiency of feed ratios wouldhave moderate effects on livestock demand and


TABLE 5.12 Total cereal demand under

the baseline and low feed ratio scenarios in

1997 and 2020

1997 2020

Base Low feed

Region year Baseline ratio


Developed world 725 822 770

Developing world 1,118 1,765 1,623


Sub-Saharan Africa 83 156 170

West Asia/North

Africa 129 196 185

South Asia 238 353 362


East Asia 415 594 549


TABLE 5.13 Net cereal trade under the

baseline and low feed ratio scenarios in 1997

and 2020

1997 2020

Base Low feed

Region year Baseline ratio



Developing world 750 1,040 1,076

Latin America −15 −3 0

Sub-Saharan Africa −13 −27 −51

West Asia/North

Africa −45 −73 −67

South Asia −3 −21 −38

Southeast Asia −7 −9 −6

East Asia −21 −67 −42


prices and important effects on cereal demandand trade in several regions. Declining feedratios would drive down maize prices and driveup consumption of maize as food by a sub-stantial amount, leading to a total cereal foodconsumption increase of 3.5 percent (36 mil-lion tons) in the developing world (Table 5.14).However, 39 percent (14 million tons) of thisincrease in consumption is in Sub-SaharanAfrica, where maize and other coarse grainsare important staple foods. The impact of thishigher cereal consumption on child malnutri-tion is actually quite significant, with a declinein the number of malnourished childrenunder five in the developing world of 2.7 mil-lion children, for a total of 128.8 millionin 2020. Indeed, child malnutrition in Sub-Saharan Africa in 2020 would fall by 1.6 mil-lion children under this scenario, comparedwith the baseline.

Full Trade Liberalization

As we have noted, most governments, bothdeveloped and developing, have been unwill-ing to fully liberalize agricultural markets.They intervene in many ways in agriculture inorder to promote domestic food production,to keep domestic food prices low, or to reducedependence on foreign suppliers. Many stud-ies have shown that these measures result inmarket distortions and inefficiencies that leavemost people worse off. Reduction of agricul-tural trade distortions has been a major thrustof recent trade negotiations. The scenariodescribed in this section simulates the effectson food production, prices, and trade ofremoving all agricultural subsidies and tradebarriers in food markets (for the commoditiescovered in IMPACT).

In the full trade liberalization scenario, allprice wedges (PSEs and CSEs) between domes-tic and international cereal prices are removed,with the reductions phased in between 2005and 2006 (see Appendix B, Tables B.11 and B.12for PSE baseline values and Tables B.13 andB.14 for CSE baseline values). Special cautionis warranted when interpreting the results forthis scenario because IMPACT is a partial equi-librium model, which does not account for thecross-sectoral linkages that would undoubtedlyaccompany widespread trade liberalization. Ageneral equilibrium model best assesses suchlinkages (see for example, Diao, Somwaru, andRoe 2001). Nevertheless, the direction and rel-ative magnitude of the changes that result fromimplementation of the full trade liberalizationscenario are instructive in assessing the impor-tance that should be placed on the agriculturaltrade liberalization agenda.

As this scenario shows, full liberalizationwould have a significant effect on cereal pricesin 2020 (Table 5.15), with moderate increasesabove the projected baseline level for all cere-

96 CHAPTER 5

TABLE 5.14 Cereal food demand under

the baseline and low feed ratio scenarios in

1997 and 2020

1997 2020

Base Low feed

Region Year Baseline ratio



Developing world 750 1,040 1,076



West Asia/

North Africa 73 107 109



East Asia 248 289 293


als. Rice will increase the most, by 14 percent,followed closely by maize, wheat, and othercoarse grains. Meat prices will respond to fulltrade liberalization with even sharper price

increases above baseline levels, because meatprices are more distorted than cereal pricesunder the baseline scenario. The removal ofdistortions therefore has a greater impact on


TABLE 5.15 World prices under the baseline and full

trade liberalization scenarios in 2020

Full trade Percentage change

Commodity Baseline liberalization from baseline


Wheat 123 133 8.1

Rice 250 285 14.0

Maize 102 111 8.8

Other coarse grains 86 93 8.1

Beef 1,740 2,044 17.5

Pork 2,239 2,484 10.9

Sheep and goat 2,832 3,368 18.9

Poultry 703 785 11.7


TABLE 5.16 Net cereal trade under the baseline and full



Region/Country Baseline liberalization from baseline

(million metric tons) (percent)

United States 119 120 1

EU15 29 23 −21

Japan −30 −38 27

Australia 30 31 3

Latin America −3 −2 −33

Sub-Saharan Africa −27 −29 7

West Asia/

North Africa −73 −77 5

South Asia −21 −21 0

Southeast Asia −9 −10 11

East Asia −67 −65 −3

––––Source: IMPACT projections, June 2001.Note: Positive figures indicate net exports; negative figuresindicate net imports.

livestock producers and consumers than oncereal producers and consumers.25 Sheep andgoat and beef prices will rise by 19 percent and18 percent respectively, while pork and poultryprices increase by a lesser amount (Table 5.15).

Major changes in regional trade balancesunder a full liberalization scenario accompanymoderate price movements in internationalmarkets (Table 5.16). Among the developed-country exporters, EU15 cereal exports areprojected to fall 21 percent (6 million tons)below baseline levels by 2020; production willfall by 4 percent in response to reduced pro-ducer subsidies. Australian and U.S. cerealexports do not change substantially. The main

developed-world cereal importer, Japan, willsee net cereal imports rise from 30 to 38 mil-lion tons. Trade shifts vary in the developingworld, but are not particularly large overall.Net cereal imports rise by 11 percent in South-east Asia, 7 percent in Sub-Saharan Africa, and5 percent in WANA, and decline by 33 percentin Latin America and 3 percent in East Asia.

Regional cereal production and demand fig-ures generally bear out the regional trade data,with full trade liberalization eliciting relativelysmall responses (Table 5.17). The one excep-tion is Sub-Saharan Africa, where cereal pro-duction is expected to rise 5 percent and cerealdemand, 6 percent, indicating that full trade

98 CHAPTER 5

TABLE 5.17 Regional cereal production and demand under the

baseline and full trade liberalization scenarios in 2020



(million metric tons) (percent)

Production

United States 424 422 0

EU15 213 204 −4



West Asia/North Africa 122 121 −1


Southeast Asia 156 155 −1

East Asia 527 528 0

Demand

United States 305 302 −1

EU15 183 181 −1



West Asia/North Africa 196 197 1



East Asia 594 593 0


liberalization would be a significant boon toSub-Saharan Africa, stimulating both cerealproduction and demand. Sub-Saharan Africarealizes substantial production and consump-tion increases from both the removal of pro-tection in the EU15 and other developedregions and from the fact that some Sub-Saharan African regions tax both consumptionand production of cereal crops.

As indicated by the larger livestock pricechanges, the full trade liberalization scenariohas a greater effect on regional livestock trade,production, and demand than it does on cere-als, because existing levels of protection aregenerally higher for livestock products than forcereals (Table 5.18). Among the major live-stock exporting regions, livestock exports willincrease substantially in Latin America and theUnited States and will fall in the EU15. Mostof the other regions will see substantial

increases in their meat imports, but WANA’snet imports of meat will decline 28 percent.

Production declines are mainly responsiblefor the EU15’s falling livestock exports, whiledemand increases are almost entirely respon-sible for the sharp rise in meat imports intoSub-Saharan Africa (Table 5.19). Slight pro-duction increases combined with slightdemand declines will be responsible for theincrease in net exports realized by both theUnited States and Latin America and will alsobe responsible for the slight decline in WANA’snet meat imports. A moderate productiondecline combined with rising demand will beresponsible for the significant increase in netimports in South Asia. While the percentageincrease in Southeast Asia’s imports appearslarge, it starts from a very low base and ismainly the result of a slight production declinecombined with a slight demand increase. Meat


TABLE 5.18 Net meat trade under the baseline and full





United States 6.1 8.9 46

EU15 2.4 1.8 −25

Japan −3.2 −3.5 9

Former Soviet Union −2.5 −3.3 32

Latin America 2.4 5.3 121

Sub-Saharan Africa −0.3 −1.5 400

West Asia/North Africa −1.8 −1.3 −28

South Asia −0.5 −2.3 360

Southeast Asia −0.5 −1.3 160

East Asia −5.4 −6.6 22

––––Source: IMPACT projections, June 2001.Note: Positive figures indicate net exports; negative figures indicatenet imports.

production and demand actually change onlya little in East Asia.

The full trade liberalization scenario willhave small overall effects on regional kilocalo-rie availability, with the most importantimprovement being an increase of nearly 2 per-cent in Sub-Saharan Africa. The number ofmalnourished children under the age of five inthe developing world in 2020 is projected todecline by 1 million children from baseline lev-els, with more than half of this improvement

occurring in Sub-Saharan Africa. In light ofthese modest welfare changes, it should bestressed again that the IMPACT model doesnot consider the dynamic efficiency improve-ments that would result from greater agricul-tural trade liberalization.

Much more dramatically, trade liberaliza-tion generates significant net economic bene-fits. In the partial equilibrium approach usedhere, the net economic benefits resulting fromfull trade liberalization are estimated as the net

100 CHAPTER 5

TABLE 5.19 Regional livestock production and demand

under the baseline and full trade liberalization scenarios in

2020




Production

United States 46.5 47.8 3

EU15 37.1 36.3 −2

Latin America 47.4 48.6 3

Sub-Saharan Africa 11.0 10.9 −1

West Asia

North Africa 11.2 11.4 2

South Asia 15.3 14.3 −7

Southeast Asia 18.2 17.7 −3

East Asia 103.3 102.8 0

Demand

United States 40.4 38.9 −4

EU15 34.7 34.5 −1

Latin America 44.9 43.3 −4

Sub-Saharan Africa 11.3 12.4 10

West Asia/

North Africa 13.0 12.7 −2

South Asia 15.8 16.6 5

Southeast Asia 18.6 19.0 2

East Asia 108.7 109.5 1


benefits to producers (change in producer sur-plus) plus the net benefits to consumers(change in consumer surplus) plus the tax sav-ings due to removals of subsidies under tradeliberalization, compared with the baselineresults in 2020. It is projected that liberaliza-tion of trade for the 16 commodities includedin the model would generate global benefitsof $35.7 billion in 2020 (Table 5.20). Bothdeveloped and developing regions benefit,with the former gaining $14.2 billion and thelatter $21.5 billion. Although these gains arenot significant by GDP, in many regions theyare significant by value of agricultural pro-duction. In proportion to their agricultural sec-tors, the biggest gainers are Japan and SouthKorea (the latter included in Other East Asiain Table 5.20). But most important, the biggest

single gainer is Sub-Saharan Africa, at $4.4 bil-lion, or 10 percent of the 2020 value of pro-duction of the commodities examined here.This is partly because African farmers wouldface less competition from subsidized exportsfrom Europe and other developed countriesunder trade liberalization. However, a signifi-cant part is also due to the removal of thecostly subsidies and taxes that many Africangovernments impose on food production andconsumption.26

OPTIMISTIC AND PESSIMISTIC SCENARIOS

Assumptions

The global pessimistic and optimistic scenariosassume alternative outcomes across a broadrange of policy-sensitive variables that affect food


TABLE 5.20 Net welfare effects of global trade liberalization for

IMPACT commodities

Net Percent value of Percent value

benefits, agricultural of GDP,

Region/Country 2020 production, 2020 2020

(US$ billion) (percent)

World 35.7 2.99 0.07

Developed world 14.2 3.02 0.04

United States 4.3 2.53 0.03

EU15 4.2 3.04 0.03

Japan 3.0 22.27 0.04



West Asia/North Africa 2.3 5.9 0.13

Sub-Saharan Africa 4.4 10.4 1.03

China 3.6 1.34 0.11

Other East Asia 2.4 28.72 0.18

India 2.1 1.93 0.14

Other South Asia 1.3 3.34 0.36



security and are subject to some degree of uncer-tainty. These include national GDP growth rates;social service investment, such as sanitation,health, and education; agricultural technologyinvestment, affecting yield growth rates; envi-ronmental degradation of both soil and water,influencing the potential for expansion of areacultivated and irrigated; removal of barriers totrade; and projected population growth over thenext 20 years. Table 5.21 summarizes all shiftsunder the optimistic and pessimistic scenarios.

Since unpredictable events and choices madeat the national level can substantially affect

GDP growth, the pessimistic scenario positsthat national GDP growth will suffer a 25 per-cent decline (and growth under the optimisticscenario will accelerate by 25 percent) from thebaseline’s moderate levels. For example, if pro-jected GDP growth for a given country is 4 per-cent per year in the baseline scenario, it is 5 per-cent in the optimistic scenario and 3 percent inthe pessimistic scenario.

Beyond food availability, a broad range offactors related to social spending stronglyinfluence rates of childhood malnutrition indeveloping countries (notably rates of school-

102 CHAPTER 5

TABLE 5.21 Percentage changes from baseline conditions under optimistic and pessimistic

scenarios

Change Pessimistic scenario Optimistic scenario

Malnutrition Multiply schooling and sanitation Multiply schooling and sanitation

indicators indicators by 0.9, subtract 0.04 indicators by 1.1, add 0.04 to life

from life expectancy ratios expectancy ratios

GDP growth rates 25 percent decline in GDP growth 25 percent increase in GDP growth

rates in developing world rates in developing world

Agricultural research Yield growth in developed countries Yield growth in developed countries

investment declines 50 percent; yield growth in increases 10 percent; yield growth in

developing countries declines China and India rises 20 percent; yield

40 percent growth in other developing Asia rises

15 percent; yield growth in other

developing countries rises 10 percent

Environmental Area growth declines by 0.15 Area growth increases by 0.10

degradation

Irrigation growth 0 percent growth in irrigated area 1 percent growth in irrigated area

above baseline levels

Population growth UN high scenario UN low scenario

Trade measures PSEs = 0.2 in 2005 through 2009; No change

PSEs = 0.4 in 2010 through 2020

CSEs = −0.20 in 2005 through 2009;

CSEs = −0.4 in 2010 through 2020


ing, the extent of sanitation, and female lifeexpectancy). The pessimistic scenario assumesa decline in both the percentage of people withaccess to safe water and the percentage offemales with access to secondary education of10 percent relative to the baseline percentagein 2020 and a decline in female life expectancyof 4 percent. The optimistic scenario assumesthat the levels of these three indicators rise insymmetric fashion with respect to the baseline(that is, gains of 10 percent in female access toeducation and access to clean water, and 4 per-cent in the female life expectancy ratio).

Yield growth scenarios depend on the out-comes of debates over the environmental andhealth implications of genetically modifiedorganisms and on the level of support for inter-national agricultural research as well as irriga-tion and water resources investment in both thedeveloped and developing worlds. Worse thanexpected future performance in the areas of agri-cultural research and development, irrigationinfrastructure and trade liberalization, amongothers, could cause even larger drops in yieldgrowth than projected under the baseline. Alter-natively, it is possible that potential threats toworld agricultural production and changes inincentives will galvanize governments, multi-lateral institutions, and private firms intoincreasing their investments in agriculturalresearch and irrigation, thus leading to signifi-cant yield increases and expansion of irrigationinfrastructure. Under the pessimistic scenario,yield growth rates are assumed to decline frombaseline levels by 50 percent in developed coun-tries and 40 percent in developing countries.Under the optimistic scenario projected yieldgrowth will increase relative to the baseline.Yield growth rates are projected to be 10 percenthigher in the developed countries, 20 percenthigher in India and China, 15 percent higher inthe rest of developing Asia, and 10 percenthigher in the rest of the developing world.

Higher upside technological potential for partsof Asia in the optimistic scenario is based on theprojected greater potential for uptake ofbiotechnology and other technological advancesin these regions. Declines in yield growth ratesunder the pessimistic scenario are higher thanthe parallel increases under the optimistic sce-nario, reflecting our assessment that there is agreater likelihood of a significant downside riskrelative to the baseline assumptions.

Much attention in recent years has focusedon the importance of soil erosion and over-farming in reducing the amount and quality ofland suitable for agricultural exploitation. Therate at which environmental degradation isactually occurring is far from clear, however,in part because good management practicescan reverse the deteriorative process and bringpreviously unsuitable land under cultivation.The pessimistic scenario models heightenedenvironmental degradation by reducing thegrowth rate of agricultural area by 15 percentin developing countries. The optimistic sce-nario assumes improved environmental man-agement in developing countries and thereforeincreases area growth rates by 10 percent overthe baseline scenario.

As we have already mentioned, water avail-ability could be the most important limitingresource to agricultural production growth overthe next 25 years. Most easily accessible irriga-tion water has already been exploited, and thescope for further large-scale irrigation develop-ment is limited. The baseline assumes a con-tinued slowdown in irrigation expansion, butstill allows for an increase in irrigated area of 30million hectares between 1997 and 2020. How-ever, it is possible that net increases in irrigatedarea over the next 25 years will be negligible if committed investment is countered by sedimentation of existing water storage andincreased rates of salinization and water-logging. The pessimistic scenario therefore


assumes zero growth in irrigated area over theprojections period. Alternatively, many nationalgovernments possessing adequate resourcesmay decide to pursue a heightened program ofirrigation development, and expansion of pri-vate investment could increase at higher thananticipated rates. The optimistic scenario mod-els this possibility by assuming an additional 1percent per year growth in developing-countryirrigated area above the baseline rate, whichwould add an additional 35 million hectares overthe baseline irrigated area in 2020.

UN population projections have undergonefrequent downward revisions over the past cou-ple of decades, as demographic shifts in the devel-oping world led to lower population growth ratesin most regions. The baseline adopts the UNmedium population growth rate projections, andour optimistic and pessimistic scenarios adoptthe low and high projections, respectively.

The slow start of the Millennium Round ofthe World Trade Organization calls into questionthe commitment of nations in both the devel-oping and developed worlds to further trade lib-eralization. The pessimistic scenario assumes aworsening of the world trade environment,together with an increased desire to protect slow-growing agriculture in developing countries.

Increased protection is represented by a phasedincrease in the PSE and a decrease in the CSE by20 percent between 2005 and 2009 and by anadditional 20 percent between 2010 and 2015.

Results

The pessimistic scenario will significantly affectworldwide agricultural production, prices,demand, trade, and child malnourishment in thedeveloping world. Worldwide cereal productionwill increase 29 percent (54 million tons)between 1997 and 2020 under the pessimistic sce-nario, compared with a 33 percent (63 millionton) increase under the baseline scenario and a38 percent (71 million ton) increase under theoptimistic scenario. Worldwide per capita cerealconsumption under the 2020 pessimistic sce-nario will be 28 kilograms per capita lower underthe pessimistic scenario than under the baseline,and it will be 30 kilograms higher than the base-line under the optimistic scenario.

Production. Per capita cereal production in2020 will decline across all regions under thepessimistic scenario (Table 5.22). Among thedeveloping regions, Sub-Saharan Africa willhave a slight decline in per capita cereal pro-duction of 4 percent from the baseline level by

104 CHAPTER 5

FIGURE 5.1 Wheat prices, alternative scenarios, 1997–2020


2020. Projected total per capita cereal produc-tion will be significantly lower than under thebaseline scenario in all other developingregions, with particularly large declines inSouth Asia at 14 percent, WANA at 9 percent,and Latin America at 10 percent.

Per capita cereal production under the opti-mistic scenario in 2020 will increase above lev-

els projected by the baseline in all regions.South Asia, Southeast Asia, and China willhave particularly strong production responses.

Prices. The optimistic and pessimistic scenarioswill dramatically affect world cereal prices (Fig-ures 5.1 to 5.4). Prices for all cereals in the opti-mistic scenario decline at an accelerating rate


FIGURE 5.2 Maize prices, alternative scenarios, 1997–2020


TABLE 5.22 Per capita cereal production in 2020 by

region, alternative scenarios

Region/Country Baseline Optimistic Pessimistic

(kilograms/capita)

United States 1,339 1,398 1,240

EU15 573 589 561

Eastern Europe 871 886 780

Former Soviet Union 487 497 456



West Asia/North Africa 243 263 220



China 355 394 328

World 335 365 307


throughout the projections period, with riceprices declining 44 percent; other grain prices, 33percent; wheat prices, 29 percent; and maizeprices, 22 percent. Cereal prices will rise in thepessimistic scenario, with other grain prices ris-ing 29 percent; wheat prices, 26 percent; and riceprices, 45 percent. Maize prices are actually pro-jected to rise 36 percent between 1997 and 2020to a price of $140 per ton in 2020, almost twicethe corresponding decline under the optimisticscenario. Thus declines for other coarse grainsand wheat will be greater under the optimisticscenario than corresponding increases under the

pessimistic scenario, while the reverse will betrue for maize and rice.

Meat production is not greatly susceptibleto the varying scenarios, with meat productionin 2020 falling an average of 5 percent belowbaseline levels in the pessimistic scenario andrising an average of 2 percent above baselinelevels in the optimistic scenario (Figure 5.5).

Trade. The pessimistic scenario is projected toaffect total world cereal trade more than theoptimistic scenario is, but the overall effects ontotal volume of trade are not large. The pes-

106 CHAPTER 5

FIGURE 5.4 Other coarse grain prices, alternative scenarios, 1997–2020


FIGURE 5.3 Rice prices, alternative scenarios, 1997–2020


simistic and optimistic scenarios begin todiverge gradually from the baseline between2000 and 2005. Projected world cereal trade in2010 under the pessimistic scenario is 7 milliontons higher than under the baseline scenario,eventually increasing to a total differencebetween the two scenarios of 20 million tonsby 2020. Projected world cereal trade underthe optimistic scenario in 2010 is more than 6million tons lower than the baseline, but in2020 this difference will be reduced to 5 mil-lion tons for a total of 296 million tons.

The impact of the scenarios on the value ofworld cereal trade is slightly different from volu-metric trends, mainly due to the effects ofchanging commodity prices. The value ofworld cereal trade under the pessimistic andbaseline scenarios diverges sharply between2000 and 2005 and is ultimately $24 billionhigher under the pessimistic scenario in 2020(Figure 5.6). The value of world cereal tradeunder the optimistic scenario trends below thevalue of trade under the baseline and pes-

simistic scenarios from the beginning of theprojections period, and it grows moderatelyfaster between 2015 and 2020. World cerealtrade in 2020 under the optimistic scenario is $9billion less than under the baseline scenario. Therelatively small changes in total trade are due tocountervailing forces within each of these alter-native scenarios. In the pessimistic scenario, ris-ing world cereal prices and declining incomes,combined with increased trade barriers in devel-oping countries, tend to drive down food importdemand, while slower production growth andhigher population growth would tend to boostimport demand.

By contrast, in the optimistic scenario, morerapidly increasing production and lower popula-tion growth tend to reduce imports; higherincome growth tends to increase import demand.

The effects that the optimistic and pessimisticscenarios have on regional imports in 2020 willvary by region, but net cereal imports into thedeveloping world as a whole will decline slightlyin both scenarios relative to the baseline. When


FIGURE 5.5 Meat production in 2020 by region, alternative scenarios


the net cereal imports are broken down byregion, the different regions have widely diver-gent trade responses to the different scenarios,depending on the balance of countervailingforces in any given region (Figure 5.7). The neteffect on imports for any given country dependsin particular on the demand and supply elastic-ities and rate of change in yield and area growth.

The volume of cereal imports into East Asia(mainly China) in 2020 will decline by 18 per-cent from baseline levels under the optimisticscenario and 21 percent under the pessimisticscenario. Cereal imports into Sub-SaharanAfrica in 2020 under the pessimistic scenariodecline 42 percent, while Latin Americabecomes a net exporter of 2 million tons of

108 CHAPTER 5

FIGURE 5.7 Net cereal trade in 2020 by region, alternative scenarios

Source: IMPACT projections, June 2001.Note: Positive figures indicate net exports; negative figures indi-cate net imports.

FIGURE 5.6 Value of world cereal trade under alternative scenarios,

1997–2020


cereals in 2020. The remaining regions willexperience rising import dependence underthe pessimistic scenario, with South Asia expe-riencing the largest percentage increase of 90percent, as well as the largest absolute increaseof 19 million tons. Cereal imports under theoptimistic scenario will increase in both LatinAmerica and Sub-Saharan Africa by significantamounts, but they will decline in all of theother regions.

As was the case for total world cereal trade,changing cereal prices somewhat modify thevolumetric cereal import picture as a result ofboth regional differences in the composition ofcereal demand and price changes over the pro-jection period (Figure 5.8). South Asia’s importbill rises from $3 billion under the baseline sce-nario to $12 billion under the pessimistic sce-nario, even though the volume of net cerealimports only approximately doubles. Mean-while, Southeast Asia, which showed a tradedeficit in volume under all three scenarios, actu-ally has a positive trade balance in value termsacross the board, due to the preponderance ofhigh-value rice in its cereal exports.

Demand. The optimistic and pessimistic scenar-ios will have significant effects on per capita kilo-calorie availability and childhood malnutrition.In terms of per capita kilocalorie availability, thepessimistic scenario will hurt South Asia, LatinAmerica, and Sub-Saharan Africa the most, withper capita kilocalorie consumption in 2020 thatis 7 percent lower than under the baseline sce-nario (Figure 5.9). (The declines in consumptionin all of the other regions are quite close to that,however.) Sub-Saharan Africa, South Asia, andLatin America will also benefit from the opti-mistic scenario, with per capita kilocalorie con-sumption in 2020 improving approximately 6percent above the baseline in these regions, withgains in the other regions close behind.

Childhood Malnutrition under Alternative Scenar-

ios. The most devastating result from the pes-simistic scenario is the impact on childhood mal-nutrition. The pessimistic scenario has a highlydetrimental effect on the welfare of children inthe developing world (Figure 5.10). Under thepessimistic scenario, the number of malnour-ished children in the developing world actually


FIGURE 5.8 Net cereal trade value in 2020 by region, alternative scenarios

Source: IMPACT projections, June 2001.Note: Positive figures indicate net exports; negative figures indicatenet imports.

increases, from 166 million children in 1997 to178 million children in 2020, a full 46 millionmore malnourished children than in the base-line 2020 projection. The optimistic scenario, incontrast, generates significant reductions in thenumber of malnourished children, down to 94million children in 2020. Whereas the baseline

scenario projected 2 million malnourished chil-dren in Latin America in 2020, a decline of 3 mil-lion children from the 5 million children mal-nourished in 1997, the pessimistic scenarioprojects this number to rise to 7 million children.WANA, which was to reduce its number of mal-nourished children under the baseline from 6 mil-

110 CHAPTER 5

FIGURE 5.10 Malnourished children in 2020, alternative scenarios

Source: IMPACT projections, June 2001

FIGURE 5.9 Per capita kilocalorie availability in 2020 by region, alternative scenarios


lion children in 1997 to 4 million in 2020, wouldexperience an increase to 7 million children in2020 under the pessimistic scenario. China, pro-jected to more than halve its number of mal-nourished children to 8 million children between1997 and 2020, will see hardly any improvementin the pessimistic scenario, with its number ofmalnourished children only declining to 15 mil-lion children. South Asia and Sub-Saharan Africawill have the largest absolute increases in the pro-jected number of malnourished children in 2020,with increases of 14 million and 10 million chil-dren above baseline levels, respectively.

The optimistic scenario projects impressiveimprovements, with Latin America completelyeliminating child malnourishment, WANAexperiencing a decline to only 1 million mal-nourished children, and China reducing thenumber of malnourished children to 3 millionby 2020. Child malnourishment in South Asiain 2020 declines from baseline levels by a size-able 13 million malnourished children in theoptimistic scenario, and an impressive 8 millionchildren in Sub-Saharan Africa.

These results indicate that the alternativescenarios have profound effects on food secu-rity as measured by childhood malnutrition.For the pessimistic scenario, reductions in percapita income and increased food pricesreduce per capita food consumption asdescribed above. The reduction in food con-sumption reduces per capita calorie con-sumption, which in turn increases the per-centage of the childhood population that is

malnourished. Reductions in public invest-ment in the social sector—including femalehealth, education, and water and sanitation—result in further increases in the percentage ofmalnourished children. The total childhoodpopulation also increases due to the higher fer-tility rates incorporated in the high populationgrowth scenario. The combined result of theseeffects is an increase in the number of mal-nourished children.

The optimistic scenario has the oppositeeffect on child malnutrition. Under the opti-mistic scenario, three broad courses combineto reduce projected levels of malnutrition: thefirst is through broad-based and rapid agricul-tural productivity growth and economicgrowth to increase effective incomes, effectivefood demand, and food availability; the secondis through a reduction in population growthrates; and the third is through investmentsthat improve access to education, female lifeexpectancy (as a proxy for quality of life forfemales), and health (the latter proxied by accessto clean water in the model). The relativecontribution of these three paths to reducingmalnutrition were estimated by undertakinga series of simulations embodying specificinterventions, including crop productivity–enhancing investments, lower populationgrowth, and higher social investments. Thesedisaggregated simulations indicate that each ofthese factors accounts for about one-third of theimprovement in childhood malnutrition.


Not all likely or potentially important devia-tions from projected baseline trends are globalin nature. Indeed, the baseline findings indicatethat the increasing importance of developingregions in the world economy could lead tounforeseen positive or negative developmentsin one region having far-reaching global con-sequences for overall food security and tradepatterns. The alternative scenarios examinedhere focus on alternative growth scenarios forAsia, because of its importance to globalgrowth, and Sub-Saharan Africa, because ofthe significant risk that this region will fail togrow at the levels projected in the baseline.Regional scenarios examine not only projectedeffects on a particular region, but also theextent to which regional disturbances rever-berate worldwide.

ASIAN SCENARIOS

India and China: Agricultural Growth

Slowdown

Continued agricultural growth in India andChina—even the slowing yield growth pro-

jected under the baseline—depends on a num-ber of assumptions regarding the possibility forcontinued yield increases, moderate (notoverly burdensome) levels of land degradation,and the country’s ability to manage pressuresfrom growing competition among the agri-cultural, urban, and industrial sectors over landand water resources. Many observers believethat these assumptions are not appropriate,and that both nations will face more severeimpediments to the expansion of agriculturalproduction. Indeed, there are a number ofsources of downside risk to crop yield growth,as described in the section on global yieldgrowth scenarios. Given these uncertainties, ascenario invoking a decline of 50 percent inarea and yield growth rates in India and China,a decline of 25 percent in livestock numbersgrowth, and a 25 percent decline in GDPgrowth (the latter due to the economy-wideeffect of the dramatic slowdown in agriculturalgrowth), represents a plausible negative sce-nario for these regions.27 As with the globalyield scenarios presented earlier, the IMPACTmodel translates such changes in yield andgrowth rate assumptions into somewhat dif-

6

ALTERNATIVE REGIONAL

SCENARIOS

112

ferent actual declines in the output value ofthese variables, because the price increasescaused by the lower yield growth induces theyields to recover partially (Table 6.1). Thus,while all assumed values were halved, thesechanges result in an actual decline of Indianannual cereal yield growth rates from 1.2 to0.7 percent for wheat, from 1.0 to 0.6 percentfor maize, from 1.0 to 0.5 percent for othergrains, and from 1.4 to 0.8 percent for rice.Actual Chinese annual yield growth ratesdecline by less than half. Area growth ratesvary little from the negligible rates projectedunder the baseline.

As expected, declining yield and areagrowth rates have a severe effect on total cropproduction in India and China. Indian cerealproduction declines 15 percent, from 254 mil-lion tons under the baseline scenario to 217million tons. As a result, India goes from nearself-sufficiency in cereals in 2020 under thebaseline scenario to imports of 30 million tons(a cost of $5.9 billion at projected internationalprices). Cereal production in China alsodeclines 15 percent from the baseline level,from 518 to 438 million tons. China’s cereal

trade deficit nearly doubles under this sce-nario, increasing from 48 million tons underthe baseline scenario to 89 million tons ($12.7billion).

Moving beyond a narrow focus on cereals,India’s overall agricultural trade balance underthe slow-growth scenario will shift from a 1997surplus of $1.7 billion to a deficit of $9.2 billionin 2020, while China’s overall trade deficit willgrow from net imports of $5.9 billion in 1997 tonet imports worth $33.5 billion in 2020 (Table6.2). In relative value terms, India’s trade deficitin 2020 will represent 7 percent of agriculturalproduction and China’s, 10 percent. These over-all trends represent the combined effects of sub-stantial increases in cereal imports as well as adecline in meat imports in China and a shift fromsmall imports of meat to small exports in India,due to lower demand caused by decliningincome growth rates and higher meat prices.

Area and yield declines in India and Chinaaffect world cereal prices by significant, but notdisastrous, amounts (Table 6.3). Wheat pricesunder this scenario increase 1 percent ($1 perton) between 1997 and 2020 to $134 per ton,compared with an 8 percent ($10 per ton) declineunder the baseline scenario. Maize and riceprices increase sharply under the slow-growthscenario and decline under the baseline scenario.Other grains decline under both scenarios.

Chinese meat production in 2020 under theslow-growth scenario will decline from 100 mil-lion tons under the baseline to 88 million tons,and Indian meat production will decline from9.2 million tons to 8.2 million tons (Table 6.4).The combination of production declines andthe decline in demand from lower incomegrowth will lead to varying price effects amongthe different livestock products (Table 6.5).World prices for beef and sheep and goat meatwill only decline by small amounts, while worldpoultry and pork prices will actually rise slightly.

ALTERNATIVE REGIONAL SCENARIOS 113

TABLE 6.1 Realized annual cereal yield

growth rates under the baseline and India and

China slow-growth scenarios, 1997–2020

Baseline Slow growth

Commodity India China India China

(percent/year)

Wheat 1.2 0.9 0.7 0.5

Maize 1.0 1.6 0.6 0.9

Other

coarse grains 1.0 1.4 0.5 0.7

Rice 1.4 0.8 0.8 0.5


Indian kilocalorie consumption in 2020declines from 2,868 kilocalories per capitaunder the baseline scenario to 2,697 kilocalo-ries per capita under the slow-growth scenario,while Chinese kilocalorie consumption declinesfrom 3,536 kilocalories per capita to 3,272 kilo-calories. More important, under the slow-growth scenario, the number of malnourishedchildren in China and India each will increaseby 2 million children over the baseline. Therate of increase in malnourished children isdependent on the ability and willingness ofIndia and China to finance massive agriculturalimports under this scenario. China, in particu-

lar, would be importing agricultural productsworth 10 percent of total agricultural produc-tion. Political exigencies may render such ahigh level of imports unacceptable. Measuresto lower agricultural import dependency inIndia and China, including high import tariffsand subsidies to domestic production, couldconsiderably worsen the ultimate impact ofnegative shocks to these agricultural produc-tion systems.

A slowdown in technological change in agri-culture and heightened resource degradationin China and India would primarily affect thecountries themselves. The implications of such

114 CHAPTER 6

TABLE 6.2 Net commodity trade under the baseline and India and

China slow-growth scenarios in 2020

India China

Slow-growth Slow-growth

Commodity scenario Baseline scenario Baseline

(US$ billion)

Wheat −2.0 −0.8 −3.0 −1.6

Maize −0.1 0 −5.6 −3.2

Rice −3.5 −0.1 −3.7 0.3

Other coarse grains −0.3 0 −0.5 −0.4

Potatoes −1.3 −0.1 −2.0 −0.3

Sweet potatoes 0 0 −1.7 −0.1

Cassava and others −0.1 0 −0.2 −0.1

Soybeans −0.6 −0.1 −3.3 −3.0

Oils −3.3 −1.2 −5.7 −4.5

Meals −0.3 0.3 −3.2 −2.0

Beef 0.2 −0.1 −0.1 −1.1

Pork 0 −0.1 −2.5 −2.6

Sheep and goat 0 −0.1 −0.3 −0.5

Poultry 0 0 −0.7 −1.6

Milk 2.1 −0.1 −0.4 −0.8

Eggs 0 0 −0.6 −0.1

Total −9.2 −2.4 −33.5 −21.6

––––Source: IMPACT projections, June 2001.Note: Positive figures indicate net exports; negative figures indicate net imports.

High Asian Feed Ratios

The future of feed efficiency trends willdepend significantly on the pace of commer-cialization in the Asian livestock sector. Indus-trial systems, mainly in the developed world,currently produce more than half of all poul-try and pork products, raising them on well-balanced mixtures of prepared commercialfeeds. Livestock systems in the developedworld, which mainly produce for slowly grow-ing domestic markets and neither producetheir own feed products nor utilize nitrogenouswastes, are concerned mainly with raising effi-ciencies to maintain profits. Grazing/mixed-


TABLE 6.3 Crop prices under the baseline

and India and China slow-growth scenarios in

1997 and 2020

1997 2020

Base Slow-growth

Cereal year Baseline scenario

(US$/metric ton)

Wheat 133 123 134

Maize 103 102 111

Other coarse

grains 97 86 94

Rice 285 250 315


TABLE 6.4 Meat production under the baseline and India and China

slow-growth scenarios in 2020

India China

Slow-growth Slow-growth

Commodity Baseline scenario Baseline scenario


Beef 5.3 4.9 10.4 8.9

Pork 1.0 0.9 60.0 54.3

Sheep and goat 1.4 1.2 3.2 3.0

Poultry 1.5 1.2 26.4 21.7

All meat 9.2 8.2 100.0 88.0


TABLE 6.5 Meat prices under the baseline

and India and China slow-growth scenarios in

2020

Slow-growth

Commodity Baseline scenario

(US$/metric ton)

Beef 1,740 1,739

Pork 2,239 2,263

Sheep and goat 2,832 2,811

Poultry 703 704


a production shock for world prices are signif-icant but not devastating. The results also indi-cate that world markets are in fact quiteresilient and can absorb large increases in Chi-nese and Indian cereal imports without hugeprice consequences. Although China in partic-ular is already a significant player in world foodmarkets and is likely to become increasinglyimportant, it does not represent a major threatto the long-term stability of global markets.Global supply response will maintain consid-erable flexibility.

farming systems, where farm households oftenproduce their own feeds and reuse wastes, alsoface strong incentives to improve efficiencies,mainly because of the rapidly growing demandfor pork, poultry, and dairy products in devel-oping countries (Steinfeld 1998). Nevertheless,despite the incentives for higher feed efficien-cies under both industrial and grazing/mixed-farming production systems, compelling evi-dence exists that the shift in livestock productionfrom small-scale households and mixed-farm-ing systems to large-scale commercial facilitiesthroughout much of Asia—the major ongoingand future trend in the regional livestock sec-tor—is leading to significantly higher feed ratios(Steinfeld 1998). Other trends act to offset theimpact on feed ratios of the shift from backyardto commercial production, particularly theongoing improvement in feeding efficiencywithin the commercial sector and the shift toless feed-intensive poultry production.

The baseline scenario incorporates a mod-erate shift in the livestock production structure,but it is possible that the pace of the transitioncould proceed more rapidly than anticipated.Notably, while Asian feed demand growth at2.7 percent annually is still expected to behigher than meat production growth at 2.6 per-cent annually during 1997–2020 under the base-line, this differential is rather small, and South-east Asian meat production growth at 2.9percent annually is significantly higher thanannual feed demand growth of 2.4 percent. Ris-ing feed ratios in East Asia, particularly China,are in large part responsible for the overall Asiantrend, with cereal feed demand growth aver-aging 2.7 percent annually and meat produc-tion growth averaging 2.5 percent.

The faster growth in meat demand relativeto feed demand in Southeast Asia stems from anumber of factors. First, relatively more effi-cient poultry production is projected to accountfor 42 percent (4 million tons) of meat demand

growth in Southeast Asia in 1997– 2020, com-pared with China’s share of 34 percent (16 mil-lion tons) over the same period. Structuralchanges in livestock production have also pro-gressed more rapidly in Southeast Asia than inEast Asia, so a significant structural shift frombackyard to commercial feeding ratios hasalready occurred. Both Indonesia and Thailandhave undergone significant poultry productionindustrialization over the last 30 years—Indone-sia since the mid 1970s, spurred by governmentproduction intensification programs such asBIMAS (Indonesia’s Agricultural Services andIntensification Agency), and Thailand throughproduction expansion and technology develop-ment led by the private sector. Thailand’s poul-try sector is now highly capital intensive andproductive, and its poultry exports are signifi-cant (Soedjana 1999; Ranong 1999).28 The rapiddevelopment of Southeast Asia’s industrial live-stock sector is indicative of the general trend ofindustrialization of the poultry sector preced-ing industrialization of the pork and dairy sec-tors (Steinfeld and Kamakawa 1999). The porksector represented 69 percent of total agricul-tural production in East Asia in 1997, with lessthan 20 percent of the sector characterized byintensive production (Bingsheng 1997).29 Thegreater importance of poultry to regional meatproduction led Southeast Asia to user fewercereals per kilogram of meat production thanEast Asia in 1997, with Southeast Asia requiring15 million tons of cereals to produce 9 milliontons of meat (a ratio of 1.7), while East Asiarequired 119 million tons of cereals to produce55 million tons of meat (a ratio of 2.2). Never-theless, East Asian smallholders—especiallypork producers—use more alternative feedcrops than relatively more industrialized South-east Asia, consuming 74 million tons of rootsand tubers for feed purposes (a ratio of 1.4),compared with Southeast Asian consumptionof only 2 million tons (a ratio of 0.2). While

116 CHAPTER 6

smallholder livestock production will not dis-appear in East Asia over the next two decades,it will undergo significant pressures fromdemand, population, and income growth.

Because the transition to higher feed ratiosin this alternative scenario represents anincrease in the pace of commercialization inlivestock production, rates of growth of live-stock production in Asia should also increasealong with any projected increase in feedratios. To simulate the shift in technology

toward higher feed ratios in Asia, we raisedmaize feed ratios by 10 percent and increasedthe production growth for poultry, pork, andeggs in all Asian countries by an additional 0.3percent per year.

These modifications will lead to an increasein Asian livestock production of 4 percent (5million tons) above levels predicted by the base-line scenario in 2020 (Table 6.6), as well as adecline of net livestock imports into Asia of 3.8million tons, from 6.3 million tons. Productionincreases are fairly uniform across regions and commodities. Higher Asian livestock pro-duction will cause some movement in meatprices, with pork prices falling 3 percent ($72per ton), poultry prices falling 1 percent ($10 perton), and egg prices falling 4 percent ($49 per ton) (Table 6.7). Meat products that are notaffected by higher rates of industrialization gen-erally experience a slight price increase abovebaseline levels under the low feed ratio sce-nario, although generally less than 1 percent.

The important ramifications of higherAsian feed ratios manifest themselves in grainmarkets (Table 6.8). It is possible that the shiftfrom traditional feeds—including household


TABLE 6.6 Meat production under the

baseline and high Asian feed ratio scenarios in

1997 and 2020

1997 2020

Base High Asian

Region year Baseline feed ratio


South Asia 8 15 15


East Asia 55 103 107

All Asia 72 137 142


TABLE 6.7 Meat and dairy prices under the

baseline and high Asian feed ratio scenarios

in 1997 and 2020

1997 2020

Base High Asian

Commodity year Baseline feed ratio

(US$/metric ton)

Beef 1,808 1,740 1,750

Pork 2,304 2,239 2,167

Sheep and goat 2,918 2,832 2,845

Poultry 735 703 693

Eggs 1,231 1,191 1,142

Milk 318 289 292


TABLE 6.8 Cereal prices under the baseline

and high Asian feed ratio scenarios in 1997

and 2020

1997 2020

Base High Asian

Commodity year Baseline feed ratio

(US$/metric ton)

Wheat 133 123 126

Maize 103 102 119

Other coarse

grains 97 86 91

Rice 285 250 256


effects on the two regions most dependent onmaize for food—Sub-Saharan Africa and LatinAmerica—where lower cereal consumptionleads to predicted increases in child malnutri-tion that are small but certainly not trivial, withthe number of malnourished children underthe age of five rising by half a million childrenfor the two regions combined.

As these results indicate, technologicalchange in the Asian livestock sector towardhigher feed intensity is important, but theimplications of this change for world cerealprices should not be overemphasized. Never-theless, it is noteworthy that maize, by far themost important feed grain in Asia, is also anessential grain for food security in Latin Amer-ica and Africa, and that rising demand for thiscrop in the Asian livestock sector could pricesome consumers in these countries out of themarket.

AFRICAN SCENARIOS

A Troubled Continent

The baseline scenario projects growing agri-cultural import needs for Sub-Saharan Africain 2020. The region’s $6.5 billion in projected2020 agricultural imports represents 8.9 per-cent of projected regional agricultural pro-duction, a slight increase above an agriculturalimport burden of 7.8 percent of regional agri-cultural production in 1997. Even theseincreases, however, rest on a fairly optimisticset of assumptions, given the recent agricul-tural performance of the region. The baselinescenario projects an annual cereal productiongrowth rate of 2.4 percent for maize, 2.8 per-cent for other grains, 2.9 percent for wheat,and 3.1 percent for rice, aggregating to a totalannual cereal production growth rate of 2.5percent. Roots and tubers production is moreimportant to Sub-Saharan Africa than to any

118 CHAPTER 6

TABLE 6.9 Maize food demand under the

baseline and high Asian feed ratio scenarios

in 2020

High Asian

Region Baseline feed ratio


Latin America 29 27

Sub-Saharan Africa 39 37

West Asia/North Africa 8 8

South Asia 13 13

Southeast Asia 12 12

East Asia 37 36


and farm wastes and roots and tubers cropssuch as sweet potatoes—to a higher propor-tion of grain feeds in Asia will cause signifi-cant upward price pressure in world grainmarkets. Raising the feed ratio for maize does,as expected, raise the price of maize by 17 per-cent above the baseline level in 2020. The effectof this price increase on maize food con-sumption is proportionally smaller, with foodconsumption declining 7 percent in LatinAmerica, 5 percent in Sub-Saharan Africa, 3percent in East Asia, and by negligibleamounts in other regions (Table 6.9). Theincrease in Asian feed demand that precipi-tated these maize price increases, however,leads to an increase in total maize demand of25 percent in East Asia (63 million tons), 15percent (6 million tons) in Southeast Asia, anda negligible amount in South Asia. Thus, aswould be expected from its position as themain livestock producer in the world, the ratesof technological and institutional change inthe Chinese livestock sector do influenceworld markets.

Declining consumption of maize for foodas a result of higher prices will have its main

other region, and production growth rates ofthese crops are also projected to average a rel-atively high 2.4 percent annually between 1997and 2020. The projected baseline rates ofgrowth for cereal crops are slightly below theannual rate of increase in cereal production of3.6 percent achieved in the region between1982 and 1997. However, this rapid rate ofgrowth coincided with the continent’s recov-ery from dismal performance between 1967and 1982, when cereal production grew at arate of only 1.8 percent annually. Projectedannual roots and tubers production growthrates also represent a decline from growthrates averaging 4.3 percent between 1982 and1997, and the trend of recovery from previousslow growth is even more extreme than forcereals, since roots and tubers production onlygrew at a rate of 1.8 percent between 1967 and1982.

As was shown in the earlier historical sec-tion, the sources of past growth in Africancereal production cause some concern for thefuture, since area expansion has driven pro-duction growth in the context of stagnantyields for most cereals in recent years. Whilethe area under cereal production in Sub-Saha-ran Africa expanded by 1.9 percent per yearover the period 1967–97, and by a rapid 3.5 per-cent per year between 1982 and 1997, cerealyield growth rates only averaged 0.8 percentper year during 1967–97 and a dismal 0.1 per-cent per year between 1982 and 1997. Part ofthe reason for poor yield performance in thepast has been the fact that rapidly expandingarea under cereal production—and the con-comitant ability to practice low-intensity cul-tivation—obviated the need for rapid yieldimprovements. Area under crop productionincreased 63 percent during 1967–97. Between1982 and 1997 alone, it surged from 94 millionhectares to 153 million hectares. Regionalgrowth in cropped area was fastest in North-

ern Sub-Saharan Africa with an increase of 93percent during 1967–97 to 45 million hectaresand slowest in Southern Sub-Saharan Africawith an increase of only 30 percent to 16 mil-lion hectares. Increases in cropped area overthe 30-year period are not, however, a goodindicator of current levels of land scarcity,since the ratio of arable and permanent crop-land to total agricultural land is particularlyhigh in Nigeria, Central and Western Sub-Saharan Africa, and Eastern Sub-SaharanAfrica, and particularly low in Southern andNorthern Sub-Saharan Africa.

Under the baseline, crop production in Sub-Saharan Africa will continue to expand ontopreviously unused land, although at a slowerrate than in the past. The increase in the areaunder cereal and roots and tubers cultivationranges from 21 percent in Nigeria to 38 per-cent in Central and Western Sub-SaharanAfrica. Nevertheless, while Africa was a landsurplus continent in the 1960s, with agricul-tural production mainly limited by sources oflabor and capital, rapid population growth hasraised population densities significantly acrossmany regions, thus circumscribing the oppor-tunities for further area expansion. Per capitaarable area regionwide has declined from 0.48hectares per capita in 1967 to 0.25 hectares in1997. Cleaver and Schreiber (1994) point outthat the figures for per capita arable area mayeven overstate the land available for cultivationbecause the figures may include land with verypoor agroclimactic and soil conditions. Despitethe increasing pressure, a detailed FAO (1994)study, summarized in Table 6.10, reveals thatthere is room for some additional area expan-sion into land of varying quality in many coun-tries across the region, perhaps with the excep-tion of water-constrained parts of NorthernSub-Saharan Africa. Nevertheless, proper tech-nologies and investment would enable someland expansion even in the north, where the


area under cereal and roots and tubers pro-duction almost doubled between 1974 and1997.

The IMPACT baseline projects relativelystrong yield growth averaging 1.5 percent forcereals in Sub-Saharan Africa, essentially dou-bling the yield growth achieved between 1967and 1997. Despite the pessimism that prevailsabout the future of agricultural productivitygrowth in Sub-Saharan African and the dismalhistorical record, such strong cereal yieldgrowth vindicates our belief that currenttrends in the region portend modest improve-ment over past performance. African yields arevery low even by other developing countrystandards, indicating that significant growthshould be possible if countries in the regionmove toward appropriate technologies, poli-cies, and programs. Although economic liber-alization has had mixed results, the policy envi-ronment is now more favorable to agriculture

than it was in the 1970s and 1980s, with pricediscrimination reduced in many countries andmarket liberalization providing the necessaryincentives for private-sector participation inthe agricultural sector. Most African govern-ments seem to have realized that agriculturalgrowth will be the key to economic develop-ment in the region in the medium term andhave embarked their countries upon develop-ment paths that support the increased pro-ductivity of the sector (World Bank 2000a).

As described above, theories of inducedtechnological innovation predict that growingpopulation pressure will lead to higher yieldgrowth rates as low-input agriculture increas-ingly ceases to be a viable option (Boserup1981). Increased population pressure in Sub-Saharan Africa is shown by the decline in totalper capita cropped area from 0.40 hectares percapita in 1967 to 0.27 hectares per capita in1997.30 However, population-induced innova-

120 CHAPTER 6

TABLE 6.10 Countries in Sub-Saharan Africa with significant amounts of

remaining high-potential arable land

Central and

Northern Western Southern Eastern

Chad Benin Angola Tanzania

Ethiopia Cameroon Lesotho Uganda

Mali Central African Republic Madagascar

Sudan Congo Republic Malawai

Ivory Coast Mozambique

Democratic Republic Swaziland

of Congo Zambia

Gabon Zimbabwe

Ghana

Guinea

Liberia

Nigeria

Togo

––––Source: FAO 1994.

tion requires access of land-constrained popu-lations to technologies and inputs. Cleaver andSchreiber (1994) point out that even as popu-lations throughout Sub-Saharan Africa are los-ing their ability to practice shifting cultivationdue to high population densities, they continueto practice other elements of extensive culti-vation, including low levels of technologicaland capital inputs, traditional land tenure andland husbandry practices, and traditionalmethods of resource acquisition. Variousstructural factors mentioned earlier have hin-dered the uptake of the methods necessary foragricultural intensification by local farmers.Exchange rate, tax, trade, and pricing policiesin the context of strict government controlhave squeezed out the private sector and sti-fled farmer investment in agricultural land(Cleaver and Schreiber 1994).

Despite the crucial importance of animproved policy framework that allows themarket to operate freely, it will not launch Sub-Saharan African agriculture on a strong andsustainable growth path without proactivemeasures at the national and internationallevel to ensure more widespread diffusion oftechnological solutions and more intensiveinput application across the region. Forinstance, Sub-Saharan African governmentsmust continue to focus substantial and well-planned effort on stimulating much higher lev-els of fertilizer use. While some researchers(Versteeg, Adegbola, and Koudokpon 1993;Janssen 1993) have pointed to the suitability oflabor-intensive techniques such as legumerotations, animal manures, and alley croppingas potential short-term fertilizer substitutes incases of labor surplus, the conclusion thatchemical fertilizer use must increase rapidly ifregional agricultural production is to respondto technological innovation is unavoidable(Byerlee and Heisey 1996). Unwise fertilizersubsidies in the 1970s and early 1980s brought

high-cost marketing and limited choice, spur-ring increased use only among politicallypowerful farmers. Subsidy reform in the1980s—along with currency devaluation andhigh world prices—led to significant local priceincreases and reduced consumption from 3.7million tons in 1988/89 to 3.5 million tons in1994/95, with more than half of the countriesin the region depending on foreign aid to sup-ply their fertilizer needs (Bumb and Baanante1996).

While the removal of subsidies is necessaryto stimulate private-sector participation in themarket, the benefits have not yet been realizeddue to a variety of factors, including trade bar-riers, political indifference, foreign exchangeshortages, low crop prices, and a lack of insti-tutional and physical infrastructure (WorldBank 2000a). A significant level of public-sector commitment will be necessary toachieve higher fertilizer use, with areas in needof focus including macroeconomic stability,price incentives, credit support, organizationalefficiency, and much higher investment ininfrastructure. Governments must encouragefertilizer use in high-potential areas, put inplace proper measures to ensure environmen-tal sustainability, and address the high cost offertilizers by lowering transport costs and rais-ing scale-economies of international purchas-ing and shipment (Byerlee and Heisey 1996;Bumb and Baanante 1996). Given that theregional supply potential is only 8.4 milliontons, Sub-Saharan Africa will have to importlarge quantities of fertilizer over the foresee-able future if it is to dramatically increase con-sumption, thus necessitating stable and timelysupplies of foreign exchange (Bumb andBaanante 1996).

The IMPACT baseline projections requirelarge investments in roads, irrigation, cleanwater, and education in Sub-Saharan Africa(see Chapter 7). Because private-sector


research in Africa is relatively undeveloped, thepublic sector is likely to be responsible for thevast majority of innovation over the next 20years (Byerlee and Heisey 1996). Areas in needof attention include the development of nutri-ent management systems for specific soils, low-cost soil rehabilitation techniques, methods forincorporating perennial crops in farming land-scapes, and innovative incentive structures toencourage long-term conservation of forestand grazing land (Scherr 1999). Byerlee andHeisey (1996) point out that the record of high-yielding maize research in Sub-Saharan Africaoffers some hope for the future—with greatpotential for raising food crop production, par-ticularly in the savanna and mid- and high-altitude areas—although “successful maizeresearch programs in Eastern and SouthernAfrica required a decade or more of sustainedeffort to produce varieties and hybrids thatwere widely adopted.” Lack of staff continu-ity and breeding strategies in many nationalprograms have been mainly responsible for thewidespread perception that local researchefforts have not been successful, althoughAlston et al. (2000) calculated the mean rate ofreturn to local research at a very respectable49.6 percent.

Future research efforts may require agreater diversification from a focus on maizeand explore opportunities for alternative cropssuch as cassava and rice that have particularproblems associated with African agroeco-logical conditions. Technological diffusion hasproven a major problem in Sub-Saharan Africa,with Goldman and Block (1993) identifying anumber of commodities for which high-yielding varieties exist but are underutilized,including cassava (with potential yieldincreases of 50 percent on half of currentlyplanted area), sweet potato, and rice (irrigatedand mangrove environments) (cited in Spencer1994). Nevertheless, the experience with maize

in Sub-Saharan Africa shows that small farm-ers will make use of improved seeds and com-plementary inputs provided the technology,infrastructure, and overall macroeconomicenvironment are appropriate. In order toaddress the conditions particular to the region,the development of improved technologypackages for all crops needs to place a pre-mium on efficient input use and maximizingreturns to labor and cash during early adop-tion. Above all, effective research must beembedded within an overall framework foragricultural development that emphasizessmallholder commercialization, private-sectorinitiative at all levels, decentralized public par-ticipation, trade, and poverty alleviation(World Bank 2000a).

Underlying the baseline IMPACT projec-tions is the implicit assumption that currentdevelopments on the continent offer theprospect of some amelioration of past trends.We take hope in the reformist currents evidentduring the 1990s throughout much of theregion and note that agriculture became farmore of a priority than it was during theimport-substituting, industry-centered days ofthe 1960s, 1970s, and early 1980s. Agriculturehas responded to macroeconomic reform andexport crop liberalization in the 1990s, with 12countries reporting agricultural GDP growthrates of 4 percent or more between 1990 and1997, and 5 more countries added after 1993(Table 6.11) (World Bank 2000a). Agriculturalvalue-added per worker also increased in 19out of 31 countries between 1979–81 and1995–97, with West Africa showing particularlystrong growth as the use of bovine animal trac-tion spread (World Bank 2000a). It is our con-tention that as the general environment con-tinues to improve—with policymakers stilldedicated to improving policies and institu-tions—agricultural productivity will continueto improve moderately as well.

122 CHAPTER 6

Despite these optimistic signs, there isarguably a very real possibility of agriculturalstagnation and deepening poverty in Sub-Saharan Africa. And recognition of a problemdoes not inevitably lead to effective action.Despite some successes, many countries inSub-Saharan Africa have struggled with thetransition to private-sector input provision andoutput markets, and uncertainty in input sup-ply and producer prices remains a major fac-tor limiting the ability of smallholders to planinvestments with long time horizons (Byerleeand Eicher 1997). A number of factors lead toconcern over the ability of regional produc-tion systems to even maintain the gains thathave been made over the last 30 years, includ-ing severe soil fertility depletion and erosionin many areas farmed without appropriatenutrient replacement or conservationist prac-tices, pest problems, overdependence on maizemonoculture, and highly variable water avail-ability. Worldwide, the decline of funding forinternational agricultural research has placedincreasing responsibility for technologicalbreakthroughs in agriculture on the privatesector, and it is likely that the great majority ofthese advances will be unsuited to the agro-climatic environment of Sub-Saharan Africa.A long record of failure and concern over theavailability of exploitable water supplies indi-

cates that the potential for irrigation expansionin Sub-Saharan Africa is limited, at least overthe medium term.

The environmental consequences of popu-lation growth and agricultural intensificationunder conditions of low input application andtechnology dissemination deserve furtherattention, given their centrality to theprospects for realization of the IMPACT base-line assumptions. Soil degradation—particu-larly erosion and nutrient depletion—hasbecome a severe threat on both marginal landsundergoing population-induced permanentcropping and high-quality rainfed lands withhigh population densities. A Global Assess-ment of Soil Degradation (GLASOD) studyfound that water and wind erosion and chem-icals had degraded 65 percent of soils on agri-cultural land in Sub-Saharan Africa since the1950s, with serious degradation affecting 19percent of all land (Oldeman et al. 1991). Sup-porting the need for far higher rates of fertil-izer application, Stoorvogel et al. (1993) extrap-olated in a continent-wide study of soilnutrient depletion that average annual nutri-ent losses over the last 30 years equaled 1.4 tonsper hectare of urea fertilizer, 375 kilograms perhectare of triple superphosphate, and 896 kilo-grams per hectare of potassium chloride(Scherr 1999). A variety of studies consistentwith these figures indicate that productivitylosses from soil degradation since World WarII have amounted to 25 percent for cropland,with African farm survey data showingdeclines in grain yields from 2–4 tons perhectare to 1 ton per hectare on originally fer-tile land (Oldeman 1998; Sanchez et al. 1997,as cited in Scherr 1999). Crop yield losses frompast erosion show cumulative crop yield reduc-tions that range from 2 to 40 percent across allSub-Saharan African countries, with a mean of6.2 percent for the region (Scherr and Yadav1996). Bojo (1996) estimates that economic loss


TABLE 6.11 Number of countries achieving

agricultural GDP growth of 4 percent or

more in Sub-Saharan Africa, 1980–97

Region 1980–90 1990–97 1993–97

Northern 0 2 4

Central and Western 3 6 7

Southern 0 4 6

Eastern 0 0 0

––––Source: World Bank (2000a).

from soil degradation ranges from less than 1percent of agricultural GDP in Madagascar,Mali, and South Africa to 2 to 5 percent of agri-cultural GDP in Ethiopia and Ghana to morethan 8 percent in Zimbabwe.

Soil degradation may indeed remain a seri-ous problem during the IMPACT projectionsperiod, particularly on marginal lands withrapidly growing populations in parts of theSahel, mountainous East Africa, and the drybelt stretching from the coast of Angola toSouthern Mozambique (Cleaver and Schreiber1994). For instance, Lal (1995) predicts thatwater erosion alone will reduce crop produc-

tivity in Sub-Saharan Africa by 14.5 percentbetween 1997 and 2020 (Scherr 1999).

Pessimistic Africa Scenario

Given the downside risks for agriculture, it isimportant to assess the implications of analternative, pessimistic scenario for Sub-Saharan Africa. In order to quantify a pes-simistic scenario for Africa in the IMPACTmodel, we cut projected crop area and yieldgrowth by 50 percent and reduced the growthin livestock, milk, and egg numbers by 30 per-cent (Table 6.12). Additionally, because agri-cultural production is so important to African

124 CHAPTER 6

TABLE 6.12 Production assumptions for Sub-Saharan Africa

under the optimistic and pessimistic scenarios

Growth rate Optimistic Pessimistic

Cereal and oilcrop yield Add 100% Subtract 50%

Cereal and oilcrop area No change Subtract 50%

Roots and tubers yield Add 50% Subtract 50%

Roots and tubers area No change Subtract 50%

Livestock yield No change Subtract 30%

Livestock area Add 30% Subtract 30%

––––Source: Authors.

TABLE 6.13 GDP growth rates in Sub-Saharan Africa under the

baseline, optimistic, and pessimistic scenarios, 1997–2020

Region Baseline Optimistic Pessimistic

(percent/year)

Nigeria 3.8 8.0 1.9




Eastern 3.5 8.0 1.8


economies, GDP growth rate assumptionsbetween 3.2 and 3.8 percent annually are cutby 50 percent (Table 6.13). Last, a variety ofsocial indicators are also expected to worsen,with national schooling and sanitation fallingby 10 percent and the life expectancy ratiofalling by 4 percent from baseline levels. Withpopulation growth projected at the UN’smedium variant level (UN 1998), regional percapita GDP in the pessimistic Africa scenariowill decline from $339 in 1997 to $297 in 2020,representing a reversal of the increase in percapita GDP growth of 1.2 percent a year pro-jected under the baseline scenario.

The pessimistic Africa scenario naturallywill adversely affect overall agricultural pro-duction, reducing cereal output in 2020 to 96million tons, or 26 percent below the baselinelevel, as cereal production growth declines to

1.3 percent annually (Table 6.14). Roots andtubers production will only reach 201 milliontons, or 20 percent below the baseline level(Table 6.15). The total agricultural productionvalue of all IMPACT commodities, estimatedunder the baseline scenario to increase 79 per-cent above 1997 values to reach $73 billion in2020, will fall to $66 billion under the pes-simistic Africa scenario. Slowing area and yieldgrowth will share responsibility for the short-fall in agricultural production, although bothcereal and roots and tubers area growth willremain slightly positive throughout the regionat 0.5 percent annually (Table 6.16).

These rates of area growth represent sharpdeclines from 1.9 percent annual cereal areagrowth and 2.2 percent annual roots and tubersarea growth achieved between 1967 and 1997,and reflect rising land scarcity and degradation


TABLE 6.14 Cereal production and growth rates in Sub-Saharan Africa

under the baseline, optimistic, and pessimistic scenarios in 2020



Cereal production

Nigeria 37 54 28

Northern 38 55 28


Southern 18 24 13

East 16 23 12

All Sub-Saharan Africa 129 187 96

Cereal growth rates (percent/year)

Nigeria 2.4 3.9 1.3




Eastern 2.4 3.9 1.3



under the pessimistic Africa scenario. Cerealyield growth will drop to 0.8 to 1.0 percentannually in all regions under the pessimisticAfrica scenario, commensurate with the 0.8 per-cent annual yield growth rate achieved region-wide during 1967–97 (Table 6.17). Sharply lowerarea growth and yield growth similar to the his-torical average represent a plausible pessimisticscenario for Sub-Saharan Africa.

On the trade side, Sub-Saharan Africa’s netcereal imports of $3.9 billion (27 million tons)in 2020 under the baseline scenario willincrease to $5.4 billion (39 million tons) under

the pessimistic Africa scenario, reaching a levelthat may not be sustainable given foreignexchange constraints (Table 6.18). Sub-SaharanAfrica’s food imports will expand precipitouslyto $11 billion of net agricultural imports in2020, representing 17 percent of the value ofagricultural production in that year.

The pessimistic Africa assumptions wouldlead to significantly worse malnutrition in Sub-Saharan Africa (Table 6.19). Overall, per capitakilocalorie availability would decline from2,232 kilocalories in 1997 to 2,162 kilocaloriesin 2020, reversing the slight improvement of

126 CHAPTER 6

TABLE 6.15 Roots and tubers production in Sub-Saharan Africa under

the baseline, optimistic, and pessimistic scenarios in 2020



Nigeria 102 126 82

Northern 10 12 8


Southern 24 29 21

East 30 37 23

All Sub-Saharan Africa 252 307 201


TABLE 6.16 Sum of cereal and roots and tubers area cropped in Sub-Saharan Africa under

the baseline, pessimistic, and optimistic scenarios, 1997 and 2020

Region 1997 Baseline Optimistic Pessimistic

(million hectares)

Nigeria 24 29 29 27

Northern 30 37 37 34

Central and Western 16 22 21 19

Southern 11 14 14 13

Eastern 9 12 12 11



TABLE 6.18 Agricultural trade balances in Sub-Saharan Africa under the

baseline, pessimistic, and optimistic scenarios in 2020

Commodity Baseline Optimistic Pessimistic

(US$ billion)

Wheat −1.7 −3.0 −1.5

Maize −0.7 −0.3 −1.1

Rice −1.4 −1.5 −1.9

Other coarse grains −0.1 1.0 −0.9

Potatoes 0.0 −0.4 −0.1

Sweet potatoes and yams 0.0 0.2 −1.2

Cassava and other roots and tubers 0.0 0.3 −1.8

Soybeans −0.1 −0.2 0.0

Oils −1.0 −1.6 −1.2

Meals 0.0 0.4 −0.1

Beef −0.1 −7.4 0.3

Pork −0.2 −2.2 −0.1

Sheep 0.2 −0.1 −0.4

Poultry −0.2 −1.2 −0.1

Milk −1.3 −8.0 −1.2

Eggs 0.0 −0.5 −0.1

Total −6.5 −25.0 −11.0

––––Source: IMPACT projections, June 2001.Note: Positive figures indicate net exports; negative figures indicate net imports.

TABLE 6.17 Cereal yield growth rates in Sub-Saharan Africa under the

baseline, pessimistic, and optimistic scenarios, 1997–2020

Region/Country Baseline Optimistic Pessimistic

(percent/year)

Nigeria 1.6 3.3 0.8




East 1.5 3.1 0.8



210 kilocalories achieved between 1997 and2020 under the baseline scenario. The numberof malnourished children in Africa would alsoincrease from 33 million in 1997 under thebaseline scenario to 49 million children (Table6.20). If budgetary and foreign exchange con-straints rendered Sub-Saharan Africa unable toincrease food imports to the high levels gener-ated under the pessimistic scenario, the num-ber of malnourished children in the regionwould increase still further.

A high level of food insecurity and heavydependence on food aid characterizes Africatoday. The region can ill afford further declinesin per capita income and simply does not pos-

sess the necessary foreign exchange to satisfy

its food needs through imports. While the

assumptions of the pessimistic Africa scenario

do result in a substantial increase in net imports

of agricultural commodities, especially cereals,

to help mitigate the projected production gap,

the region still will experience a significant

decline in per capita kilocalorie availability and

a significant increase in child malnourishment

even at these high and potentially economically

unsustainable import levels.

An assessment of the investment require-

ments under the pessimistic Africa scenario is

presented in Chapter 7 of this volume.

128 CHAPTER 6

TABLE 6.19 Kilocalorie consumption in Sub-Saharan Africa under the baseline,

pessimistic, and optimistic scenarios in 2020

All

Central Sub-Saharan

Scenario Northern and Western Southern Eastern Nigeria Africa

(kilocalories per capita)

Baseline 2,317 2,370 2,219 2,170 3,168 2,442

Optimistic 3,012 2,976 3,207 2,838 4,323 3,232

Pessimistic 2,090 2,117 1,946 1,933 2,721 2,162


TABLE 6.20 Numbers of malnourished children in Sub-Saharan Africa under the

baseline, pessimistic, and optimistic scenarios in 2020

All

Central Sub-Saharan

Scenario Nigeria Northern and Western Southern Eastern Africa

(millions)

Baseline 7.5 13.7 8.6 4.3 5.2 39.3

Optimistic 4.4 9.0 4.5 1.4 2.5 22.0

Pessimistic 9.3 16.0 11.9 5.7 6.6 48.6


Optimistic Africa Scenario

While IMPACT baseline results predict anincrease in the number of malnourished chil-dren in Sub-Saharan Africa from 33 millionchildren in 1997 to 39 million in 2020, the devel-oping world, excluding Sub-Saharan Africa,will achieve a decline in the absolute numberof malnourished children of 31 percent, from134 million children in 1997 to 92 million chil-dren in 2020. What kind of transformations interms of economic and agricultural growth,education, and health will be necessary forSub-Saharan Africa to battle childhood mal-nutrition as effectively as the rest of the devel-oping world? Is it feasible to hope that even amajor economic transformation on the conti-nent could lead to real inroads against child-hood malnutrition?

In order to provide a sense of the magni-tude of the challenges facing Sub-SaharanAfrica and the necessity for national govern-ments and international organizations to payconcerted attention to these challenges, wepresent the following scenario as representa-tive of what will be needed to achieve a 33 per-cent reduction in the number of malnourishedchildren in Sub-Saharan Africa, that is, to gofrom 33 million children in 1997 to 22 millionin 2020 (Tables 6.19 and 6.20). Such a reduc-tion would bring malnutrition trends in Sub-Saharan Africa more in line with thoseprojected elsewhere in the developing world.To achieve such a dramatic improvement inchildhood malnutrition, total GDP growthin the five Sub-Saharan African subregionsin IMPACT (Nigeria, Northern, Central andWestern, Eastern, and Southern) would haveto increase from the baseline-estimated rangeof 3.2 to 3.8 percent per year during 1997–2020to an annualized rate of growth of 8 percent.Realized crop yield growth rates regionwidewould have to rise from between 1.3 and 1.8

percent per year under the baseline to an annu-alized rate of 2.7 to 3.6 percent in 1997–2020.31

Yield growth rates of this magnitude would benecessary to satisfy rising cereal demand onthe continent without increasing the level ofcereal imports. If yield growth rates werelower, higher cereal imports could meet thedeficit between regional demand and supply,although it is difficult to see how the regioncould achieve 8 percent GDP growth withouttremendous growth in agricultural productiv-ity. Foreign exchange constraints, of concernin the baseline and pessimistic Sub-SaharanAfrica scenarios, would probably not restrictmoderate levels of net cereal imports if theeconomies of the region were growing at arate of 8 percent per year.

The high yield rates (and high investmentrequirements presented in Chapter 7) neededto generate an optimistic Africa scenario serveas a caution against assuming that agriculturewill take off in Sub-Saharan Africa the same asit did during the Green Revolution in Asia.Various authors have argued the necessity ofseeking different paths to rapid and sustainableagricultural productivity growth. Spencer(1994) points to higher seasonal labor produc-tivity, new production technologies relyingminimally on transported inputs, new tech-nologies for disease and pest resistance, and afocus on rainfed systems as keys to overcom-ing the unique challenges that Sub-Saharanagriculture poses. A perfect example of thetype of low-cost technology that can play arole in overcoming seemingly daunting obsta-cles is the bicycle. In the Asian countryside,ownership of bicycles often approaches 40percent. Bicycle ownership rates in Sub-Saharan Africa are only about 3.5 percent dueto the paucity of local production and variousgovernment import impediments in the formof taxes and surcharges (Hayami and Platteau1997).


Growing evidence has accumulated in recentyears indicating that GDP and agricultural pro-duction growth alone are not sufficient to drivesignificant improvements in childhood malnu-trition (Smith and Haddad 2000). A variety ofadditional variables concerning the status andeducation of women and the access of the gen-eral population to clean water have proven cru-cial determinants of childhood well-being, andthese would have to improve drastically in Africato significantly improve malnutrition. Under theoptimistic Africa scenario, rates of femaleschooling are projected to increase by 20 per-cent, compared with the baseline, with cleanwater access and female life expectancy bothincreasing by 10 percent. Achievement of theimpressive advances in quality-of-life indicatorsrepresented by the optimistic Africa scenariowould require a tremendous level of commit-ment and investment at all levels, and a majoreffort to focus on the status, education, andhealth of women. Neither African governmentsnor the international community have yet dis-played a willingness to provide the investmentsthat would be necessary to establish a decentlife for the great many impoverished and mal-nourished children in the region. Tragically, therapidly increasing costs of HIV/AIDS in mostof Sub-Saharan Africa will only raise the chal-lenges impeding improved human developmentin the region. Brown (1996) presents the remark-able finding that while HIV/AIDS does notimpact overall population growth rates to a sig-nificant degree, the disease could lower percapita GDP growth rates in severely affectedSub-Saharan Africa by 0.6 to 1.4 percent annu-ally. Any outcome in this range would obviouslyrepresent a huge blow to the prospects for sus-tainable economic development.

The optimistic Africa scenario shows thatsignificant poverty alleviation in Sub-SaharanAfrica is an immense task that will require anequally immense level of commitment on both

130 CHAPTER 6

the national and international levels. East andSoutheast Asia have realized impressive gainsover the last three decades in combatingpoverty and generating economic growth, anda similarly remarkable transformation will benecessary for real improvements in childhoodmalnutrition to take place in Sub-SaharanAfrica. For example, substantial crop yieldgrowth at the level indicated by the optimisticAfrica scenario would require an estimated 8to 10 percent annual growth in fertilizer useacross the region, a level commensurate withthe 9 percent annual growth achieved in Asiabetween 1959/60 and 1994/95 (Bumb andBaanante 1996).32 For an idea of what will berequired in the way of investments under theoptimistic Africa scenario, see Chapter 7.

HIGH MEAT DEMAND IN INDIA—TURMOIL

IN WORLD MARKETS?

Many have voiced concern that revolutionarygrowth in meat demand in India over the com-ing decades could lead to massive Indianimports of meat or cereals (for feed), withresultant upward pressure on world prices (see,for example, Bhalla, Hazell, and Kerr 1999).Current levels of meat and egg consumptionin India are quite low, even taking into accountthe country’s low per capita incomes. Indiansconsumed only 4 million tons of meat and 1million ton of eggs in 1997, translating intoannual per capita consumption of 4.4 kilo-grams of meat and 1.6 kilograms of eggs. Bycontrast, Pakistanis consumed 13.6 kilogramsper capita of meat per year and 2.1 kilogramsper capita of eggs in 1997, and the Chinese con-sumed 38.6 kilograms of meat per year and 5.6kilograms of eggs, albeit at a significantlyhigher level of per capita income. On the otherhand, per capita consumption of milk in Indiais high at 71.2 kilograms, with only Pakistaniconsumption higher in Asia at 139.8 kilograms.

The IMPACT baseline predicts significantincreases in Indian meat and dairy demand,with income elasticities of demand compara-ble to those in other Asian countries. However,because meat consumption levels will be lowinitially, total Indian meat demand will onlyincrease to 9 million tons in 2020, which trans-lates into per capita demand of 7.1 kilo-grams—still very low. By contrast, per capitameat demand will reach 17.3 kilograms in Pak-istan and 64.4 kilograms in China by 2020.Indian egg demand will also increase under thebaseline, reaching 2.5 kilograms per capita in2020. This figure will be higher than in Pak-istan, where per capita egg demand will onlyincrease to 2.5 kilograms per capita, but sig-nificantly lower than per capita egg demand of8.5 kilograms per capita in China in 2020.

But a more extreme change in the Indiandietary pattern than foreseen in the baselinescenario is possible (Bhalla, Hazell, and Kerr1999). Meat demand in India (and South Asiagenerally) differs from that in the rest of Asia

in being lower and heavily weighted towardruminant meat, mainly because beef is widelyavailable and lower priced than other meats(Rutherford 1999).33 Nevertheless, urban con-sumers in India do consume significantly moremeat than rural consumers, and urbanizationis expected to rise over the projections period(Abdulai 1999).34 In addition, Bhalla, Hazell,and Kerr (1999) report that income elasticitiesfor meat and egg products in India haveincreased over time in rural areas. It is possiblethat as Indian consumers become richer, bet-ter integrated into the world economy, andmore aware of the dietary options available tothem, they will abandon their traditional veg-etarianism and shift demand preferencestoward animal protein. Bhalla, Hazell, andKerr (1999) report some evidence that this shifthas occurred already, with 44 percent of urbanand 32 percent of rural households consum-ing meat products in 1987/88 and 50 percentof both doing so in 1993/94. Similarly, averagebudget shares for meat increased by 19 percent


TABLE 6.21 Livestock production and demand assumptions underlying the

high India meat demand and baseline scenarios, including comparison with

historical growth in China

Production growth

High India

Income elasticity meat

High India Baseline demand China

Commodity Baseline meat demand (1997–2020) (1997–2020) (1967–97)

(percent/year)

Beef 0.63 1.25 2.9 6.6 11.1

Pork 0.58 1.50 2.9 6.9 6.1

Sheep and goat 0.58 1.50 3.0 7.2 8.2

Poultry 0.96 1.50 4.8 8.3 8.6

Milk 0.58 1.00 3.2 5.5 5.7

Eggs 0.55 1.50 3.4 7.3 8.4

––––Source: FAO (2000a) for China and expert estimates for all other data.

in rural areas between 1987/88 and 1993/94,despite a minimal increase in incomes and ris-ing relative prices for these goods, comparedwith the consumer price index (National Sam-ple Survey, as reported in Bhalla, Hazell, andKerr 1999).

Under the high India meat demand sce-nario, the income elasticities of Indian meatdemand, feed ratios, and rates of growth inlivestock production are increased (Tables 6.21and 6.22). The income elasticities of demandfor pork, sheep and goats, poultry, and eggs areall increased to 1.50, the income elasticity ofdemand for beef is increased to 1.25, and thatfor milk is increased to 1.00, with all kept con-stant for the entire projections period. Rates ofgrowth in livestock numbers and feed ratios forwheat, maize, and other grains are also raised,reflecting demand-induced increases in live-stock production and the rapid shift from back-yard to commercial production that wouldhave to accompany rising livestock production.The rate of growth of total meat productionof 3.1 percent under the baseline scenario

increases to 6.4 percent under the high Indiameat demand scenario during 1997–2020. Suchproduction growth rates are much higher thanthe annual projected rates of growth of 2.6percent in East Asia and 3.0 percent in South-east Asia, but comparable to annual produc-tion growth rates in Asia of 5.1 percent in tem-perate and highland zones and 7.2 percent inthe humid and subhumid tropics between 1985and 1997 (Hoffman 1999). China set an histor-ical precedent of 7.4 percent annual meat pro-duction growth between 1967 and 1997.

A massive increase in Indian meat demandand production under the high meat demandscenario would have a major impact on Indiantrade balances for meats and cereals. Worldmeat and cereal prices would also be affectedto varying degrees. Total meat demand in Indiais projected to rise to 23 million tons, or 18.0kilograms per capita (Table 6.23). This level ofper capita meat demand is comparable to thePakistani projected per capita demand of 18.2kilograms per capita and the Indonesian of 16.0kilograms per capita in 2020, but it is signifi-cantly higher than per capita demand in otherSouth Asian countries such as Bangladesh(Table 6.24). We assume that Indian meat pro-duction will have to rise sufficiently to covermost massive quantities of meat products.Therefore we estimate net meat imports of 1.8million tons as necessary to cover excessdemand. Egg demand under the high Indiameat demand scenario will increase to a pro-jected 10 million tons, representing per capitaconsumption of 7.6 kilograms. Such a level ofper capita egg consumption would place India’sprojected per capita egg consumption on a parwith WANA’s (7.0 kilograms) and SoutheastAsia’s (6.7 kilograms).

Higher meat production, through its effectson feed demand, will place a significant strainon Indian agriculture. With meat productionincreasing to 20.9 million tons, Indian cereal

132 CHAPTER 6

TABLE 6.22 Ratio of cereal feed demand to

livestock production (not including milk)

under the high India meat demand scenario,

1997 and 2020

High India meat

Region/Country 1997 demand, 2020

China 2.1 2.2


India 0.3 1.2

Indonesia 0.7 0.5


Pakistan 0.4 0.4

Philippines 2.2 2.2

Thailand 2.7 2.4


demand will increase from 181 million tons in1997 to 281 million tons in 2020, with 25 mil-lion tons of 2020 demand accounted for byfeed demand. The baseline scenario projectedtotal Indian cereal demand in 2020 at only 260million tons, with 4 million tons accounted forby feed demand. Cereal production shortfallsunder the high India meat demand scenariowill necessitate cereal imports of 26 milliontons ($3.1 billion in value) in 2020 (Table 6.25).Meal demand will also increase significantlyover the projections period, rising from 11 mil-lion tons in 1997 to 58 million tons in 2020,while production will only increase to 27 mil-lion tons, thus necessitating imports of 31.4


TABLE 6.23 Total and per capita meat demand in India

under high India meat demand and baseline scenarios,

1997 and 2020

1997 2020

Base High India

Commodity year Baseline meat demand


Meat demand

Beef 2.6 5.4 12.5

Pork 0.5 1.0 3.1

Sheep and goat 0.7 1.4 4.0

Poultry 0.5 1.5 3.2

Total meat demand 4.3 9.4 22.7

Milk 71.4 147.5 246.8

Eggs 1.6 3.4 9.6

Per capita demand (kilograms/capita)

Beef 2.7 4.3 9.9

Pork 0.5 0.8 2.4

Sheep and goat 0.7 1.1 3.2

Poultry 0.5 1.2 2.5

Milk 74.3 116.5 195.0

Eggs 1.7 2.7 7.6


TABLE 6.24 Per capita meat demand

in various countries under the high India

meat demand scenario, 1997 and 2020

Country 1997 2020

(kilograms/capita)

Bangladesh 3.3 5.0

China 42.8 70.7

India 4.5 18.0

Indonesia 9.9 16.0

Pakistan 14.7 18.2

Thailand 28.9 49.5


million tons ($5.6 billion) in 2020. Total pro-jected meal demand in 2020 under the baselinescenario is only 24 million tons.

Higher meat demand in India will affectworld meat and cereal prices, but the impactsare rather small. Table 6.26 shows meat andcereal prices under the baseline and high Indiameat demand scenarios. Meat prices will be onlyslightly higher in 2020 under the high India meatdemand scenario than under the baseline sce-nario. Under the high demand scenario, beefprices will be 1 percent higher in 2020 and poul-try prices 3 percent higher. Price changes amongthe cereals will also be small, with wheat prices5 percent higher in 2020 than under the baselinescenario, rice prices 2 percent higher, and maizeprices 6 percent higher. High Indian meatdemand will have the largest impact on inter-national meals prices, raising them 28 percentabove the baseline 2020 values.

Higher meat consumption in India will leadto higher kilocalorie consumption and lower

childhood malnutrition. Per capita kilocalorieconsumption in 2020 will rise from 2,868 kilo-calories under the baseline to 3,060 kilocalo-ries under the high India meat demandscenario, and the projected number of mal-nourished children under the age of five in2020 will decline from 44 million childrenunder the baseline to 42 million children underthe high India meat demand scenario. This pos-itive shift, however, will be highly dependenton the ability of Indian livestock producers andfarmers to meet the production targets thatour scenario assumes. Indian livestock willhave to undergo a significant commercialtransformation if it is to satisfy a significantproportion of projected demand under thehigh India meat demand scenario. Indiaalready faces significant land and water short-ages, rendering the feasibility of the projected6.2 percent growth in livestock numbers highlyquestionable. While higher meat importscould presumably take pressure off the Indian

134 CHAPTER 6

TABLE 6.25 Net Indian cereal and meal trade under

high India meat demand and baseline scenarios, 1997

and 2020

1997 2020

Base High India

Commodity year Baseline meat demand


Wheat −0.8 −6.7 −15.7

Maize 0.0 0.2 −5.5

Other coarse

grains 0.0 0.5 −4.7

Rice 2.6 −0.3 −0.3

All cereals 1.8 −6.4 −26.2

Meals 4.2 1.8 −31.4

––––Source: IMPACT projections, June 2001.Note: Positive figures indicate net exports; negative figuresindicate net imports.

livestock sector to meet domestic demand,India’s projected agricultural trade deficitunder current assumptions of high feedimports for domestic production alreadyreaches $17 billion in 2020, representing 8.9percent of total agricultural production.35

Should domestic Indian production prove

unable to supply a rapidly expanding domes-tic market, Indian consumers would turn tothe international market, resulting in a furtherincrease in livestock prices, thus precipitatinga move back down the demand curve to lowerlevels of meat consumption. We have not mod-eled this scenario, however.


TABLE 6.26 Prices under the baseline and high India meat demand

scenarios, 1997 and 2020

1997 2020

Base High India Price change

Commodity year Baseline meat demand over baseline


Beef 1,808 1,740 1,772 2

Pork 2,304 2,239 2,289 2

Sheep and goat 2,918 2,832 3,043 7

Poultry 735 703 723 3

Milk 318 289 294 2Eggs 1,231 1,191 1,262 6

Wheat 133 123 129 5

Maize 103 102 108 6

Other coarse grains 97 86 93 8

Rice 285 250 254 2

Meals 199 191 244 28


IMPACT projections of agricultural productionand malnourishment depend on assumptionsregarding investment expenditures on a varietyof sectors generally considered important foragricultural development. The most importantinvestment drivers in the IMPACT model areirrigation, rural roads, education, clean waterprovision, and agricultural research. In thischapter we attempt to place dollar values on theassumptions embedded in the various scenar-ios: the baseline, the global pessimistic and opti-mistic scenarios, and the regional pessimisticand optimistic scenarios for Sub-Saharan Africa.

REQUIREMENTS FOR

THE BASELINE SCENARIO

For the baseline scenario, the cost of improve-ments in the five sectors during 1997–2020 areevaluated and aggregated into total investmentcosts for each region of the developing world.The dollar values thus represent what will haveto be spent in each of the five sectors between1997 and 2020 if the baseline projections are tobecome a reality. Table 7.1 summarizes invest-ments across the five sectors.

Any comprehensive investment assessmenton a global basis requires synthesis of sparsedata and many simplifying assumptions. Whenreliable, disaggregated country-level data arenot available, we have chosen to employ globalaverages as a second-best approximation of realcosts. The analysis presented here should thusbe viewed as a guide to the level of investmenteffort required to generate baseline outcomeson a regional and large-country basis, and thefollowing figures are intended as ballpark esti-mates of total costs and a basis for regionalcomparisons. Additional analysis will berequired to synthesize and prioritize invest-ment decisions in individual countries.

Total Investments

As the estimates in this chapter make clear, thedeveloping nations and international donorsface major challenges in order to achieve eventhe modest levels of agricultural productiongrowth and human welfare improvement envi-sioned in the IMPACT baseline scenario. Pro-jected expenditures on the five categories ofinvestment total $578.9 billion between 1997and 2020. This translates on an annual basis to

7

INVESTMENT REQUIREMENTS:

WHAT WILL THE COSTS BE?

136

3.6 percent of 1997 developing-world govern-ment spending of $706.1 billion (World Bank2000b). Projected spending by South Asia ishighest at $148.2 billion, followed by LatinAmerica at $140.4 billion. This represents 11percent per year of total government expendi-tures of $58.1 billion for South Asia, and 2.2percent of $272.7 billion in government expen-ditures for Latin America. Sub-Saharan Africawill face investment requirements totaling$106.9 billion, representing a sizeable 19 per-cent a year of government expenditures of$25.0 billion in 1997. China, despite doomsay-ers’ frequent warnings of impending agricul-tural disaster, will only require investmentstotaling $41.4 billion to achieve baseline results.

Different regions face different challenges.Overall, irrigation will represent 30 percent oftotal investment in the five sectors, with agri-cultural research and rural roads accountingfor 21 percent each. Education receives thelowest investment at 13 percent of the total.While irrigation expansion—particularly damconstruction—has been much maligned inrecent years for its possibly adverse impacts onthe environment and local populations, it isprojected that investment in irrigation willremain a large item, especially in India andLatin America; India mainly because of thehigh rate of increase in irrigated area, andLatin America because costs of irrigationdevelopment are relatively high there.

Rural road construction will be particularlyimportant in Sub-Saharan Africa and LatinAmerica, where it will represent 35 and 26 per-cent of total investment in the five investmentsectors, respectively, because new roads areneeded to support relatively rapid area andcrop yield growth in those regions. Despitehigh levels of investment, Sub-Saharan Africawill still have an extremely underdevelopedtransportation system, and further improve-ments in this sector are essential.

Because of rapid population growth, edu-cation will represent 27 percent of totalinvestment expenditures in WANA. Cleanwater, representing 15 percent of total invest-ment expenditures across the developingworld, is projected to account for 35 percentof China’s expenditures in 1997–2020. Agri-cultural research will account for particularlylarge shares of China’s and WANA’s totalinvestment expenditures, at 35 percent and 31percent, respectively. Public agriculturalresearch will account for only 7 percent of Sub-Saharan Africa’s total investment expendituresbetween 1997 and 2020.

Irrigation

Methodology. Total irrigation investments arecalculated by multiplying the estimated areaincreases in irrigation in 1997–2020 projectedin the IMPACT baseline by the cost of irriga-tion per hectare ( Jones 1995; Rosegrant andSvendsen 1993; Rosegrant et al. 1997). Totalirrigated area data are adjusted for croppingintensity (FAO 2000a), because the IMPACTmodel data include multiple cropping seasons.Investment costs per hectare are on a 1997 realbasis, with data from other years adjusted to1997 real dollars, utilizing the manufacturingunit value (MUV) index (World Bank 1998).

Projected Investment. Irrigation investments indeveloping countries between 1997 and 2020are projected to total $175 billion under thebaseline scenario (Table 7.1). South Asia willaccount for a significant share of this invest-ment, accounting for 35 percent of total irri-gation investment in the developing world;India alone will account for 24 percent of thisshare. Latin America’s share is 26 percent, Sub-Saharan Africa’s 16 percent, and WANA andEast Asia’s approximately 10 percent each.India has the lowest estimated costs per hectareof irrigation development in the developing

INVESTMENT REQUIREMENTS: WHAT WILL THE COSTS BE? 137

world at $3,074, indicating the extent to whichlarge projected total expenditures can be attrib-uted to massive increases in irrigated area.India is projected to expand physical irrigatedarea by 13.8 million hectares during 1997–2020,almost twice the projected expansion of 7.0million hectares in Latin America over thisperiod, despite the roughly equal projectedtotal costs in these two regions. China, whichhad the largest irrigated area in the world in1997, will only spend $3.2 billion on 0.5 millionhectares of irrigation expansion between 1997and 2020, accounting for 1.8 percent of thedeveloping world total. India is projected tosurpass China as the most heavily irrigatedcountry in the world by 2020, with 63.3 mil-lion hectares under irrigation, compared withChina’s 57.6 million hectares.

Rural Roads

Methodology. Rural road investments are cal-culated by multiplying the incremental roadlength required in 1997–2020 by road invest-

ment costs per kilometer, with the latter costsestimated by Spencer (1994). Unfortunately, alack of country-level data on this variablerequires application of the world average of$50,000 per kilometer to all regions. Rural roadlength data are available for 1997 (InternationalRoad Federation 1999). The increase in ruralroad length required under the baseline sce-nario is composed of two components. First,the road length is assumed to increase at therate necessary to maintain the same density ofroads for new cropland as exists for presentcropland. Second, the additional road lengthrequired to generate the crop yield increaseattributable to road investment is estimatedand added to the road length required for crop-land expansion. The contribution of roadinvestment to yield growth is estimated basedon the projected proportion of future yieldgrowth attributable to road expansion and theelasticity of productivity growth with respectto road investment (the latter is based on Fan,Hazell, and Thorat 1999).

138 CHAPTER 7

TABLE 7.1 Total projected investments, baseline scenario, 1997–2020

National

Rural Clean agricutural Total

Region/Country Irrigation roads Education water research a investments

(US$ billion)

Latin America 44.8 36.7 12.1 9.8 37.0 140.4



South Asia 61.3 27.4 14.5 27.0 18.0 148.2

India 42.5 23.5 10.5 18.4 15.6 110.5

Southeast Asia 18.6 3.9 6.8 9.4 14.1 52.6

China 3.2 6.8 2.4 14.4 14.6 41.4


––––Source: IMPACT projections, June 2001.aIn addition, international agricultural research expenditures by the CGIAR centers are projected to beUS$9.7 billion.

Projected Investment. Cumulative rural roadinvestments in the developing world are pro-jected to total $120.3 billion between 1997 and2020 (Table 7.1). Sub-Saharan Africa and LatinAmerica together account for more than 60percent of total expenditures in developingcountries over the projections period. Thesetwo regions account for most of the projectedexpansion in crop area harvested, as well as rel-atively rapid crop yield growth, with significantrequirements for investments in roads under-lying both of these developments. India alonewould have to invest $23.5 billion of SouthAsia’s $27.4 billion investment in rural roads.Southeast Asia, China, and WANA are pro-jected to have much lower road investments.

Public Agricultural Research

Methodology. Both national research expendi-tures and international expenditures of theCGIAR are estimated and summed to producetotal public agricultural research expenditures.Data on annual agricultural research expendi-tures are not available for the year 1997 (Tabor,Janssen, and Bruneau 1998). Therefore, themost recent data available (for the period1991–96, depending on the country), togetherwith growth rates in agricultural researchexpenditures for recent periods, are used tointerpolate forward to estimate 1997 real agri-cultural research expenditures. Because theoriginal national level data were standardizedon a 1993 purchasing power parity basis, theywere inflated with the MUV index (WorldBank 1998) to obtain 1997 values. The datawere then converted from purchasing powerparity basis to market prices (World Bank2000b) to maintain comparability with theother investment data. In order to projectannual expenditures on agricultural researchto 2020, future rates of growth in agriculturalresearch expenditures are then estimated basedon recent past trends in research expenditures

and crop yields, the projected contribution ofresearch to future crop yield growth, and elas-ticities of productivity with respect to agricul-tural research expenditures. All expendituresbetween 1997 and 2020 are subsequentlysummed to obtain total agricultural researchexpenditures over the projections period.

Projected Investment. Public national agricul-tural research expenditures in the developingworld are projected to total $121.7 billion dur-ing 1997–2020 (Table 7.1). Expenditures inLatin America will total $37.0 billion, repre-senting 30 percent of total expenditures in thedeveloping world and far outstripping the sec-ond highest region, WANA. The large expen-ditures in Latin America are mainly due toremarkably high agricultural research expen-ditures in Brazil, which are estimated toaccount for $22.7 billion. Sub-Saharan Africais expected to invest only $8.0 billion inresearch between 1997 and 2020, signalingcontinued underinvestment in agriculturalresearch in that region. Moderate agriculturalresearch expenditures are projected in India($15.6 billion) and China ($14.6 billion).

In addition to national agricultural research,the international research conducted by theCGIAR system has played a key role in cropproductivity growth (Evenson, Pray, andRosegrant 1999; Pardey et al. 1996). Theresearch conducted in the CGIAR is notincluded in the estimates above for public agri-cultural research expenditures or total nationalinvestment expenditures, so these are esti-mated separately. Expenditures by the CGIARsystem for international agricultural researchwere $319.2 million in 1997, with an annualrate of real expenditure growth of 3.8 percentbetween 1990 and 1997. On the assumptionthat rates of funding growth for internationalagricultural research will continue to declineover the projections period, we used an esti-


mated annual growth rate of 1.9 percent toproject that expenditures for international agri-cultural research will reach $9.7 billion underthe baseline scenario in 1997–2020. Thus, com-bined international and national agriculturalresearch will total $131.4 billion over the pro-jections period.

Education

Methodology. The estimated incremental invest-ments in education are based on the cumulatedannual costs of the additional number of femalestudents required to improve the percentage offemales with access to secondary education tothe levels projected in the baseline. In order tocalculate education investment costs, it is firstnecessary to obtain estimates of the additionalnumber of students that will enroll for second-ary school education between 1997 and 2020.This value is calculated by multiplying the num-ber of females of secondary school age in 1997(UN 1998) by the percentage of females enrolledin secondary school in 1997, and subtractingthat number from the number of females of sec-ondary school age in 2020 (UN 1998), multipliedby projected enrollment in the IMPACT base-line for 2020. Assuming a constant rate ofgrowth in school enrollment, the annual incre-ment in enrollment necessary to achieve theprojected enrollment in 2020 is computed. Theannual increase in enrollment is subsequentlymultiplied by the annual cost of secondaryschool education per student (UNESCO 2000;Sayed 1996) and cumulated over the period togenerate the total incremental costs for sec-ondary school education.

Projected Investment. Under the baseline sce-nario, incremental expenditures on educatingwomen in the developing world will total $75.9billion during 1997–2000 (Table 7.1). WANA’ssecondary school education costs per stu-dent—ranging from $132 to $327—and the

need to educate 73.2 million women between1997 and 2020 will cause the region to have thehighest projected education costs, represent-ing 28 percent of the developing world total.India’s costs per student are relatively low at$74, but India will have to educate an estimated142.4 million women between 1997 and 2020.China’s projected education investment is verylow at only $2.4 billion to educate an additional32.9 million women, an indication of both therelatively good performance in female educa-tion achieved to this point and the low educa-tion costs per student in China.

Clean Water

Methodology. Incremental investment costs forclean water are based on the investmentrequired to increase the percentage of peoplehaving access to clean water to the levels pro-jected under the baseline scenario. Theseinvestment costs are calculated by multiplyingthe number of people gaining access to cleanwater between 1997 and 2020 by the invest-ment costs per person of providing cleanwater. Per capita costs of providing clean water(WSSCC 2000) and estimates of the numberof people gaining access to clean water are dis-aggregated by urban and rural components(UN 1998). The per capita unit cost of provid-ing water is estimated at $75 for urban areasand $25 for rural areas (WSSCC 2000).

Projected Investment. The baseline scenario willrequire a projected $86.5 billion investment inclean water provision, with 1.8 billion additionalpeople gaining access to clean water through-out the developing world, as the percentage ofthose in the developing world who have accessto clean water rises from 69 percent in 1997 to79 percent in 2020. South Asia’s requiredinvestment will be the highest at $27.0 billion,representing 31 percent of total expenditures inthe developing world; India will account for

140 CHAPTER 7

$18.4 billion of that investment, as an additional387 million people gain access to clean water(an increase from 81 percent of the total popu-lation to 92 percent). Sub-Saharan Africa’s pro-jected expenditures for clean water are also largeat $17.3 billion, with an additional 359 millionpeople gaining access to clean water. Never-theless, only 65 percent of the population ofSub-Saharan Africa will have access to cleanwater in 2020. China will also have significantexpenditures of $14.4 billion on clean water pro-vision, reaching an additional 303 million peo-ple, with coverage extending to 78 percent ofthe population by 2020.

REQUIREMENTS FOR OPTIMISTIC

AND PESSIMISTIC SCENARIOS

As expected, the optimistic scenario requireslarge increases in investment in the key driv-ers, while the pessimistic scenario envisionssignificant reductions in investment. In thissection, we summarize the major changes in

investment required: Table 7.2 gives projectedinvestments for the optimistic scenario, andTable 7.3 for the pessimistic. Under the opti-mistic scenario, increases in irrigated areacause projected irrigation expenditures in thedeveloping world to nearly double between1997 and 2020. South Asia’s irrigation invest-ment costs rise to $101 billion, compared withthe baseline of $61 billion. Chinese irrigationinvestments increase explosively in bothabsolute and relative terms vis-à-vis the base-line, rising from $3.2 to $37.3 billion, or from1.8 to 10.9 percent of developing world expen-ditures. Latin America’s share is 19 percent;Sub-Saharan Africa’s, 14 percent; WANA’s, 13percent; and Southeast Asia’s, 9 percent. Withno expansion in irrigation assumed under thepessimistic scenario, the projected increase inirrigation expenditures is zero.

The optimistic scenario projects total ruralroad expenditures in the developing countriesto rise 7 percent above the baseline level to$128.5 billion. China’s expenditures will rise 6


TABLE 7.2 Total projected investments under the optimistic scenario, 1997–2020

National

Rural Clean agricultural Total

Region/Country Irrigation roads Education water researcha investments

(US$ billion)

Latin America 66.4 37.9 24.3 11.1 39.5 179.1



South Asia 101.0 29.9 17.1 30.0 20.4 198.4

India 61.0 25.6 12.7 20.4 17.8 137.8

Southeast Asia 30.9 4.2 9.2 10.2 15.2 69.7

China 37.3 7.2 5.0 17.5 16.4 83.5


––––Source: IMPACT projections, June 2001.aIn addition, international agricultural research investment of the CGIAR centers is projected to beUS$10.4 billion.

percent above baseline levels, while Sub-Saharan Africa, South Asia, and Southeast Asiaare all projected to increase their expendituresby approximately 9 percent. Under the pes-simistic scenario, rural road investments in thedeveloping world are projected to decline 17percent below baseline levels. Investments inSub-Saharan Africa will fall particularly sharplyby 24 percent, from $37.9 to $28.6 billion, whileinvestments in WANA and Latin America willonly fall 11 and 9 percent from the baseline,respectively.

Total education costs rise 36 percent abovethe baseline level to $102.9 under the optimisticscenario. Increases will be particularly high inLatin America, with projected expenditures ris-ing 101 percent, and China, at 105 percent.Expenditure increases for education under theoptimistic scenario will be rather low in SouthAsia and WANA, and extremely low in Sub-Saharan Africa, where educational expendi-tures are projected to increase by only 13 per-cent. WANA’s share of the developing world’s

total education investment falls to 25 percentunder the optimistic scenario.

Total projected education investment in thedeveloping world declines 28 percent from thebaseline to $54.6 billion under the pessimisticscenario. Declines are generally symmetricalto the increases generated by the optimisticscenario, with the notable exceptions of LatinAmerica and China. In Latin America expen-ditures will decline 50 percent from the base-line to $6 billion (whereas investments doubleunder the optimistic scenario). In China,expenditures will decline to 0, compared withthe baseline of $2.4 billion because the num-ber of children of secondary school age willbe greatly reduced.

Projected expenditures on clean waterprovision in the developing world between1997 and 2020 rise 13 percent above baselinelevels to $97.5 billion under the optimistic sce-nario, as 183.6 million additional people gainaccess to clean water. The percentage of thedeveloping world population with access to

142 CHAPTER 7

TABLE 7.3 Total projected investments under the pessimistic scenario, 1997–2020

National

Rural Clean agricutural Total

Region/Country Irrigation roads Education water researcha investments

(US$ billion)

Latin America 0 33.5 6.0 8.2 27.4 75.0

West Asia/North Africa 0 6.5 16.4 7.4 21.6 52.0

Sub-Saharan Africa 0 28.6 13.6 14.9 6.8 63.9

South Asia 0 22.4 11.6 23.2 13.2 70.3

India 0 19.3 8.2 15.7 11.3 54.5

Southeast Asia 0 3.2 4.6 8.2 11.1 27.1

China 0 5.6 0 10.2 11.0 26.8

Developing world 0 100.2 54.6 72.6 95.4 322.7

––––Source: IMPACT projections, June 2001.aIn addition, international agricultural research investment of the CGIAR centers is projected to be US$7.5billion.

clean water in 2020 rises to 87 percent from 79percent in the baseline scenario. Additionalexpenditures are particularly large inChina, with an increase of 22 percent; LatinAmerica, with an increase of 14 percent; andIndia, with an increase of 11 percent. The pop-ulation with access to clean water in 2020increases by 57.8 million people above thebaseline level in China (from 78 to 86 percent),18.4 million people above the baseline in LatinAmerica (from 81 to 89 percent), and 32.7 mil-lion people above the baseline in India (from92 percent to full coverage).

Expenditure shifts for clean water provisionunder the pessimistic scenario are approxi-mately 3 to 4 percent lower than the corre-sponding increases under the optimistic sce-nario, with the exception of China, whereclean water expenditures decline by 29 percentto $10.2 billion.

During 1997–2020, agricultural researchexpenditures in the developing world willrise 7.4 percent above baseline levels underthe optimistic scenario (to $130.7 billion).Increases are pronounced in India, rising 14percent, and China, with an increase of 12 per-cent. Increases in all other regions remainbelow 10 percent.

Declines under the pessimistic scenario aremore dramatic, indicating the unfortunatereality that the scope for cutting public agri-cultural research budgets remains greaterthan the scope for increasing them. Overallexpenditures on public agricultural researchbetween 1997 and 2020 are projected to decline22 percent from baseline levels under the pes-simistic scenario. Declines are again particu-larly pronounced in Asia, with expenditurein India falling 28 percent and expenditures inChina falling 25 percent. Declines are also quitelarge in Latin America, with expendituresfalling 26 percent.

In terms of international agriculturalresearch, CGIAR expenditures during 1997–2020 are projected to increase 7 percent underthe optimistic scenario ($10.4 billion) anddecline 22 percent under the pessimistic sce-nario ($7.5 billion) (CGIAR 1999).

Total investment expenditures for the opti-mistic and pessimistic scenarios show both thepromising future that could result from higherinvestment in agricultural production andhuman health as well as the depressing realityof a more likely future of serious under-investment in these sectors. Total investmentin the optimistic scenario increases to $802.4billion, 39 percent higher than under the base-line scenario. However, $168.1 billion of the$223.5 billion investment increase under theoptimistic scenario is devoted to expansion ofirrigated area. The other four categories ofinvestment go up by only $55.4 billion, anincrease of 14 percent over the cumulativelevel of investment for these expendituresunder the baseline. Total investment under thepessimistic scenario declines to $322.7 billion,a decrease of 44 percent from the baselinelevel. Irrigation again accounts for the largestshift, falling by $174.6 billion, while otherinvestments combined fall by $81.8 billion, or20 percent.

REGIONAL INVESTMENT REQUIREMENTS

The Pessimistic Scenario for Africa

The development challenges facing Sub-Saharan Africa are clearly the most dauntingof those facing any region in the world. Con-certed investment in agricultural and humandevelopment will be necessary to overcomethese challenges over the next two decades,but pervasive poverty, mismanagement, andcorruption in the region have limited thescope for internally generated investment,


while donor fatigue and the perceived failuresof much of past multilateral lending and bilat-eral aid may limit the availability of externalfunds. The baseline scenario requires totalregional investment expenditures for Sub-Saharan Africa of about $107 billion duringthe period 1997–2020, which on an annualbasis represents a sizeable 19 percent of total

government consumption in the region in1997 (Table 7.4). Rural roads will account forprojected spending of $38 billion, irrigationfor $28 billion, provision of access to cleanwater for $17 billion, education for $16 billion,and agricultural research for $8 billion. South-ern Sub-Saharan Africa is projected to have thelargest expenditures in the region at $27 bil-

144 CHAPTER 7

TABLE 7.4 Total projected investment in Sub-Saharan Africa under the baseline,

pessimistic, and optimistic scenarios, 1997–2020

Central and All

Investment scenario Nigeria Northern Western Southern Eastern Africa

(US$ billion)

Irrigation

Baseline 7.4 13.2 0.3 6.1 1.2 28.1

Optimistic 18.3 21.2 1.8 11.6 2.1 54.4

Pessimistic 0 0 0 0 0 0

Rural roads

Baseline 4.6 3.1 16.2 9.1 4.9 37.9

Optimistic 6.7 4.6 25.2 12.8 7.0 56.3

Pessimistic 2.7 1.8 9.0 5.4 2.9 21.8

Agricultural research

Baseline 0.8 0.7 1.5 2.2 2.7 8.0

Optimistic 1.4 1.3 2.9 3.7 4.7 14.1

Pessimistic 0.6 0.5 1.1 1.6 2.0 5.8

Education

Baseline 3.4 1.7 1.7 6.5 2.5 15.7

Optimistic 6.4 6.0 4.5 12.7 7.5 37.2

Pessimistic 2.7 1.4 1.3 5.4 2.1 12.9

Access to clean water

Baseline 3.1 4.1 4.7 2.6 2.8 17.3

Optimistic 3.8 5.1 5.6 3.0 3.2 20.7

Pessimistic 2.8 3.6 4.2 2.3 2.5 15.3

Total Investments

Baseline 19.2 22.7 24.3 26.5 14.2 106.9

Optimistic 36.6 38.7 39.4 44.0 24.5 182.8

Pessimistic 8.8 7.9 15.6 14.7 9.4 55.8


lion, followed by Central and Western Sub-Saharan Africa at $24 billion and NorthernSub-Saharan Africa at $23 billion. Eastern Sub-Saharan Africa will have the lowest projectedinvestment expenditures under the baseline at$14 billion.

Investment expenditures are sharply lowerunder the pessimistic Sub-Saharan Africa sce-nario, declining 48 percent to $56 billion,which on an annual basis represents 10 percentof total government consumption in theregion in 1997. Rural road investment declines42 percent from the baseline level; irrigationinvestment, 100 percent to zero; investmentin clean water provision, 12 percent; edu-cation investment, 18 percent; and agriculturalresearch investment, 28 percent. Investmentexpenditures decline in all regions under thepessimistic scenario, with Central and West-ern Africa investing the most at $16 billion, fol-lowed by Southern Africa at $15 billion andEastern Africa at $9 billion. Northern Sub-Saharan Africa’s investment decline will be par-ticularly precipitous under the pessimistic sce-nario, with total expenditures falling 65 percentto $8 billion, while Nigeria’s expenditures fall54 percent to $9 billion.

The Optimistic Scenario for Africa

Unfortunately the obstacles to achieving theoptimistic Africa vision are in many ways moredaunting than those facing Asia in the 1950s.The optimistic Africa scenario requires a 71percent increase in projected investment acrossthe region between 1997 and 2020 to reach

$183 billion, which on an annual basis repre-sents 32 percent of total government con-sumption in the region in 1997 (Table 7.4).Under the optimistic Africa scenario, ruralroad investment jumps 49 percent above thebaseline level, irrigation investment jumps 94percent, investment in provision of clean waterrises 20 percent, and agricultural researchinvestment increases 77 percent. SouthernSub-Saharan Africa will have the highest invest-ment expenditures in the region, followed byCentral and Western Sub-Saharan Africa.

These results show that far higher invest-ments will be necessary under the optimisticAfrica scenario to secure significant improve-ments in agricultural productivity and humanwell-being. The optimistic Africa scenario rep-resents a self-reinforcing, virtuous cycle ofhigher growth and higher investment, whichAfrica has been unable to initiate up to thispoint. The destructive cycle of declining invest-ment and stagnating GDP growth outlined inthe pessimistic Africa scenario, with per capitaGDP actually declining due to rapid popula-tion growth, may turn out to be just as self-reinforcing as the virtuous cycle outlined in theoptimistic Africa scenario. The poor prospectsfor Sub-Saharan Africa under the baseline sce-nario show that more aggressive policy actionson multiple fronts are needed to improve foodsecurity prospects in this region, and the opti-mistic scenario for Africa confirms that mas-sive increases in investment will be necessaryto achieve serious inroads against child mal-nutrition.


LESSONS FROM THE BASELINE SCENARIO

Fundamental changes are occurring in theglobal structure of food demand, driven inlarge part by economic growth, rising incomes,and rapid urbanization in the developingeconomies. As population growth rates indeveloping countries decline over the next 20years, rising incomes and rapid urbanization,particularly in Asia, will change the composi-tion of demand.36 World cereal markets willstrengthen from historical levels under theIMPACT baseline, with real internationalcereal prices declining only slowly. The strongerprice picture results in significant part from thecontinued gradual slowing of the rate of cerealproduction growth: with the exception ofLatin America and Sub-Saharan Africa, cerealarea will grow but little. Cereal yield growthwill thus account for the preponderance of pro-duction growth. However, in most countriesand regions—Sub-Saharan Africa being thenotable exception—the gradual slowdown ingrowth of crop yields that began in much ofthe world during the early 1980s will persist.Livestock production will grow considerably

faster than crop production but will also slowrelative to recent production growth.

Direct per capita consumption of maize andcoarse grains for food will decline as risingincomes lead consumers to shift to wheat andrice. As incomes rise further and lifestyleschange with urbanization, a continued grad-ual shift in demand from rice to wheat will con-tinue. Increasing incomes in developing coun-tries will also drive strong growth in per capitaand total meat consumption, which will in turninduce strong growth in consumption of cere-als for feed, particularly maize. In contrast, percapita meat consumption in developed coun-tries will remain nearly constant, with smalldeclines in beef and pork consumption offsetby growth in poultry consumption. Thesetrends will lead to an increase in the impor-tance of developing countries in global foodmarkets. Thus, even with continued slowing ofpopulation growth rates, urbanization and ris-ing incomes in the developing world will drivefood demand growth higher than productiongrowth in most regions over the next twodecades. Greater regional disparities betweensupply and demand will result in more agri-

8

SUMMARY AND CONCLUSIONS

146

cultural trade, with developing countries play-ing an increasingly significant role in globalfood markets. East and South Asia will fuel theboom in cereal import demand, althoughother regions will be important growth cen-ters for imports of certain cereal crops (WANAfor wheat, for example). On the export side,the United States will benefit more from futureexport opportunities than the EU15. Indeed,China, South Asia, and WANA will becomemuch more dependent on cereal imports, withSouth Asia shifting to a cereal trade deficit in2020 from near trade balance in 1997. Meatdemand in China will also help drive increasedworld meat trade, with the United States,EU15, and Latin America the primaryexporters of meat products. Several individualcommodities are the force behind higher agri-cultural trade volumes: wheat will lead theupward trend in cereal trade and poultry willdo the same for livestock trade.

Concurrently with overall expanding tradevolumes, projected trade patterns will enlargeexisting trade surpluses and deficits at theregional level. On the surplus side, the exportsof the United States, Latin America, andSoutheast Asia will account for the highest netvalues of agricultural exports. More dramaticincreases are in store for the net agriculturalimporters, however. WANA will continue topost the highest import bill as a percentage ofthe total value of agricultural production (at34 percent in 2020), although China will easilymove past WANA as the primary importer ofagricultural goods in absolute terms. Never-theless, China’s imports will remain a modest6 percent of the total value of agricultural pro-duction because of the tremendous size of theChinese agricultural sector. Overall, China’simport bill will increase more than threefold,and WANA’s will rise slightly more than 50 per-cent. Sub-Saharan Africa’s import bill willgrow to 9 percent of agricultural production

in value terms by 2020. In general, althoughtrade volumes and values will be higher,regional differences—in dietary patterns,income, population growth, and productionpossibilities—will persist.

Despite declining real food prices andexpanding world trade, food security for thepoor will only improve slowly in many regions.Sub-Saharan Africa, for example, will experi-ence little improvement in per capita calorieavailability, and the region’s number of mal-nourished children will increase by 9 millionchildren. Thus, even with relatively abundantfood in the world, there will not be enoughgrowth in effective per capita demand for foodin Sub-Saharan Africa to improve its food secu-rity situation. More progress can be seen forSouth Asia, home to more than one-half of theworld’s malnourished children, but some 64million children will still be malnourished inthe region in 2020, down from 85 million in1997. Slowly declining world food prices andbuoyant international trade will coexist withcontinuing—and even rising—malnutritionthroughout much of the world.

LESSONS FROM ALTERNATIVE SCENARIOS

The IMPACT baseline scenario presents foodsecurity outcomes based on our best estimatesof a large number of underlying drivers ofworld food markets. Complex interactionsamong technology, policy, investments, envi-ronment, and human behavior in turn influ-ence them. The alternative scenarios presentedin this volume show that plausible changes inthese drivers, and the influences underlyingthem, can have dramatic effects on food sup-ply, demand, prices, trade, and malnutrition,although the good news is that internationalagricultural markets have proven quite suc-cessful in dampening the effects of even sig-nificant shocks to the system.

SUMMARY AND CONCLUSIONS 147

The global scenarios show that no singlephenomenon at the global level can, in and ofitself, determine the world food picture in2020. Much public attention over the last sev-eral decades has focused on population growthin the developing world as the “fundamental”global problem in need of immediate atten-tion, and indeed the low population growthscenario shows that slower population growthover the next two decades would indeed havesubstantial, beneficial effects on levels of childmalnutrition in the developing world. Never-theless, while slower population growth willreduce overall food demand and alleviate stresson fragile production systems, it cannot fix thedeeply rooted structural and technologicalchallenges that confront the poor—the ruralpoor in particular—even in the present day.

The future of trade liberalization has beenanother area of recent public focus, with dis-agreements over globalization stalling the mil-lennium round of trade negotiations. Thetrade liberalization scenario shows small-to-moderate increases in world prices, comparedwith the baseline in 2020. Changes in supplyand demand in most of the developing regionsare small, with the exception of Sub-SaharanAfrica. More important, trade liberalizationwould generate significant net economic ben-efits for developing countries. Taking intoaccount the benefits to producers and con-sumers and the tax savings due to removals ofsubsidies, liberalization of trade for the 16IMPACT commodities would generate globalbenefits of $35.7 billion in 2020, including$21.5 billion for developing countries. Sub-Saharan Africa would get the highest benefitsof any region, $4.4 billion, or 10 percent of thevalue of domestic production of the IMPACTcommodities.

Perhaps the most surprising results arethose for the low feed ratio scenario, whichindicate that improvements in feeding effi-

ciency could have a substantial impact on childmalnutrition by lowering international cerealprices, particularly maize and other grainprices, and raising food consumption of thesecrops in the developing world. Sub-SaharanAfrica benefits greatly under this scenario, withthe number of malnourished children in theregion falling by 1.6 million children—animprovement of 4 percent from the baselinenumber of 39.3 million children. As theseresults show, the integration of world agricul-tural markets means that broad trends at theglobal level can have important secondaryeffects, both for good and ill, on individualregions that are relatively unaffected by theoriginal trends themselves.

The wide price swings associated with thealternative yield scenarios show that yieldgrowth will be a key determinant in ensuringthat food is available at affordable prices to theworld’s poor over the next several decades. Theyield scenarios assume that the downsidepotential is greatest in the developed world andthe upside potential is greatest in the develop-ing world. However, the developing world ischaracterized by both greater uncertaintyabout ultimate productivity gains and a greaterstake in outperforming expectations, particu-larly in high-poverty regions such as South Asiaand Sub-Saharan Africa. Should rising pro-ductivity in the more advanced developing anddeveloped countries leave less-developedregions behind, international cereal prices willfall and poor net cereal consumers will bene-fit, but the net cereal producers who comprisemuch of the rural population in poor countrieswill see their incomes decline. Greater atten-tion to the needs of producers in difficult agro-climatic and low-potential production systemsis urgently needed.

The regional Asian scenarios provide insightinto the ability of world markets, particularlygrain markets, to respond flexibly to local pro-

148 CHAPTER 8

duction shocks. For India and China both, theslow-growth scenarios as well as the high Asianfeed ratio scenario have significant impacts onvarious measures of local agricultural produc-tion, demand, trade, and food security but onlymodest effects on international prices. Thisresult points to the need to react sensibly to theprospect that both India and China couldbecome much larger net cereal-importingcountries over the coming decades than is pro-jected in the baseline. Far more important thanthe simple fact that these countries are placingsignificant demands on world markets shouldbe the underlying reasons behind such a devel-opment. Clearly, production shortfalls due todisappointing yield performance and wide-spread degradation would be cause for concernin the two Asian giants. But rising importdependence can be a positive development if itresults from a reduction in trade distortions anda realignment of domestic incentives towardsectors in which these countries possess a com-parative advantage. If India and China welcomethe benefits of supplying their domestic foodneeds through global markets, the rest of theworld should certainly be able to supply theseneeds without major dislocations.

Intense concern over the future prospectsfor Sub-Saharan Africa is a theme echoedthroughout this volume. It reflects theunavoidable expectation that worsening foodinsecurity in the region will continue over thecoming decades. Many experts see the pes-simistic scenario as the most likely outcomefor Sub-Saharan Africa. Deteriorating naturalresources, stagnant technologies, and risingpopulation densities remain common featuresof the rural landscape throughout much ofSub-Saharan Africa. These problems will onlybe alleviated with the advent of a major struc-tural transformation from subsistence agri-culture to a commercialized and highly pro-ductive agricultural economy, capable of

supporting a large urban population. The opti-mistic scenario—comprising an admittedlyhighly stylized set of hypothetical assumptionsthat lead to rather modest declines in childmalnutrition—invokes a rate of productivitygrowth that in itself would represent a remark-able transformation. Nevertheless, while theability of African production systems to pro-duce the hypothesized levels of agriculturalproduction in the time frame of the optimisticscenario remains questionable, surely financialand political commitment to this goal shouldremain at the forefront of the developmentagenda. In light of the ominous threats thatconfront the health and well-being of its chil-dren, Africa cannot afford to fall furtherbehind, but a sense of urgency and commit-ment to the challenge of rural developmenthas been lacking.

The broader global optimistic and pes-simistic scenarios show that better policy andmore rapid economic and agricultural growthcan lead to substantial food security improve-ments, but significantly worse outcomes arealso possible if key drivers perform even slightlyworse. As investment calculations show, foodsecurity is vulnerable to relatively smalldeclines in policy efforts relative to the base-line. Conversely, the developing countries canachieve significant reductions in childhood mal-nutrition within the boundaries of plausiblelong-term economic performance. It is pro-jected that an increase in investment of $224billion—about 40 percent above the baselineprojection—in irrigation, agricultural research,rural roads, female education, and develop-ment of clean water would be required toreduce the number of malnourished childrenin 2020 by 38 million children (29 percent),compared with the baseline projection for 2020.A concerted effort to eliminate childhood mal-nutrition will require policy reform and greaterpublic investment, producing dramatic long-


term gains in income growth, agricultural pro-ductivity, and social indicators.

LESSONS FOR INVESTMENT

Irrigation, agricultural research, and ruralroads are the three main investments that willdrive agricultural production growth in themodel, and it is projected that together thesethree areas will require $416.6 billion of invest-ment between 1997 and 2020 to achieve base-line agricultural production growth rates. Irri-gation is projected to account for 42 percent ofinvestment in these three areas under the base-line and nearly 60 percent of productivity-enhancing investments under the optimisticscenario, not taking into account the costs ofoperating and maintaining existing systems.Given the relatively low rates of returnachieved on many recent irrigation projects( Jones 1995), the extent to which governmentsmay be overinvesting in irrigation projects atthe expense of other worthwhile productiveinvestments remains unclear. Cutbacks on irri-gation investment under the pessimistic sce-nario, however, impose significant reductionsin crop yield and area growth. Moreover, ini-tial results from a prototype model linking theIMPACT model to a global water simulationmodel support the assessment that the down-side impacts on food production of cutbacksin irrigation investment—and more broadly inwater supply, demand management, and waterinfrastructure—can be very large (Rosegrantand Cai 2000). Given the aggregate level of theglobal analysis here, more detailed analyses ofcountry-specific opportunities in irrigation andwater infrastructure are required to further pri-oritize irrigation and water investments.

Ultimately, it will be up to individual policy-makers at specific moments in time to deter-mine whether scarce public resources arebetter spent on expanding or repairing irriga-

tion infrastructure, investing in agriculturalresearch to develop locally appropriate cropvarieties, or providing farmers with betterextension services. Recent studies (Alston etal. 2000) have shown the high returns thataccrue to investments in agricultural research,underpinning a strong case that national gov-ernments are underinvesting. Additionally,national expenditures have only increased inimportance as funding for international agri-cultural research declines, and the private sec-tor increasingly serves as the engine drivingtechnological change in agriculture. The devel-oping countries will lose out on the tremen-dous benefits from emerging advances if theydo not commit enough of their own funds toadapting these technologies to their specificneeds. While private investment in agriculturalresearch in developed countries is likely togrow significantly, an important, and probablydominant, role for public investment willremain because of incentive problems that dis-courage private investment in research ori-ented to the crops and agroclimatic conditionsof developing countries.

New research on India (Fan, Hazell, andThorat 1999) and China (Fan, Zhang, andZhang 2000) shows that investments in ruralroads have large impacts on productivity andpoverty alleviation. The baseline projectionsindicate that the construction of rural roads inSub-Saharan Africa necessary to support cropyield growth and area expansion will requiremassive investments, accounting for 35 percentof total projected investments in the region.Failure to invest heavily in rural roads wouldmake it difficult for the region to move beyondsubsistence agriculture to a more commer-cialized system, with all the dynamic effectsthat agricultural commercialization could haveon rural incomes.

Together with Sub-Saharan Africa, SouthAsia faces the biggest challenges in improving

150 CHAPTER 8

food security and reducing malnutrition. SouthAsia, with the largest projected investmentexpenditures of any region, will also have tomake difficult allocation decisions, especiallygiven the projected need to invest heavily inirrigation infrastructure, rural roads, and cleanwater provision. Policymakers in Sub-SaharanAfrica and South Asia—and throughout thedeveloping world—must carefully assessinvestment options, and they will require thepolitical will and foresight necessary to makethe appropriate allocation decisions amongcompeting sectors.

Given the key roles played by female edu-cation and clean water provision in reducingchildhood malnutrition, it is essential thatdeveloping countries reinvigorate their com-mitments to these spending areas. The base-line scenario projects cumulative spending of$75.9 billion on female education and $86.5 onclean water provision. Additional investmentof $27 billion for education and $11 billion inclean water provision contributes a large shareof the reduction in childhood malnutritionrealized under the optimistic scenario. Whilethese investments have value beyond their con-tribution to reducing childhood malnutrition,it is important to note that increased food pro-duction alone will not lead to major improve-ments in food security in the absence of suchcomplementary investments. WANA and Sub-Saharan Africa in particular will require sig-nificant investments in the education ofwomen, even under the baseline scenario.

Finally, economic growth will be an essen-tial component of all investment decisions,both because it can provide the resources nec-essary to make investments, and because onlythrough economic growth can employmentopportunities expand for those leaving agri-culture and taking advantage of expandingeducational opportunities. Although the pre-cise set of policy reforms and the prioritiesand magnitudes of increases in investmentrequired to eliminate childhood malnutritionneed to be determined in detail for each coun-try, IMPACT results confirm that the threefoundations for success are broad-based eco-nomic growth, growth in agricultural pro-duction, and investment in social services,including education and health. Failure in anyof these three areas will severely hamperefforts to eliminate childhood malnutrition.

The challenges to improving food securityworldwide are not intractable. As the baselineand alternative scenarios presented heredemonstrate, considerable flexibility in worldfood markets provides a substantial bufferagainst catastrophic downside outcomes.Scope exists for national and international pol-icy to improve regional outcomes whileadding to the worldwide buffer. Taking advan-tage of the opportunities available and deliv-ering on the possibility of a more food-secureworld in 2020 will remain the primary chal-lenge for policymakers at all levels over thenext 20 years.


153

Definitions of IMPACT Regions and Countries

Developed regions/countries

Australia

European Union (EU 15): Austria, Belgium, Denmark, Finland, France, Germany, Greece,Ireland, Italy, Luxembourg, the Netherlands, Portugal, Spain, Sweden, and the UnitedKingdom

Japan

United States

Other developed countries: Canada, Iceland, Israel, Malta, New Zealand, Norway, SouthAfrica, and Switzerland

Former Soviet Union (FSU) Regions/Countries

Eastern Europe: Albania, Bosnia-Herzegovina, Bulgaria, Croatia, Czech Republic, Hungary, Macedonia, Poland, Romania, Slovakia, Slovenia, and Yugoslavia

Central Asia: Kazakhstan, Kyrgyzstan, Tajikistan, Turkmenistan, Uzbekistan

Rest of the Former Soviet Union: Armenia, Azerbaijan, Belarus, Estonia, Georgia, Latvia, Lithuania, Moldova, Russian Federation, and Ukraine

Developing regions/countries

Central and Latin America

Argentina

Brazil

Colombia

Mexico

Appendix A

COUNTRIES

AND COMMODITIES

INCLUDED IN THE IMPACT

MODEL

154 APPENDIX A

Other Latin America: Antigua and Barbuda, Bahamas, Barbados, Belize, Bolivia, Chile,Costa Rica, Cuba, Dominica, Dominican Republic, Ecuador, El Salvador, FrenchGuiana, Grenada, Guadeloupe, Guatemala, Guyana, Haiti, Honduras, Jamaica,Martinique, Netherlands Antilles, Nicaragua, Panama, Paraguay, Peru, Saint Kittsand Nevis, Saint Lucia, Saint Vincent, Suriname, Trinidad and Tobago, Uruguay, andVenezuela

Sub-Saharan Africa

Central and Western Sub-Saharan Africa: Benin, Cameroon, Central African Republic,Comoros Island, Congo Democratic Republic, Congo Republic, Gabon, Gambia,Ghana, Guinea, Guinea-Bissau, Ivory Coast, Liberia, Sao Tome and Principe, Senegal,Sierra Leone, and Togo

Eastern Sub-Saharan Africa: Burundi, Kenya, Rwanda, Tanzania, and Uganda

Nigeria

Northern Sub-Saharan Africa: Burkina Faso, Chad, Djibouti, Eritrea, Ethiopia, Mali,Mauritania, Niger, Somalia, and Sudan

Southern Sub-Saharan Africa: Angola, Botswana, Lesotho, Madagascar, Malawi,Mauritius, Mozambique, Namibia, Réunion, Swaziland, Zambia, and Zimbabwe

West Asia/North Africa (WANA)

Egypt

Turkey

Other West Asian/North African countries: Algeria, Cyprus, Iran, Iraq, Jordan, Kuwait,Lebanon, Libya, Morocco, Saudi Arabia, Syria, Tunisia, United Arab Emirates, andYemen

South Asia

Bangladesh

India

Pakistan

Other South Asian countries: Afghanistan, Maldives, Nepal, and Sri Lanka

Southeast Asia

Indonesia

Malaysia

Myanmar

Philippines

Thailand

Viet Nam

Other Southeast Asian countries: Brunei, Cambodia, and Laos

COUNTRIES AND COMMODITIES 155

East Asia

China (including Taiwan and Hong Kong)

Korea, Republic of

Other East Asian countries: Democratic People’s Republic of Korea; Macao; and Mongolia

Rest of the world: Cape Verde, Fiji, French Polynesia, Kiribati, New Caledonia, Papua NewGuinea, Seychelles, and Vanuatu

Definitions of IMPACT Commodities

Beef (beef and buffalo meat)

Pork (pig meat)

Sheep/goat (sheep and goat meat)

Poultry (chicken meat)

Eggs

Milk (raw milk containing all its constituents: not concentrated, pasteurized, sterilized orotherwise preserved, homogenized, or peptonized)

Wheat

Maize

Rice

Other coarse grains (barley, millet, oats, rye, and sorghum)

Potatoes

Sweet potatoes and yams

Cassava and other roots and tubers (cassava and other tubers, roots or rhizomes)

Soybeans

Meals (copra cake, cottonseed cake, groundnut cake, other oilseed cakes, palm kernelcake, rape and mustard seed cake, sesame seed cake, soybean cake, sunflower seedcake, fish meal, meat and blood meal)

Oils (vegetable oils and products, animal fats and products)

157

Appendix B

SUPPLEMENTARY

PRODUCTION, DEMAND,

AND TRADE DATA

TABLE B.1 Net maize and wheat trade, selected countries, 1967 and 1997

Wheat Maize

Region/Country 1967 1997 1967 1997


Asia –12.3 –21.7 0.3 –13.9

China −5.5 −4.6 −0.4 −2.0

Indonesia −0.3 −3.7 0.1 −0.5

Japan −4.0 −5.8 −4.2 −16.1

Korea, Republic of −0.7 −3.4 0.0 −8.0

Malaysia −0.3 −0.8 −0.2 −2.3

Philippines −0.6 −2.1 0.0 −0.4

Latin America –3.9 –7.8 5.6 −2.3

Brazil −2.5 −6.7 0.7 −0.6

Colombia −0.2 −1.0 0.0 −1.8

Mexico 0.1 −1.8 1.0 −4.4

Other Latin American −3.4 −6.2 −0.4 −5.0

West Asia/North Africa –5.8 –25.9 –0.4 –9.7

Egypt −2.3 −6.8 −0.2 −2.9

Other West Asia/North Africa −3.1 −18.2 −0.2 −5.8

Sub-Saharan Africa −1.2 −6.1 0.7 −1.8


Note: Positive figures indicate net exports; negative figures indicate net imports.

TA

BL

E B

.2

Gro

wth

rat

es o

f dem

and

for

vari

ou

s ce

real

s, 1

967–

97

Whe

at d

eman

dM

aize

dem

and

Oth

er c

oar

se g

rain

s de

man

dR

ice

dem

and

Reg

ion

1967–

82

1982–

90

1990–

97

1967–

82

1982–

90

1990–

97

1967–

82

1982–

90

1990–

97

19

67

–8

21

98

2–

90

19

90

–9

7

(per

cen

t/yea

r)

Latin

Am

eric

a3.

40.

82.

33.

62.

34.

07.

1−0

.71.

93.

52.

01.

7

Sub-

Saha

ran

Afr

ica

6.2

1.7

5.6

2.9

5.3

2.7

1.4

2.8

3.6

5.1

3.3

2.9

Wes

t Asi

a/N

orth

Afr

ica

4.6

3.1

1.9

6.1

4.7

3.9

2.7

4.5

0.6

4.6

3.3

4.6

Asi

a5.

83.

62.

33.

11.

7–3

.5–0

.0–2

.6–2

.33.

32.

61.

5

Sout

h A

sia

4.9

3.8

3.4

1.3

3.1

2.5

0.6

−0.8

−0.5

2.8

3.4

1.8

Sout

heas

t Asi

a6.

23.

26.

94.

46.

44.

78.

16.

54.

73.

62.

62.

1

Eas

t Asi

a6.

53.

51.

36.

14.

14.

3−1

.2−3

.30.

13.

51.

81.

0

Dev

elop

ed w

orld

1.6

1.5

−0.9

2.8

0.8

1.1

1.3

−0.2

−3.3

0.7

0.5

0.6

Dev

elop

ing

wor

ld5.

23.

12.

34.

53.

84.

01.

40.

71.

33.

42.

51.

6

––––

Sou

rce:

Bas

ed o

n FA

OST

AT

dat

a (F

AO

200

0a).

SUPPLEMENTARY PRODUCTION, DEMAND, AND TRADE DATA 159

TABLE B.3 Global meat trade, 1967 and

1997

Meat 1967 1997

(milion metric tons)

Beef 2.4 6.9

Pork 1.6 6.7

Sheep and goat 0.6 0.9

Poultry 0.4 6.4

Total 5.0 20.9


TABLE B.4 Roots and tubers demand, 1967–97

Region 1967 1982 1990 1997

(milion metric tons)

Latin America 44.2 44.7 47.0 49.5

Sub-Saharan Africa 56.2 73.5 110.6 148.8

West Asia/North Africa 3.4 8.4 12.8 15.6

Asia 145.7 188.6 197.9 242.8

South Asia 11.5 20.6 25.4 32.2

Southeast Asia 18.6 25.0 26.0 28.4

East Asia 115.6 143.0 146.5 182.2

Developing world 251.1 316.7 370.9 449.1

Developed world 258.7 226.8 219.7 195.0


TABLE B.5 Cereal yields, Sub-Saharan

Africa, 1967–97

Cereal 1967 1982 1990 1997

(kilograms/hectare)

Maize 945 1,158 1,201 1,267

Wheat 928 1,394 1,589 1,655

Other coarse

grains 633 809 712 751

Rice 877 916 1,107 1,074


TA

BL

E B

.6

In

com

e d

eman

d e

last

icit

ies,

199

7 an

d 2

020

So

uthe

ast

Sub

-Sah

aran

Lat

inD

evel

ope

d

Co

mm

odi

tyY

ear

So

uth

Asi

aE

ast

Asi

a A

sia

WA

NA

Afr

ica

Am

eric

aco

untr

ies

Bee

f19

970.

570.

820.

810.

450.

840.

380.

17

2020

0.39

0.63

0.69

0.35

0.72

0.28

0.05

Pork

1997

0.28

0.61

0.64

0.33

0.64

0.41

0.27

2020

0.12

0.52

0.46

0.16

0.56

0.35

0.09

Shee

p an

d go

at m

eat

1997

0.42

0.54

0.33

0.40

0.33

0.48

0.24

2020

0.25

0.37

0.17

0.23

0.21

0.32

0.16

Poul

try

1997

0.91

0.91

0.81

0.72

0.72

0.81

0.64

2020

0.77

0.82

0.66

0.59

0.64

0.70

0.51

Egg

s19

970.

560.

510.

510.

410.

460.

210.

03

2020

0.47

0.36

0.35

0.21

0.34

0.12

–0.0

9

Milk

1997

0.50

0.63

0.59

0.32

0.73

0.30

0.27

2020

0.40

0.46

0.42

0.20

0.61

0.20

0.12

Whe

at19

970.

180.

260.

320.

100.

500.

090.

17

2020

0.13

0.13

0.28

–0.0

10.

490.

030.

11

Ric

e19

970.

180.

050.

140.

200.

380.

220.

33

2020

0.04

–0.2

00.

010.

130.

380.

190.

22

Mai

ze19

970.

02–0

.16

0.04

–0.1

40.

300.

05–0

.05

2020

–0.0

3–0

.26

–0.0

7–0

.22

0.26

–0.0

2–0

.15

Oth

er c

oars

e gr

ains

1997

0.01

–0.0

9–0

.04

–0.1

20.

33–0

.10

–0.1

8

2020

–0.0

6–0

.15

–0.1

2–0

.24

0.26

–0.2

1–0

.24

Pota

toes

1997

0.56

0.42

0.45

0.45

0.42

0.31

0.21

2020

0.45

0.30

0.36

0.34

0.40

0.23

0.14

Swee

t pot

atoe

s an

d ya

ms

1997

–0.0

8–0

.06

0.00

0.10

0.26

–0.0

5–0

.10

2020

–0.1

8–0

.16

–0.1

00.

020.

18–0

.15

–0.2

0

Cas

sava

and

oth

er r

oots

1997

0.06

–0.0

50.

040.

080.

31–0

.06

–0.1

8

and

tube

rs20

20–0

.08

–0.1

4–0

.06

–0.0

90.

19–0

.16

–0.2

7

––––

Sou

rce:

Exp

ert e

stim

ates

.

SUPPLEMENTARY PRODUCTION, DEMAND, AND TRADE DATA 161

TABLE B.7 Crop area (or animal slaughter numbers) elasticity with respect to own-crop

price, average by region

South East Southeast Sub-Saharan Latin Developed

Commodity Asia Asia Asia WANA Africa America countries

Beef 0.27 0.35 0.26 0.26 0.29 0.39 0.34

Pork 0.23 0.36 0.45 0.13 0.25 0.42 0.35

Sheep and goat meat 0.26 0.22 0.24 0.26 0.22 0.20 0.29

Poultry 0.38 0.37 0.36 0.38 0.30 0.35 0.35

Eggs 0.28 0.38 0.35 0.37 0.28 0.28 0.23

Milk 0.34 0.37 0.34 0.40 0.38 0.32 0.31

Wheat 0.10 0.15 0.02 0.20 0.21 0.21 0.23

Rice 0.12 0.10 0.10 0.16 0.15 0.15 0.12

Maize 0.13 0.12 0.12 0.20 0.19 0.17 0.19

Other coarse grains 0.13 0.13 0.11 0.21 0.26 0.22 0.19

Potatoes 0.15 0.14 0.10 0.15 0.15 0.15 0.16

Sweet potatoes and yams 0.09 0.08 0.10 0.08 0.12 0.10 0.11

Cassava and other roots

and tubers 0.09 0.07 0.11 0.08 0.15 0.09 0.08

Soybeans 0.16 0.15 0.13 0.12 0.10 0.18 0.14

––––Source: Expert estimates.

TABLE B.8 Crop yield elasticity with respect to own-crop price, average by region


Crop Asia Asia Asia WANA Africa America countries

Wheat 0.18 0.16 0.13 0.14 0.19 0.15 0.12

Rice 0.12 0.13 0.10 0.13 0.18 0.15 0.11

Maize 0.15 0.14 0.11 0.13 0.17 0.13 0.14

Other coarse grains 0.12 0.11 0.09 0.10 0.14 0.10 0.10

Potatoes 0.14 0.12 0.10 0.12 0.16 0.12 0.11

Sweet potatoes and yams 0.11 0.09 0.08 0.08 0.14 0.09 0.09


and tubers 0.11 0.09 0.08 0.08 0.14 0.09 0.09

Soybeans 0.14 0.12 0.10 0.11 0.16 0.12 0.11


162 APPENDIX B

TABLE B.9 Crop yield elasticity with respect to wage of labor, average by region



Wheat –0.11 –0.07 –0.08 –0.07 –0.17 –0.09 –0.04

Rice –0.07 –0.05 –0.06 –0.07 –0.15 –0.09 –0.03

Maize –0.11 –0.06 –0.07 –0.06 –0.15 –0.07 –0.04

Other coarse grains –0.09 –0.05 –0.06 –0.05 –0.12 –0.06 –0.03

Potatoes –0.10 –0.06 –0.07 –0.07 –0.15 –0.08 –0.04

Sweet potatoes and yams –0.09 –0.05 –0.06 –0.06 –0.13 –0.07 –0.04


and tubers –0.09 –0.05 –0.06 –0.06 –0.13 –0.07 –0.04

Soybeans –0.09 –0.06 –0.07 –0.06 –0.14 –0.08 –0.04


TABLE B.10 Crop yield elasticity with respect to price of capital, average by region



Wheat –0.06 –0.09 –0.05 –0.07 –0.02 –0.06 –0.09

Rice –0.05 –0.07 –0.04 –0.07 –0.03 –0.06 –0.08

Maize –0.05 –0.08 –0.04 –0.07 –0.02 –0.05 –0.10

Other coarse grains –0.04 –0.06 –0.03 –0.05 –0.01 –0.04 –0.07

Potatoes –0.04 –0.06 –0.03 –0.05 –0.02 –0.04 –0.07

Sweet potatoes and yams –0.02 –0.04 –0.02 –0.03 –0.01 –0.03 –0.06


and tubers –0.02 –0.04 –0.02 –0.03 –0.01 –0.02 –0.06

Soybeans –0.04 –0.06 –0.03 –0.05 –0.02 –0.05 –0.07


TABLE B.11 Baseline IMPACT PSE estimates, livestock products

Sheep

Region/Country Beef Pork and goat Poultry Eggs Milk

Developed countries

United States 0.04 0.03 0.04 0.03 0.00 0.60

EU15 0.63 0.10 0.63 0.19 0.00 0.57

Japan 0.32 0.34 0.32 0.34 0.00 0.80

Australia 0.04 0.03 0.04 0.03 0.00 0.31

Other developed countries 0.12 0.13 0.14 0.15 0.00 0.47

Eastern Europe 0.14 0.10 0.06 0.21 0.00 0.29

Central Asia 0.23 0.22 0.21 0.24 0.00 0.37

Other Former Soviet Union 0.39 0.38 0.38 0.38 0.00 0.45

Latin America

Mexico 0.04 0.05 0.04 0.05 0.00 0.44

Brazil 0.12 0.12 0.12 0.12 0.00 0.20

Argentina 0.07 0.04 0.07 0.04 0.00 0.10

Colombia 0.08 0.06 0.06 0.06 0.00 0.20

Other Latin America 0.06 0.06 0.06 0.06 0.00 0.18

Sub-Saharan Africa

Nigeria 0.25 0.25 0.25 0.25 0.00 0.00

Northern 0.28 0.08 0.16 0.08 0.00 0.00

Central and Western 0.43 0.28 0.28 0.10 0.00 0.00

Southern 0.24 0.10 0.20 0.05 0.00 −0.29

Eastern 0.50 0.30 0.50 0.30 0.00 −0.10

West Asia/North Africa

Egypt 0.15 0.15 0.15 0.15 0.00 0.30

Turkey 0.34 0.28 0.34 0.28 0.00 0.54

Other West Asia/North Africa 0.12 0.12 0.12 0.12 0.00 0.23

Asia

India 0.29 0.29 0.29 0.29 0.00 0.26

Pakistan 0.10 0.10 0.10 0.10 0.00 0.22

Bangladesh 0.10 0.10 0.10 0.10 0.00 0.22

Other South Asia 0.05 0.05 0.05 0.05 0.00 0.19

Indonesia 0.11 0.18 0.13 0.20 0.00 0.15

Thailand 0.13 0.13 0.13 0.13 0.00 0.50

Malaysia 0.13 0.40 0.13 0.35 0.00 0.50

Philippines 0.25 0.30 0.25 0.25 0.00 0.50

Viet Nam 0.27 0.22 0.27 0.15 0.00 0.27

Myanmar 0.15 0.16 0.15 0.16 0.00 0.08

Other Southeast Asia 0.15 0.16 0.15 0.16 0.00 0.08

China 0.22 0.10 0.22 0.10 0.00 0.25

––––Source: See note 16.

TA

BL

E B

.12

B

asel

ine

IMP

AC

T P

SE

est

imat

es, c

erea

ls a

nd

ro

ots

an

d t

ub

ers

Cas

sava

Sw

eet

and

oth

er

Oth

erpo

tato

es

roo

ts a

nd

Reg

ion/

Co

untr

yW

heat

Mai

zeR

ice

grai

nsP

ota

toes

and

yam

stu

bers

So

ybea

nsM

eals

Oils

Dev

elop

ed c

ount

ries

Uni

ted

Stat

es0.

170.

060.

140.

140.

000.

000.

000.

100.

100.

10

EU

150.

460.

290.

460.

460.

000.

000.

000.

380.

480.

38

Japa

n0.

860.

840.

680.

680.

000.

000.

000.

360.

360.

36

Aus

tral

ia0.

050.

050.

000.

000.

000.

000.

000.

020.

020.

02

Oth

er d

evel

oped

cou

ntri

es0.

060.

070.

010.

010.

000.

000.

000.

100.

100.

10

Eas

tern

Eur

ope

0.02

0.00

0.05

0.00

0.00

0.00

0.00

−0.0

9−0

.09

−0.0

9

Cen

tral

Asi

a−0

.15

0.20

0.20

0.20

0.00

0.00

0.00

−0.0

4−0

.04

−0.0

4

Oth

er F

orm

er S

ovie

t Uni

on−0

.15

−0.3

00.

000.

000.

000.

000.

00−0

.04

−0.0

4−0

.04

Latin

Am

eric

a

Mex

ico

0.29

0.01

0.10

0.10

0.00

0.00

0.00

0.06

0.06

0.06

Bra

zil

0.24

0.20

0.05

0.05

0.00

0.00

0.00

0.09

0.05

0.05

Arg

entin

a0.

040.

060.

050.

050.

000.

000.

000.

010.

000.

00

Col

ombi

a0.

210.

280.

050.

050.

000.

000.

000.

11−0

.01

−0.0

1

Oth

er L

atin

Am

eric

a0.

050.

110.

050.

050.

000.

000.

000.

160.

280.

28

Sub-

Saha

ran

Afr

ica

Nig

eria

−0.3

00.

30−

0.10

−0.1

00.

000.

000.

00−0

.10

−0.1

0−0

.10

Nor

ther

n−0

.29

−0.1

7−0

.17

−0.1

70.

000.

000.

00−0

.03

−0.0

3−0

.03

Cen

tral

and

Wes

tern

−0

.20

0.41

−0.2

3−0

.23

0.00

0.00

0.00

−0.0

7−0

.07

−0.0

7

Sout

hern

−0

.25

0.12

−0.2

0−0

.20

0.00

0.00

0.00

−0.0

7−0

.07

−0.0

7

Eas

tern

−0

.23

0.30

−0.0

6−0

.06

0.00

0.00

0.00

−0.0

2−0

.02

−0.0

7

Wes

t Asi

a/N

orth

Afr

ica

Egy

pt0.

460.

120.

030.

230.

000.

000.

000.

300.

040.

04

Tur

key

0.42

0.05

0.46

0.46

0.00

0.00

0.00

0.45

0.23

0.45

Oth

er W

est A

sia/

Nor

th A

fric

a0.

070.

070.

130.

070.

000.

000.

000.

180.

180.

18

Indi

a0.

100.

180.

020.

020.

000.

000.

000.

300.

200.

20

Paki

stan

0.15

0.10

0.28

0.26

0.00

0.00

0.00

0.30

0.18

0.18

Ban

glad

esh

0.10

0.20

0.20

0.10

0.00

0.00

0.00

0.20

0.12

0.12

Oth

er S

outh

Asi

a0.

050.

100.

100.

060.

000.

000.

000.

250.

200.

20

Indo

nesi

a0.

100.

090.

040.

040.

000.

000.

000.

380.

190.

19

Tha

iland

0.60

0.10

0.10

0.10

0.00

0.00

0.00

0.24

0.10

0.10

Mal

aysi

a0.

280.

490.

250.

450.

000.

000.

000.

050.

170.

17

Phili

ppin

es0.

280.

270.

600.

320.

000.

000.

000.

290.

100.

10

Vie

t Nam

0.05

0.25

0.09

0.23

0.00

0.00

0.00

0.10

0.17

0.17

Mya

nmar

0.10

0.20

0.09

0.10

0.00

0.00

0.00

0.10

0.10

0.10

Oth

er S

outh

east

Asi

a0.

100.

200.

090.

100.

000.

000.

000.

100.

100.

10

Chi

na0.

070.

070.

080.

080.

000.

000.

000.

020.

020.

02

Rep

ublic

of K

orea

0.13

0.71

0.70

0.70

0.00

0.00

0.00

0.77

0.77

0.77

Oth

er E

ast A

sia

0.05

0.20

0.10

0.38

0.00

0.00

0.00

0.10

0.20

0.20

Res

t of t

he w

orld

0.10

0.20

0.10

0.19

0.00

0.00

0.00

0.10

0.20

0.20

––––

Sou

rce:

See

note

16.

TA

BL

E B

.13

B

asel

ine

IMP

AC

T C

SE

est

imat

es, l

ives

tock

pro

du

cts

She

ep

Reg

ion/

Co

untr

yB

eef

Po

rkan

d go

atP

oul

try

Egg

sM

ilk

Dev

elop

ed C

ount

ries

Uni

ted

Stat

es−0

.01

0.00

−0.0

10.

000.

00−0

.57

EU

15−0

.34

−0.0

8−0

.34

−0.2

10.

00−0

.53

Japa

n−0

.29

−0.3

3−0

.29

−0.3

30.

00−0

.71

Aus

tral

ia0.

000.

000.

000.

000.

00−0

.28

Oth

er d

evel

oped

cou

ntri

es−0

.09

−0.1

1−0

.05

−0.1

30.

00−0

.47

Eas

tern

Eur

ope

−0.0

5−0

.05

−0.0

1−0

.18

0.00

−0.2

6

Cen

tral

Asi

a−0

.23

−0.2

2−0

.19

−0.2

30.

00−0

.36

Res

t of F

orm

er S

ovie

t Uni

on−0

.39

−0.3

8−0

.37

−0.3

70.

00−0

.45

Latin

Am

eric

a

Mex

ico

−0.0

2−0

.04

−0.0

2−0

.04

0.00

−0.4

0

Bra

zil

−0.1

2−0

.12

−0.1

2−0

.12

0.00

−0.2

0

Arg

entin

a−0

.07

−0.0

4−0

.07

−0.0

40.

00−0

.10

Col

ombi

a−0

.08

−0.0

6−0

.06

−0.0

60.

00−0

.20

Oth

er L

atin

Am

eric

a−0

.06

−0.0

6−0

.06

−0.0

60.

00−0

.18

Sub-

Saha

ran

Afr

ica

Nig

eria

−0.2

5−0

.25

−0.2

5−0

.25

0.00

0.00

Nor

ther

n −0

.28

−0.0

8−0

.16

−0.0

80.

000.

00

Cen

tral

and

Wes

tern

−0

.43

−0.2

8−0

.28

−0.1

00.

000.

00

Wes

t Asi

a −0

.24

−0.1

0−0

.20

−0.0

50.

00−0

.09

Nor

th A

fric

a −0

.50

−0.3

0−0

.50

−0.3

00.

00−0

.15

Egy

pt−0

.15

−0.1

5−0

.15

−0.1

50.

00−0

.30

Tur

key

−0.2

5−0

.25

−0.2

5−0

.25

0.00

−0.5

0

Oth

er W

est A

sia/

Nor

th A

fric

a−0

.12

−0.1

2−0

.12

−0.1

20.

00−0

.23

Asi

a

Indi

a−0

.29

−0.2

9−0

.29

−0.2

90.

00−0

.26

Paki

stan

−0.0

5−0

.05

−0.0

5−0

.05

0.00

−0.2

2

Ban

glad

esh

−0.0

5−0

.05

−0.0

5−0

.05

0.00

−0.2

2

Oth

er S

outh

Asi

a−0

.05

−0.0

5−0

.05

−0.0

50.

00−0

.19

Indo

nesi

a−0

.11

−0.1

5−0

.15

−0.2

00.

00−0

.15

Tha

iland

−0.1

3−0

.13

−0.1

3−0

.13

0.00

−0.5

0

Mal

aysi

a−0

.13

−0.4

0−0

.13

−0.2

50.

00−0

.50

Phili

ppin

es−0

.25

−0.3

0−0

.25

−0.3

50.

00−0

.50

Vie

t Nam

−0.2

7−0

.22

−0.2

7−0

.15

0.00

−0.2

7

Mya

nmar

−0.1

5−0

.16

−0.1

5−0

.16

0.00

−0.0

8

Oth

er S

outh

east

Asi

a−0

.15

−0.1

6−0

.15

−0.1

60.

00−0

.08

Chi

na−0

.22

−0.1

0−0

.22

−0.1

00.

00−0

.25

Rep

ublic

of K

orea

−0.3

8−0

.31

−0.3

8−0

.31

0.00

−0.2

0

Oth

er E

ast A

sia

−0.2

0−0

.15

−0.1

8−0

.05

0.00

−0.6

4

Res

t of t

he w

orld

−0.2

0−0

.15

−0.1

5−0

.05

0.00

−0.4

0

––––

Sou

rce:

See

note

16.

TA

BL

E B

.14

B

asel

ine

IMP

AC

T C

SE

est

imat

es, c

erea

ls a

nd

ro

ots

an

d t

ub

ers

pro

du

cts

Cas

sava

Oth

erS

wee

tan

d o

ther

coar

sepo

tato

esro

ots

and

Reg

ion/

Co

untr

yW

heat

Mai

zeR

ice

grai

nsP

ota

toes

and

yam

stu

bers

So

ybea

nsM

eals

Oils

Uni

ted

Stat

es0.

000.

000.

000.

000.

000.

000.

000.

000.

000.

00

EU

15−0

.09

−0.2

5−0

.13

−0.1

30.

000.

000.

000.

000.

000.

00

Japa

n−0

.76

−0.7

4−0

.55

−0.5

50.

000.

000.

000.

000.

000.

00

Aus

tral

ia0.

00−0

.02

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

Oth

er d

evel

oped

cou

ntri

es−0

.03

−0.0

5−0

.05

−0.0

10.

000.

000.

00−0

.07

−0.0

7−0

.07

Eas

tern

Eur

ope

0.01

0.00

−0.2

00.

000.

000.

000.

00−0

.05

−0.0

5−0

.05

Cen

tral

Asi

a0.

15−0

.20

0.00

−0.2

00.

000.

000.

00−0

.07

−0.0

7−0

.07

Res

t of F

orm

er S

ovie

t Uni

on0.

15−0

.03

−0.0

90.

000.

000.

000.

00−0

.07

−0.0

7−0

.07

Mex

ico

−0.1

8−0

.01

−0.0

5−0

.09

0.00

0.00

0.00

−0.0

4−0

.04

−0.0

4

Bra

zil

−0.2

4−0

.20

−0.0

5−0

.05

0.00

0.00

0.00

−0.0

9−0

.10

−0.1

0

Arg

entin

a−0

.04

−0.0

6−0

.05

−0.0

50.

000.

000.

00−0

.01

−0.0

2−0

.02

Col

ombi

a−0

.21

−0.2

8−0

.05

−0.0

50.

000.

000.

00−0

.11

−0.0

1−0

.01

Oth

er L

atin

Am

eric

a−0

.05

−0.1

1−0

.15

−0.1

50.

000.

000.

00−0

.16

−0.2

8−0

.28

Nig

eria

−0.3

0−0

.30

−0.2

5−0

.15

0.00

0.00

0.00

−0.1

0−0

.10

−0.1

0

Nor

ther

n Su

b-Sa

hara

n A

fric

a−0

.20

−0.2

1−0

.28

−0.2

50.

000.

000.

00−0

.07

−0.0

7−0

.07

Cen

tral

& W

este

rn S

ub-S

ahar

an A

fric

a−0

.10

−0.4

1−0

.20

−0.2

80.

000.

000.

00−0

.22

−0.2

2−0

.22

Sout

hern

Sub

-Sah

aran

Afr

ica

−0.1

5−0

.12

−0.1

8−0

.20

0.00

0.00

0.00

−0.2

2−0

.22

−0.2

2

Eas

tern

Sub

-Sah

aran

Afr

ica

−0.0

5−0

.30

−0.0

3−0

.18

0.00

0.00

0.00

−0.2

2−0

.22

−0.2

2

Egy

pt−0

.46

−0.1

2−0

.03

−0.2

30.

000.

000.

00−0

.25

−0.0

4−0

.04

Indi

a−0

.10

−0.0

9−0

.02

−0.0

20.

000.

000.

00−0

.30

−0.2

0−0

.20

Ban

glad

esh

−0.0

5−0

.20

−0.2

0−0

.10

0.00

0.00

0.00

−0.2

0−0

.10

−0.1

0

Paki

stan

−0.1

5−0

.10

−0.2

8−0

.26

0.00

0.00

0.00

−0.3

0−0

.16

−0.1

6

Tur

key

−0.3

2−0

.01

−0.3

8−0

.38

0.00

0.00

0.00

−0.3

7−0

.21

−0.3

7

Oth

er W

est A

sia/

Nor

th A

fric

a−0

.06

−0.0

7−0

.14

−0.0

70.

000.

000.

00−0

.15

−0.1

8−0

.18

Indi

a−0

.10

−0.0

9−0

.02

−0.0

20.

000.

000.

00−0

.30

−0.2

0−0

.20

Paki

stan

−0.1

5−0

.10

−0.2

8−0

.26

0.00

0.00

0.00

−0.3

0−0

.16

−0.1

6

Ban

glad

esh

−0.0

5−0

.20

−0.2

0−0

.10

0.00

0.00

0.00

−0.2

0−0

.10

−0.1

0

Oth

er S

outh

east

Asi

a−0

.05

−0.1

7−0

.10

−0.0

50.

000.

000.

00−0

.25

−0.2

0−0

.20

Indo

nesi

a−0

.10

−0.0

9−0

.04

−0.0

40.

000.

000.

00−

0.38

−0.1

2−0

.12

Tha

iland

−0.6

0−0

.05

−0.1

0−0

.10

0.00

0.00

0.00

−0.2

4−0

.10

−0.1

0

Mal

aysi

a−0

.28

−0.4

9−0

.25

−0.4

50.

000.

000.

00−0

.05

−0.0

1−0

.01

Phili

ppin

es−0

.28

−0.2

7−0

.60

−0.3

20.

000.

000.

00−0

.29

−0.1

0−0

.10

Vie

t Nam

−0.0

5−0

.25

−0.0

9−0

.12

0.00

0.00

0.00

−0.1

0−0

.11

−0.1

1

Mya

nmar

−0.0

5−0

.20

−0.0

9−0

.10

0.00

0.00

0.00

−0.1

0−0

.17

−0.1

7

Oth

er S

outh

east

Asi

a−0

.05

−0.2

0−0

.08

−0.1

00.

000.

000.

00−0

.10

−0.1

7−0

.17

Chi

na−0

.07

−0.0

7−0

.08

−0.0

80.

000.

000.

00−0

.02

−0.0

2−0

.02

Rep

ublic

of K

orea

−0.1

3−0

.67

−0.6

7−0

.67

0.00

0.00

0.00

−0.3

8−0

.38

−0.3

8

Oth

er E

ast A

sia

−0.0

5−0

.20

−0.0

9−0

.30

0.00

0.00

0.00

−0.1

0−0

.20

−0.2

0

Res

t of t

he w

orld

−0.1

0−0

.20

−0.0

9−0

.19

0.00

0.00

0.00

−0.1

0−0

.20

−0.2

0

––––

Sou

rce:

See

note

16.

171

Appendix C

REGIONAL FOOD SUPPLY AND

DEMAND DATA AND ANNUAL

GROWTH RATES

ASIA

TABLE C.1 Food supply and demand indicators, East Asia, 1967–97

Indicator 1967 1982 1990 1997

Population (millions) 812.2 1,085.6 1,220.8 1,320.4

Per capita area harvested (hectares) 0.12 0.09 0.08 0.07

Cereal production (million metric tons) 153.9 271.3 339.1 391.0

Cereal yield (kilograms/hectare) 1,637 2,887 3,576 4,157

Per capita cereal production (kilograms) 191.9 268.9 293.9 314.0


TABLE C.2 Annual growth rates in East Asia, 1967–97

Indicator 1967–82 1982–90 1990–97

(percent/year)

Population 2.0 1.5 1.1

Per capita area harvested −1.9 −1.3 −1.2

Cereal production 3.9 2.8 2.1

Cereal yield 3.9 2.7 2.2

Cereal imports 10.5 0.2 −3.2


172 APPENDIX C

TABLE C.3 Food supply and demand indicators, South Asia, 1967–97

Indicator 1967 1982 1990 1997

Population (millions) 663.7 935.6 1,117.1 1,288.5



Cereal yield (kilograms/hectare) 827 1,235 1,574 1,816



TABLE C.4 Annual growth rates in South Asia, 1967–97

Indicator 1967–82 1982–90 1990–97

(percent/year)


Per capita area harvested −1.6 −2.3 2.0



Cereal imports −8.9 1.3 −8.8


TABLE C.5 Food supply and demand indicators, Southeast Asia, 1967–97

Indicator 1967 1982 1990 1997

Population (millions) 263.2 372.8 436.9 492.4






REGIONAL FOOD SUPPLY AND DEMAND DATA AND ANNUAL GROWTH RATES 173

LATIN AMERICA

TABLE C.6 Annual growth rates in Southeast Asia,

1967–97

Indicator 1967–82 1982–90 1990–97

(percent/year)





Per capita cereal production 1.6 0.7 1.1


TABLE C.7 Food supply and demand indicators, Latin America, 1967–97

Indicator 1967 1982 1990 1997





Per capita cereal production (kilograms) 225 262 222 253


TABLE C.8 Annual growth rates in Latin America,

1967–97

Indicator 1967–82 1982–90 1990–97

(percent/year)



Cereal production 3.5 −0.1 3.7


Cereal imports n.a. 15.8 3.5


174 APPENDIX C

SUB-SAHARAN AFRICA

WEST ASIA/NORTH AFRICA

TABLE C.9 Food supply and demand indicators, Sub-Saharan Africa, 1967–97

Indicator 1967 1982 1990 1997




Cereal yield (kilograms/hectare) 746 937 924 948



TABLE C.10 Annual growth rates in Sub-Saharan

Africa, 1967–97

Indicator 1967–82 1982–90 1990–97

(percent/year)


Per capita area harvested −2.5 1.4 −0.1


Cereal yield 1.5 −0.2 0.3


TABLE C.11 Food supply and demand indicators, West Asia/North Africa, 1967–97

Indicator 1967 1982 1990 1997







REGIONAL FOOD SUPPLY AND DEMAND DATA AND ANNUAL GROWTH RATES 175

TABLE C.12 Annual growth rates, West Asia/North

Africa, 1967–97

Indicator 1967–82 1982–90 1990–97

(percent/year)





Cereal imports 3.4 3.7 2.0


177

Appendix D

PRODUCTION, DEMAND,

AND TRADE DATA BY

COMMODITY, 1997 AND 2020

TABLE D.1 Beef production, demand, and trade data, 1997 and 2020

1997 2020

Net Net

Region/Country Production Demand trade Production Demand trade

(1,000 metric tons)

Developed world 30,682 30,198 152 34,887 33,731 1,156

United States 11,755 11,928 −167 14,363 14,125 238

EU15 7,860 7,237 503 7,844 7,506 339

Former Soviet Union 4,872 5,578 −731 5,293 5,920 −628

Developing world 27,000 27,037 −152 50,529 51,685 −1,156

Latin America 13,013 12,450 500 21,192 19,369 1,823

Sub-Saharan Africa 2,470 2,452 11 5,148 5,212 −63

West Asia/North Africa 1,388 1,747 −377 2,352 3,095 −744

South Asia 4,125 3,953 158 8,040 8,241 −201

Southeast and East Asia 5,984 6,398 −425 13,771 15,696 −1,926

World 57,681 57,235 0 85,415 85,415 0

––––Source: IMPACT projections, June 2001.Note: For net trade, positive figures indicate exports, negative figures indicate imports.

178 APPENDIX D

TABLE D.2 Pork production, demand, and trade data, 1997 and 2020

1997 2020

Net Net


(1,000 metric tons)

Developed world 36,331 36,171 7 40,612 39,210 1,402

United States 8,074 7,943 156 10,132 9,057 1,074

EU15 16,818 15,704 992 17,786 16,441 1,345

Former Soviet Union 3,025 3,689 −646 3,411 3,882 −471

Developing world 46,740 46,676 −7 78,698 80,101 −1,402

Latin America 3,798 3,878 −105 6,488 6,439 49

Sub-Saharan Africa 775 804 −43 1,553 1,645 −92

West Asia/North Africa 60 66 −6 89 105 −16

South Asia 520 520 0 996 1,066 −70

Southeast and East Asia 41,531 41,344 157 69,495 70,703 −1,208

World 83,071 82,846 0 119,311 119,311 0


TABLE D.3 Sheep and goat production, demand, and trade data, 1997 and 2020

1997 2020

Net Net


(1,000 metric tons)

Developed world 3,366 3,147 211 4,207 3,725 482

United States 118 157 −39 153 189 −36

EU15 1,139 1,365 −222 1,278 1,467 −189

Former Soviet Union 606 635 −21 885 863 22

Developing world 7,578 7,758 −211 12,982 13,465 −482

Latin America 436 454 −24 708 731 −23

Sub-Saharan Africa 1,205 1,197 6 2,292 2,216 76


South Asia 1,767 1,760 7 3,363 3,413 −50

Southeast and East Asia 2,399 2,453 −48 3,626 3,914 −289

World 10,944 10,906 0 17,189 17,189 0


PRODUCTION, DEMAND, AND TRADE DATA BY COMMODITY, 1997 AND 2020 179

TABLE D.4 Poultry production, demand, and trade data, 1997 and 2020

1997 2020

Net Net


(1,000 metric tons)

Developed countries 29,665 28,150 701 40,748 37,593 3,155

United States 14,913 12,423 2,109 21,882 17,063 4,819

EU15 8,399 7,574 689 10,158 9,271 887

Former Soviet Union 1,070 2,154 −1,251 1,359 2,807 −1,449

Developing countries 28,838 29,197 −701 64,238 67,393 −3,155

Latin America 9,363 9,261 −60 18,983 18,395 588

Sub-Saharan Africa 957 1,052 −127 1,965 2,179 −214


South Asia 1,109 1,110 −1 2,935 3,064 −129


World 58,503 57,347 0 104,986 104,986 0


TABLE D.5 All meats production, demand, and trade data, 1997 and 2020

1997 2020

Net Net


(1,000 metric tons)

Developed world 100,044 97,666 1,071 120,454 114,259 6,195

United States 34,860 32,451 2,059 46,530 40,434 6,095

EU15 34,216 31,880 1,962 37,066 34,685 2,382

Former Soviet Union 9,573 12,056 −2,649 10,948 13,472 −2,526

Developing world 110,156 110,668 −1,071 206,447 212,644 −6,195

Latin America 26,610 26,043 311 47,371 44,934 2,437

Sub-Saharan Africa 5,407 5,505 −153 10,958 11,252 −293

West Asia/North Africa 6,368 7,140 −946 11,196 12,962 −1,767

South Asia 7,521 7,343 164 15,334 15,784 −450


World 210,199 208,334 0 326,901 326,901 0


TA

BL

E D

.6

Wh

eat

pro

du

ctio

n, d

eman

d, a

nd

tra

de

dat

a, 1

997

and

202

0

19

97

20

20

Reg

ion/

Co

untr

yY

ield

Pro

duct

ion

Dem

and

Net

tra

deY

ield

Pro

duct

ion

Dem

and

Net

tra

de

(kil

ogra

ms/

cap

ita

)(1

,000 m

etri

c to

ns)

(kil

ogra

ms/

cap

ita

)(1

,00

0 m

etri

c to

ns)

Dev

elop

ed w

orld

2,58

131

6,69

824

4,78

162

,318

2,99

037

4,09

926

8,15

510

5,94

4

Uni

ted

Stat

es2,

661

66,2

8335

,309

26,1

353,

326

85,9

5544

,771

41,1

84

EU

155,

793

99,4

2480

,578

12,9

966,

134

103,

185

84,9

1118

,274

Form

er S

ovie

t Uni

on1,

445

68,1

9573

,010

−2,7

471,

674

81,3

0875

,303

6,00

5

Dev

elop

ing

wor

ld2,

657

280,

002

340,

138

−62,

318

3,41

138

5,73

749

1,68

1−1

05,9

44

Latin

Am

eric

a2,

439

23,0

9429

,635

−7,8

453,

357

39,2

8541

,273

−1,9

88

Sub-

Saha

ran

Afr

ica

1,65

52,

633

9,41

1−6

,580

2,26

65,

081

19,0

60−1

3,98

0

Wes

t Asi

a/N

orth

Afr

ica

1,91

550

,487

75,1

09−2

5,90

82,

541

73,1

9411

0,96

7−3

7,77

3

Sout

h A

sia

2,35

888

,829

93,4

73−5

,261

3,17

212

7,37

314

7,05

7−1

9,68

4

Sout

heas

t and

Eas

t Asi

a4,

733

114,

959

132,

210

−16,

430

5,73

514

0,80

417

2,78

9−3

1,98

5

Wor

ld2,

616

596,

700

584,

919

03,

190

759,

836

759,

836

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

TA

BL

E D

.7

Ric

e p

rod

uct

ion

, dem

and

, an

d t

rad

e d

ata,

199

7 an

d 2

020

19

97

20

20

Reg

ion/

Co

untr

yY

ield

Pro

duct

ion

Dem

and

Net

tra

deY

ield

Pro

duct

ion

Dem

and

Net

tra

de

(kil

ogra

ms/

cap

ita

)(1

,000 m

etri

c to

ns)

(kil

ogra

ms/

cap

ita

)(1

,00

0 m

etri

c to

ns)

Dev

elop

ed w

orld

4,01

517

,025

17,2

0454

4,62

618

,124

18,7

90−6

66

Uni

ted

Stat

es4,

396

5,47

03,

533

2,20

35,

551

7,13

34,

666

2,46

7

EU

154,

271

1,79

62,

105

−396

4,94

02,

147

2,51

2−3

65

Form

er S

ovie

t Uni

on1,

557

781

1,26

4−5

231,

874

970

1,56

5−5

95

Dev

elop

ing

wor

ld2,

494

367,

053

363,

623

−54

3,20

248

5,08

848

4,42

266

6

Latin

Am

eric

a2,

144

13,0

8914

,751

−1,3

573,

199

20,9

4121

,709

−768

Sub-

Saha

ran

Afr

ica

1,07

47,

466

10,9

32−3

,798

1,66

015

,185

20,9

43−5

,758

Wes

t Asi

a/N

orth

Afr

ica

3,99

35,

453

8,15

1−3

,069

5,47

98,

275

13,4

18−5

,143

Sout

h A

sia

1,89

611

1,00

310

8,38

62,

931

2,61

115

7,94

215

8,71

1−7

69

Sout

heas

t and

Eas

t Asi

a6,

388

230,

030

221,

172

5,51

77,

967

282,

732

269,

251

13,4

81

Wor

ld2,

536

384,

078

380,

827

03,

238

503,

212

503,

212

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

TA

BL

E D

.8

Mai

ze p

rod

uct

ion

, dem

and

, an

d t

rad

e d

ata,

199

7 an

d 2

020

19

97

20

20

Reg

ion/

Co

untr

yY

ield

Pro

duct

ion

Dem

and

Net

tra

deY

ield

Pro

duct

ion

Dem

and

Net

tra

de

(kil

ogra

ms/

cap

ita

)(1

,000 m

etri

c to

ns)

(kil

ogra

ms/

cap

ita

)(1

,00

0 m

etri

c to

ns)

Dev

elop

ed w

orld

6,82

532

9,35

729

0,52

827

,507

8,14

541

0,64

834

3,67

766

,971

Uni

ted

Stat

es8,

123

238,

759

183,

057

44,8

109,

673

297,

186

227,

330

69,8

56

EU

158,

722

36,9

4737

,633

−1,1

669,

586

39,1

2940

,021

−891

Form

er S

ovie

t Uni

on2,

649

7,08

67,

464

−303

3,63

110

,351

7,57

12,

780

Dev

elop

ing

wor

ld2,

881

266,

347

294,

985

−27,

507

4,10

944

0,97

950

7,95

0−6

6,97

1

Latin

Am

eric

a2,

639

74,0

9375

,562

−2,3

043,

880

123,

197

118,

116

5,08

1

Sub-

Saha

ran

Afr

ica

1,26

726

,218

28,6

50−1

,578

1,76

445

,564

52,1

07−6

,543

Wes

t Asi

a/N

orth

Afr

ica

4,80

69,

488

18,2

96−9

,703

5,78

513

,413

27,7

77−1

4,36

4

Sout

h A

sia

1,64

513

,222

13,7

16−1

022,

115

18,7

5918

,965

−206

Sout

heas

t and

Eas

t Asi

a7,

270

143,

307

158,

687

−13,

765

10,5

2924

0,02

429

0,87

7−5

0,85

3

Wor

ld4,

233

595,

704

585,

513

05,

399

851,

627

851,

627

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

TA

BL

E D

.9

Oth

er c

oar

se g

rain

s p

rod

uct

ion

, dem

and

, an

d t

rad

e d

ata,

199

7 an

d 2

020

19

97

20

20

Reg

ion/

Co

untr

yY

ield

Pro

duct

ion

Dem

and

Net

tra

deY

ield

Pro

duct

ion

Dem

and

Net

tra

de

(kil

ogra

ms/

cap

ita

)(1

,000 m

etri

c to

ns)

(kil

ogra

ms/

cap

ita

)(1

,00

0 m

etri

c to

ns)

Dev

elop

ed w

orld

2,45

719

0,91

617

2,46

714

,455

2,83

322

1,52

119

1,63

729

,884

Uni

ted

Stat

es3,

507

27,3

1722

,531

3,61

64,

406

34,2

0728

,425

5,78

3

EU

154,

386

65,7

3352

,283

8,60

84,

774

68,0

6655

,594

12,4

71

Form

er S

ovie

t Uni

on1,

361

48,2

0249

,549

668

1,44

951

,851

51,7

8071

Dev

elop

ing

wor

ld1,

101

103,

653

119,

256

−14,

455

1,48

216

1,03

119

0,91

4−2

9,88

4

Latin

Am

eric

a2,

572

14,3

0818

,122

−3,7

563,

546

23,9

7729

,809

−5,8

32

Sub-

Saha

ran

Afr

ica

751

32,9

8633

,512

−418

1,12

162

,698

63,7

62−1

,065

Wes

t Asi

a/N

orth

Afr

ica

1,47

520

,025

27,4

35−6

,400

1,84

627

,592

43,4

23−1

5,83

1

Sout

h A

sia

878

22,4

2322

,750

−588

1,09

227

,723

28,5

65−8

42

Sout

heas

t and

Eas

t Asi

a3,

568

13,9

0817

,379

−3,2

774,

769

19,0

3725

,271

−6,2

33

Wor

ld1,

714

294,

569

291,

724

02,

048

382,

552

382,

552

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

TA

BL

E D

.10

A

ll ce

real

s p

rod

uct

ion

, dem

and

, an

d t

rad

e d

ata,

199

7 an

d 2

020

19

97

20

20

Reg

ion/

Co

untr

yY

ield

Pro

duct

ion

Dem

and

Net

tra

deY

ield

Pro

duct

ion

Dem

and

Net

tra

de

(kil

ogra

ms/

cap

ita

)(1

,000 m

etri

c to

ns)

(kil

ogra

ms/

cap

ita

)(1

,00

0 m

etri

c to

ns)

Dev

elop

ed w

orld

3,37

785

3,99

672

4,98

010

4,33

43,

976

1,02

4,39

282

2,25

920

2,13

3

Uni

ted

Stat

es5,

334

337,

829

244,

430

76,7

646,

469

424,

481

305,

192

119,

290

EU

155,

540

203,

900

172,

599

20,0

425,

970

212,

527

183,

038

29,4

89

Form

er S

ovie

t Uni

on1,

448

124,

264

131,

287

−2,9

051,

647

144,

480

136,

219

8,26

1

Dev

elop

ing

wor

ld2,

316

1,01

7,05

51,

118,

002

−104

,334

3,06

51,

472,

835

1,67

4,96

7−2

02,1

33

Latin

Am

eric

a2,

531

124,

584

138,

070

−15,

262

3,65

420

7,40

021

0,90

7−3

,507

Sub-

Saha

ran

Afr

ica

948

69,3

0382

,505

−12,

374

1,38

012

8,52

815

5,87

2−2

7,34

6

Wes

t Asi

a/N

orth

Afr

ica

1,97

585

,453

128,

991

−45,

080

2,57

412

2,47

419

5,58

5−7

3,11

1

Sout

h A

sia

1,81

423

5,47

723

8,32

5−3

,020

2,46

033

1,79

735

3,29

8−2

1,50

1

Sout

heas

t and

Eas

t Asi

a6,

395

502,

204

529,

448

−27,

955

8,48

068

2,59

775

8,18

8−7

5,59

0

Wor

ld2,

703

1,87

1,05

11,

842,

983

03,

383

2,49

7,22

72,

497,

227

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

TA

BL

E D

.11

P

ota

toes

pro

du

ctio

n, d

eman

d, a

nd

tra

de

dat

a, 1

997

and

202

0

19

97

20

20

Reg

ion/

Co

untr

yY

ield

Pro

duct

ion

Dem

and

Net

tra

deY

ield

Pro

duct

ion

Dem

and

Net

tra

de

(kil

ogra

ms/

cap

ita

)(1

,000 m

etri

c to

ns)

(kil

ogra

ms/

cap

ita

)(1

,00

0 m

etri

c to

ns)

Dev

elop

ed w

orld

17,2

2218

4,70

818

5,45

81,

052

19,7

2920

1,09

819

7,37

93,

720

Uni

ted

Stat

es38

,815

21,7

7221

,655

8248

,434

27,7

1327

,253

460

EU

1533

,535

47,8

7247

,933

847

38,6

0750

,563

49,0

261,

537

Form

er S

ovie

t Uni

on10

,922

68,5

5569

,293

−184

11,8

6270

,449

69,0

701,

379

Dev

elop

ing

wor

ld15

,279

117,

185

119,

101

−1,0

5223

,004

206,

558

210,

278

−3,7

20

Latin

Am

eric

a13

,877

15,2

0615

,821

−394

19,2

9123

,943

23,9

70−2

7

Sub-

Saha

ran

Afr

ica

5,92

22,

650

2,71

9−6

98,

644

4,88

95,

094

−205

Wes

t Asi

a/N

orth

Afr

ica

20,0

2815

,189

15,0

5634

529

,062

24,8

2924

,817

12

Sout

h A

sia

15,6

2824

,280

24,3

17−4

26,0

2455

,599

56,3

52−7

52

Sout

heas

t and

Eas

t Asi

a29

,147

59,8

5161

,142

−901

42,6

2897

,283

99,9

73−2

,690

Wor

ld16

,412

301,

892

304,

559

021

,263

407,

656

407,

656

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

TA

BL

E D

.12

S

wee

t p

ota

toes

an

d y

ams

pro

du

ctio

n, d

eman

d, a

nd

tra

de

dat

a, 1

997

and

202

0

19

97

20

20

Reg

ion/

Co

untr

yY

ield

Pro

duct

ion

Dem

and

Net

tra

deY

ield

Pro

duct

ion

Dem

and

Net

tra

de

(kil

ogra

ms/

cap

ita

)(1

,000 m

etri

c to

ns)

(kil

ogra

ms/

cap

ita

)(1

,00

0 m

etri

c to

ns)

Dev

elop

ed w

orld

18,6

652,

033

2,13

8−6

822

,116

2,33

22,

266

66

Uni

ted

Stat

es17

,480

589

607

−15

22,9

1073

469

540

EU

1510

,921

5485

−12

13,5

0066

92−2

6

Form

er S

ovie

t Uni

on0

00

00

00

0

Dev

elop

ing

wor

ld12

,963

165,

485

165,

506

6816

,292

216,

437

216,

503

−66

Latin

Am

eric

a7,

848

2,67

22,

685

3011

,283

3,98

53,

762

223

Sub-

Saha

ran

Afr

ica

7,75

139

,911

39,9

406

11,8

0673

,547

73,4

1812

9

Wes

t Asi

a/N

orth

Afr

ica

22,7

2920

219

84

35,1

9732

130

219

Sout

h A

sia

8,49

91,

662

1,66

11

10,1

582,

007

1,95

849

Sout

heas

t and

Eas

t Asi

a25

,183

120,

347

120,

331

2832

,232

135,

629

136,

086

−458

Wor

ld13

,011

167,

518

167,

644

016

,338

218,

769

218,

769

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

TA

BL

E D

.13

C

assa

va a

nd

oth

er r

oo

ts a

nd

tu

ber

s p

rod

uct

ion

, dem

and

, an

d t

rad

e d

ata,

199

7 an

d 2

020

19

97

20

20

Reg

ion/

Co

untr

yY

ield

Pro

duct

ion

Dem

and

Net

tra

deY

ield

Pro

duct

ion

Dem

and

Net

tra

de

(kil

ogra

ms/

cap

ita

)(1

,000 m

etri

c to

ns)

(kil

ogra

ms/

cap

ita

)(1

,00

0 m

etri

c to

ns)

Dev

elop

ed w

orld

12,4

3233

810

,354

−10,

624

16,7

9344

14,

465

−4,0

24

Uni

ted

Stat

es13

,000

324

5−2

5718

,926

427

4−2

70

EU

1516

,308

69,

111

−9,6

5520

,226

83,

349

−3,3

42

Form

er S

ovie

t Uni

on0

01

−10

00

0

Dev

elop

ing

wor

ld9,

539

178,

981

165,

232

10,6

2413

,077

274,

806

270,

782

4,02

4

Latin

Am

eric

a11

,384

31,1

7630

,996

221

16,8

2245

,223

43,4

021,

822

Sub-

Saha

ran

Afr

ica

8,00

097

,385

97,3

3649

11,6

9117

3,52

317

3,31

221

1

Wes

t Asi

a/N

orth

Afr

ica

57,6

6519

820

9−5

979

,230

268

276

−8So

uth

Asi

a19

,907

6,57

56,

599

−23

24,4

538,

440

8,64

4−2

03

Sout

heas

t and

Eas

t Asi

a28

,309

42,9

5329

,410

10,4

2936

,254

46,3

2544

,183

2,14

1

Wor

ld9,

543

179,

319

175,

587

013

,081

275,

248

275,

248

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

TA

BL

E D

.14

A

ll ro

ots

an

d t

ub

ers

pro

du

ctio

n, d

eman

d, a

nd

tra

de

dat

a, 1

997

and

202

0

19

97

20

20

Reg

ion/

Co

untr

yY

ield

Pro

duct

ion

Dem

and

Net

tra

deY

ield

Pro

duct

ion

Dem

and

Net

tra

de

(kil

ogra

ms/

cap

ita

)(1

,000 m

etri

c to

ns)

(kil

ogra

ms/

cap

ita

)(1

,00

0 m

etri

c to

ns)

Dev

elop

ed w

orld

17,2

2518

7,07

919

7,95

0−9

,640

19,7

4720

3,87

120

4,11

0−2

38

Uni

ted

Stat

es37

,587

22,3

6422

,507

−190

47,1

0428

,451

28,2

2223

0

EU

1533

,449

47,9

3257

,129

−8,8

2038

,507

50,6

3752

,467

−1,8

31

Form

er S

ovie

t Uni

on10

,922

68,5

5569

,294

−185

11,8

6270

,449

69,0

701,

379

Dev

elop

ing

wor

ld11

,777

461,

651

449,

839

9,64

016

,124

697,

801

697,

563

238

Latin

Am

eric

a11

,749

49,0

5449

,502

−143

17,0

8373

,151

71,1

342,

018

Sub-

Saha

ran

Afr

ica

7,87

613

9,94

613

9,99

5−1

411

,645

251,

959

251,

824

135

Wes

t Asi

a/N

orth

Afr

ica

20,2

4515

,589

15,4

6329

029

,351

25,4

1825

,395

23

Sout

h A

sia

15,6

3332

,517

32,5

77−2

624

,653

66,0

4666

,954

−906

Sout

heas

t and

Eas

t Asi

a28

,631

223,

151

210,

883

9,55

636

,610

279,

237

280,

242

−1,0

07

Wor

ld12

,959

648,

729

647,

790

016

,821

901,

673

901,

673

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

TA

BL

E D

.15

E

ggs

pro

du

ctio

n, d

eman

d, a

nd

tra

de

dat

a, 1

997

and

202

0

19

97

20

20

Reg

ion/

Co

untr

yP

rodu

ctio

nD

eman

dN

et t

rade

Pro

duct

ion

Dem

and

Net

tra

de

(1,0

00 m

etri

c to

ns)

Dev

elop

ed w

orld

17,8

9017

,827

7417

,860

17,7

2613

4

Uni

ted

Stat

es4,

615

4,49

512

24,

929

4,78

414

5

EU

155,

354

5,26

290

4,94

24,

812

130

Form

er S

ovie

t Uni

on2,

853

2,87

5−1

32,

785

2,83

2−4

7

Dev

elop

ing

wor

ld32

,884

32,9

55−7

448

,805

48,9

39−1

34

Latin

Am

eric

a4,

445

4,46

2−1

96,

395

6,35

738

Sub-

Saha

ran

Afr

ica

901

905

−91,

727

1,72

34

Wes

t Asi

a/N

orth

Afr

ica

2,21

52,

227

−10

3,50

73,

574

−67

Sout

h A

sia

2,14

22,

134

114,

560

4,55

19

Sout

heas

t and

Eas

t Asi

a23

,166

23,2

12−4

632

,588

32,7

02−1

13

Wor

ld50

,773

50,7

820

66,6

6666

,666

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

TA

BL

E D

.16

M

ilk p

rod

uct

ion

, dem

and

, an

d t

rad

e d

ata,

199

7 an

d 2

020

19

97

20

20

Reg

ion/

Co

untr

yP

rodu

ctio

nD

eman

dN

et t

rade

Pro

duct

ion

Dem

and

Net

tra

de

(1,0

00 m

etri

c to

ns)

Dev

elop

ed w

orld

338,

719

317,

630

20,0

1438

2,28

434

8,58

533

,699

Uni

ted

Stat

es70

,690

73,9

33−3

,269

86,0

8789

,477

−3,3

90

EU

1512

5,35

411

4,19

610

,639

138,

771

117,

575

21,1

97

Form

er S

ovie

t Uni

on66

,620

65,8

6673

369

,507

69,9

25−4

19

Dev

elop

ing

wor

ld20

7,96

322

7,67

5−2

0,01

438

6,01

541

9,71

4−3

3,69

9

Latin

Am

eric

a55

,670

61,3

32−5

,767

87,3

6791

,906

−4,5

39

Sub-

Saha

ran

Afr

ica

15,8

2618

,135

−2,2

7933

,519

38,1

15−4

,596

Wes

t Asi

a/N

orth

Afr

ica

25,4

6730

,167

−4,8

8541

,424

49,2

89−7

,864

Sout

h A

sia

96,3

5697

,097

−711

194,

208

198,

332

−4,1

23

Sout

heas

t and

Eas

t Asi

a14

,569

20,7

29−6

,227

29,3

6741

,674

−12,

308

Wor

ld54

6,68

254

5,30

50

768,

299

768,

299

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

TA

BL

E D

.17

M

eals

pro

du

ctio

n, d

eman

d, a

nd

tra

de

dat

a, 1

997

and

202

0

19

97

20

20

Reg

ion/

Co

untr

yP

rodu

ctio

nD

eman

dN

et t

rade

Pro

duct

ion

Dem

and

Net

tra

de

(1,0

00 m

etri

c to

ns)

Dev

elop

ed w

orld

71,7

3583

,702

−12,

873

104,

966

109,

829

−4,8

63

Uni

ted

Stat

es35

,600

29,0

695,

644

56,0

7943

,087

12,9

91

EU

1520

,645

35,9

81−1

5,09

428

,371

43,2

68−1

4,89

7

Form

er S

ovie

t Uni

on2,

832

2,89

0−5

63,

613

3,77

9−1

65

Dev

elop

ing

wor

ld85

,039

71,4

8312

,873

152,

684

147,

820

4,86

3

Latin

Am

eric

a34

,550

15,5

9618

,770

60,9

6629

,191

31,7

75

Sub-

Saha

ran

Afr

ica

2,96

32,

497

469

5,64

55,

410

235

Wes

t Asi

a/N

orth

Afr

ica

2,74

26,

024

−3,3

604,

568

10,9

73−6

,405

Sout

h A

sia

17,0

0412

,771

4,07

329

,215

28,4

1380

2

Sout

heas

t and

Eas

t Asi

a27

,725

34,5

50−7

,090

52,2

0773

,751

−21,

544

Wor

ld15

6,77

415

5,18

50

257,

649

257,

649

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

TA

BL

E D

.18

O

ils p

rod

uct

ion

, dem

and

, an

d t

rad

e d

ata,

199

7 an

d 2

020

19

97

20

20

Reg

ion/

Co

untr

yP

rodu

ctio

nD

eman

dN

et t

rade

Pro

duct

ion

Dem

and

Net

tra

de

(1,0

00 m

etri

c to

ns)

Dev

elop

ed w

orld

50,7

1051

,079

−604

61,5

2559

,530

1,99

5

Uni

ted

Stat

es16

,583

14,2

502,

205

21,9

2317

,767

4,15

6

EU

1518

,954

19,7

66−1

,147

20,8

3721

,618

−781

Form

er S

ovie

t Uni

on4,

178

5,64

6−1

,169

4,94

66,

452

−1,5

06

Dev

elop

ing

wor

ld63

,112

61,1

5660

411

4,35

711

6,35

2−1

,995

Latin

Am

eric

a14

,073

11,8

182,

108

24,7

4019

,762

4,97

8

Sub-

Saha

ran

Afr

ica

4,39

55,

338

−916

8,85

710

,868

−2,0

10

Wes

t Asi

a/N

orth

Afr

ica

2,57

36,

968

−4,7

224,

274

11,9

38−7

,664

Sout

h A

sia

10,2

9713

,824

−4,2

4820

,040

26,2

84−6

,245

Sout

heas

t and

Eas

t Asi

a31

,428

23,1

198,

124

55,9

2247

,314

8,60

8

Wor

ld11

3,82

211

2,23

50

175,

882

175,

882

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

TA

BL

E D

.19

S

oyb

ean

s p

rod

uct

ion

, dem

and

, an

d t

rad

e d

ata,

199

7 an

d 2

020

19

97

20

20

Reg

ion/

Co

untr

yP

rodu

ctio

nD

eman

dN

et t

rade

Pro

duct

ion

Dem

and

Net

tra

de

(1,0

00 m

etri

c to

ns)

Dev

elop

ed w

orld

75,8

4771

,949

3,11

710

1,68

396

,709

4,97

4

Uni

ted

Stat

es70

,852

45,9

6923

,614

94,8

6162

,879

31,9

82

EU

151,

381

16,6

31−1

4,67

61,

862

21,3

26−1

9,46

4

Form

er S

ovie

t Uni

on31

639

5−4

835

445

8−1

04

Dev

elop

ing

wor

ld69

,063

73,0

53−3

,117

125,

031

130,

006

−4,9

74

Latin

Am

eric

a45

,140

40,7

915,

151

81,8

3267

,823

14,0

09

Sub-

Saha

ran

Afr

ica

642

667

−25

1,07

81,

331

−252

Wes

t Asi

a/N

orth

Afr

ica

239

863

−633

365

1,62

2−1

,258

Sout

h A

sia

6,22

66,

317

−22

12,3

9413

,017

−623

Sout

heas

t and

Eas

t Asi

a16

,815

24,4

13−7

,586

29,3

6246

,208

−16,

846

Wor

ld14

4,91

014

5,00

10

226,

714

226,

714

0

––––

Sou

rce:

IMPA

CT

pro

ject

ions

, Jun

e 20

01.

Note

:Fo

r ne

t tra

de, p

ositi

ve fi

gure

s in

dica

te e

xpor

ts, n

egat

ive

figur

es in

dica

te im

port

s.

1 Unless otherwise noted, figures in this sectioncome from ACC/SCN (1992) and WHO (1997), ascited in Smith and Haddad (2000, Table 1).

2 Two standard deviations below a median value ofweight-for-age is considered a sign of malnutrition,using U.S. National Center for Health Statistics/World Health Organization standards.

3 Population statistics are calculated from FAO(2000a), with three-year averages centered on theend years used to calculate growth.

4 See Appendix A for a list of the countries includedin each region in this study.

5 Although the 12-member European Communitydid not become the 15-member European Unionuntil 1993, we will refer to the EU15 throughout thisreport. The member states are Austria, Belgium, Den-mark, Finland, France, Germany, Greece, Ireland,Italy, Luxembourg, the Netherlands, Portugal, Spain,Sweden, and the United Kingdom.

6 A revision of Chinese statistics would also lowergrowth rates of meat demand between 1990 and1997. However, while China has certainly drivenmeat demand and production increases in recentyears, other regions have also expanded demandrapidly, including Latin America and Southeast Asia.In any event, China’s production growth rates are

so high that even a revision downward would notchange the overall picture. Nevertheless, we mustawait better data for a definitive answer.

7 However, Smil (2000) points out that much of thearea removed from production is basically unsuitedto agriculture and produces a very low-yielding crop.

8 Other grain per capita demand declined during thisperiod, thus accounting for the fact that the sum ofthe wheat and maize increase is greater than the totalof 23 kilograms per capita.

9 All dollar amounts in this volume refer to U.S. dol-lars.

10 Southern Africa, in particular Zimbabwe duringthe early 1980s, serves as a good example.

11 Coarse grains other than maize.

12 However, it should be noted that other grainyields in Nigeria were extraordinarily high in 1982.

13 The treatment of WANA in the IMPACT modelis relatively aggregated. Other WANA includes Alge-ria, Cyprus, Iran, Iraq, Jordan, Kuwait, Lebanon,Libya, Morocco, Saudi Arabia, Syria, Tunisia, UnitedArab Emirates, and Yemen.

14 WANA actually had the fastest population growthrate of any region in the world between 1982 and 1990at 3.0 percent.

NOTES

195

196 NOTES

15 All population projections are from UN (1998).

16 Sources include OECD (1999); Ingco and Ng(1998); Fan and Tuan (1998); Finger et al. (1996);McDougall et al. (1998); UNCTAD various years;Valdes (1996); Valdes and Schaeffer (1995a); Valdesand Schaeffer (1995b); Valdes and Schaeffer (1995c);Valdes and Schaeffer (1995d).

17 Calculations of value for total agricultural trade,total agricultural production, and so forth, refer onlyto the IMPACT commodities; they do not includethe large number of agricultural products not sub-sumed within the IMPACT model. Calculations invalue terms are produced by multiplying projectedIMPACT volume by price.

18 The baseline irrigated area data were obtainedfrom FAO (2000b) for all developing countriesexcept China. For China, national data were used,and various sources were used for developed coun-tries. All values are calibrated to IMPACT baselineharvested area and production.

19 The moral issues surrounding high populationgrowth in food-insecure regions—including the dif-ficult question of the intrinsic worth of life evenunder conditions of extreme poverty—lie outsidethe scope of this report. Nevertheless, the authorswish to recognize the importance of these issues.

20 An alternative specification for a low populationgrowth scenario would maintain per capita incomegrowth at the level projected under the baseline sce-nario, necessitating a decline in total income growthrates to compensate for slower population growth.This alternative scenario is somewhat more consis-tent with the dynamic view of population growth thatemphasizes labor’s role as a factor of production andsource of additional knowledge. Under this scenario,per capita income in the developing world in 2020declines to $2,458, from $2,599 under the standardlow population growth scenario. This decline in percapita income corresponds in turn to an increase inthe number of children under the age of five who aremalnourished of 1 million children in the developingworld in 2020. One other notable impact of this alter-native low population growth scenario is a decline innet developing world meat imports of 1 million tons,from 7 million tons under the standard low popula-tion growth scenario to 6 million tons. Other ele-

ments of food consumption remain similar betweenthese two scenarios.

21 It should be noted, however, that slower popu-lation growth rates in the developed world wouldintensify a host of nonagricultural problems in thesecountries related to the costs of an aging populationat near zero growth. Measuring these effects isbeyond the scope of this paper.

22 However, the sword cuts both ways: the baselevel for demand growth has also increased, andgrowth in total demand requirements is projectedto be as high over the next 25 years as over the pre-vious 25 years.

23 Also note that the 1997 base yield remains con-stant in all scenarios, so that the shock to yieldgrowth rates only affects projected yield increasesbetween 1997 and 2020.

24 Free-range chickens consume up to 20 percentmore feed than their caged counterparts (Smil 2000).

25 It should be noted that the net effect on con-sumers of an increase in prices due to full trade lib-eralization depends on the level of distortions cur-rently facing these consumers under the currenttrading regime. While international cereal and live-stock prices will increase under trade liberalization,consumers living under particularly heavily taxedsystems will pay lower prices overall.

26 The world price and global net benefits estimatedhere are similar in magnitude to those estimated byDiao, Somwaru, and Roe (2001) using a generalequilibrium model for full agricultural trade liber-alization including a few additional commoditiessuch as sugar and fruits and vegetables. Diao,Somwaru, and Roe estimated static welfare net ben-efits of $31.1 billion, and an increase in the index ofworld agricultural prices by 11.6 percent.

27 For crops in which area actually declined duringthe projections period, a doubling of this decline wasassumed.

28 For example, the average size of broiler produc-tion units grew from 345 animals in the mid-1970sto 14,000 animals in the mid-1990s (Steinfeld andKamakawa 1999).

29 While the precise percentage of pork producedunder intensive conditions is limited, the number of

NOTES 197

pig-holding households decreased from 1.076 mil-lion in 1975 to 0.636 million in 1995, while the totalnumber of pigs increased by more than 100 percent(Riethmueller 1997, as quoted in Steinfeld andKamakawa 1999).

30 Although Sub-Saharan Africa has the highest percapita cereal area harvested of any IMPACT regionat 0.13 hectares per capita, this figure has actuallyincreased from 0.12 hectares per capita in 1982.

31 Yield growth rates vary by crop.

32 Cleaver and Schreiber (1994), however, estimatethat fertilizer use would have to rise 15 percentannually to achieve 3.5 percent annual yield growth..

33 With per capita consumption in 1997 equal to 2.8kilograms per capita for beef, 0.5 kilograms percapita for pork, 0.7 kilograms per capita for sheepand goat meat, and 0.5 kilograms per capita forpoultry products.

34 India’s urban population as a percentage of itstotal population is expected to rise from 26 percentin 1993 to 35 percent in 2020.

35 Importing livestock would be by definition moreexpensive than importing the feed necessary todomestically produce livestock, since the value-added would no longer occur domestically.

36 Even the East Asian economic crisis will not slowthe pace of urbanization in Asia.

REFERENCES

199

Abdulai, A. 1999. Household food demand analy-sis in India. Journal of Agricultural Economics 50(2):316–327.

Abdulai, A., and P. Hazell. 1995. Prerequisites andpriorities for sustainable economic develop-ment. In Re-establishing agriculture as a priority for

development policy in Sub-Saharan Africa, ed. A.Abdulai and C. Delgado. IFPRI OccasionalPaper. Washington, D.C.: International FoodPolicy Research Institute.

Abrol, I. P. 1987. Salinity and food production inthe Indian sub-continent. In Water and water pol-

icy in world food supplies, ed. W. R. Jordan. Col-lege Station,Tex., U.S.A.: Texas A&M Univer-sity Press.

ACC/SCN (United Nations Administrative Com-mittee on Coordination/Sub-Committee onNutrition). 1992. Second report on the world nutri-

tion situation,vol. I. Geneva.Ahlburg, D. A. 1998. Julian Simon and the popula-

tion growth debate. Population and Development

Review 24(2): 317–327.Ahlburg, D. A., A. C. Kelley, and K. O. Mason, eds.

1996. The impact of population growth on well-being

in developing countries. Berlin: Springer-Verlag.Alexandratos, N. 1995. World agriculture: Towards

2010: An FAO study. Rome, Italy: Food and Agri-culture Organization of the United Nations.

Alston, J. M., C. Chan-kang, M. C. Marra, P. G.Pardey, and TJ Wyatt. 2000. A meta-analysis of

rates of return to agricultural R&D: Ex pede her-

culem? IFPRI Research Report 113. Washington,D.C.: International Food Policy Research Insti-tute.

Bhadra, D., and A. S. P. Brandão. 1993. Urbaniza-

tion, agricultural development, and land allocation.

World Bank Discussion Paper 201. Washington,D.C.: World Bank.

Bhalla, G. S., P. Hazell, and J. Kerr. 1999. Prospects

for India’s cereal supply and demand to 2020. 2020Vision for Food, Agriculture, and the Environ-ment Discussion Paper 29. Washington, D.C.:International Food Policy Research Institute.

Bingsheng, K. E. 1997. Recent developments in thelivestock sector in China and changes in the live-stock/feed relations. A draft background paperprepared by the Food and Agriculture Organi-zation of the United Nations, Rome.

Bojo, J. 1996. The costs of land degradation in Sub-Saharan Africa. Ecological Economics 16: 161–173.

Bongaarts, J. 1996. Population pressure and thefood supply system in the developing world. Pop-

ulation and Development Review 22(3): 483–503.

Bongaarts, J., and R. A. Bulatao. 1999. Completingthe demographic transition. Population and Devel-

opment Review 25(3): 515–529.

200 REFERENCES

Boserup, E. 1981. Population and technological

change: A study of long-term trends. Chicago: Uni-versity of Chicago Press.

Bouis, H. 1994. Changing food consumption pat-terns in Asia and prospects for improvements innutrition: Implications for agricultural produc-tion policies. Paper presented at a symposiumsponsored by the Asian Productivity Organiza-tion, Changing Dietary Intake and Food Con-sumption, 13–20 December, Tokyo, Japan.

Braun, J. von, E. Serova, H. tho Seeth, and O.Melyukhina. 1996. Russia’s food economy in transi-

tion: Current policy issues and the long-term outlook.

2020 Vision for Food, Agriculture, and the Envi-ronment Discussion Paper 18. Washington, D.C.:International Food Policy Research Institute.

Brown, Lester R. 1995. Who will feed China? Wake-

up call for a small planet. New York: W. W. Nor-ton & Company.

Brown, Lester R., and H. Kane. 1994. Full house:

Reassessing the earth’s population carrying capac-

ity. New York: W. W. Norton & Company.Brown, Lynn R. 1996. The potential impact of AIDS

on population and economic growth rates. 2020Vision for Food, Agriculture, and the Environ-ment Discussion Paper 15. Washington, D.C.:International Food Policy Research Institute.

Bumb, B. L., and C. A. Baanante. 1996. The role of

fertilizer in sustaining food security and protecting

the environment to 2020. 2020 Vision Food, Agri-culture, and the Environment Discussion Paper17. Washington, D.C.: International Food Pol-icy Research Institute.

Buringh, P., and R. Dudal. 1987. Agricultural landuse in space and time. In Land transformation in

agriculture, ed. M. G. Wolman and F. G. A.Fournier. New York: John Wiley.

Byerlee, D., and C. K. Eicher, eds. 1997. Africa’s

emerging maize revolution. Boulder, Colo., U.S.A.:Lynne Reiner.

Byerlee, D., and P. W. Heisey. 1996. Past andpotential impacts of maize research in Sub-Saharan Africa: A critical assessment. Food Pol-

icy 21(3): 255–277.Caldwell, J. C. 2000. Rethinking the African AIDS

epidemic. Population and Development Review

26(1): 117–135.Capps, O., Jr., and J. D. Schmitz. 1991. A recogni-

tion of health and nutrition factors in fooddemand analysis. Western Journal of Agricultural

Economics (U.S.A.) 6 (July): 21–35.

Cassman, K. 1999. Ecological intensification ofcereal production systems: Yield potential, soilquality, and precision agriculture. ColloquiumPaper. Proceedings of the National Academy of Sci-

ence 96(11): 5952–5959.Cassman, K., and P. L. Pingali. 1996. Intensification

of irrigated rice systems: Learning from the pastto meet future challenges. In Explorations in agri-

cultural resource economics: Rice case studies and

reviews (1987–1995). Los Baños, Laguna, Philip-pines: International Rice Research Institute.

CGIAR (Consultative Group on International Agri-cultural Research). 1999. Annual report 1999.

Washington, D.C.Chambers, R. 1988. Managing canal irrigation:

Practical analysis from South Asia. Press Syndi-cate of the University of Cambridge, Cam-bridge, U.K.

Cleaver, K. M., and G. A. Schreiber. 1994. Revers-

ing the spiral: The population, agriculture, and envi-

ronment nexus in Sub-Saharan Africa. Directionsin Development Series. Washington, D.C.:World Bank.

Crosson, P. 1986. Sustainable food production:Interactions among natural resources, technol-ogy, and institutions. Food Policy 11(2): 143–156.

———. 1995. Soil erosion and its on-farm productiv-

ity consequences: What do we know? DiscussionPaper 95-29. Washington, D.C.: Resources forthe Future.

Davis, G. C. 1997. The logic of testing structuralchange in meat demand: A methodologicalanalysis and appraisal. American Journal of Agri-

cultural Economics 79 (November): 1186–1192.de Haan, C., H. Steinfeld, and H. Blackburn. 1996.

Livestock and the environment: Finding a balance.

Report of a study sponsored by the Commissionof the European Communities, the World Bank,and the governments of Denmark, France, Ger-many, the Netherlands, United Kingdom, andUnited States of America. Rome: Food and Agri-culture Organization of the United Nations.

Delgado, C. 1996. Agricultural transformation: Keyto broad-based growth and poverty alleviationin Africa. Agenda for Africa’s Economic Renewal:

U.S.–Third World Policy Perspectives No. 21. Wash-ington, D.C.: Overseas Development Council.

Delgado, C., M. Rosegrant, H. Steinfeld, S. Ehui,and C. Courbois. 1999. Livestock to 2020: The next

food revolution. 2020 Vision for Food, Agricul-ture, and the Environment Discussion Paper 28.

REFERENCES 201

Washington, D.C.: International Food PolicyResearch Institute.

Diao, X., A. Somwaru, and T. Roe. 2001. A global

analysis of agricultural trade reform in WTO mem-

ber countries. Minneapolis, St. Paul, Minn.,U.S.A.: Economic Development Center, Uni-versity of Minneapolis.

Díaz-Bonilla, E. 1990. Politicas macroeconomicasy sectoriakes y les estrategias de desarrollo rural.In Politicas Diferenciadas Para e Desarrollo Rural,

ed. A. Barril. IICA–Ministerio de Agricultura,Chile.

———. 1997. South American wheat markets andMERCOSUR. Trade and Macroeconomics Divi-sion Discussion Paper 23. International FoodPolicy Research Institute, Washington, D.C.

———. 1999. South American wheat markets. InEconomics of world wheat markets, ed. J. M. Antleand V. H. Smith. Oxon, U.K.: CAB International.

Dogra, B. 1986. The Indian experience with largedams. In The social and environmental effects of

large dams, vol. 2, ed. E. Goldsmith and N. Hild-yard. London: Wadebridge Ecological Center.

Dyson, T. 1996. Population and food: Global trends

and future prospects. London and New York:Routledge.

Eales, J., J. Hyde, and L. F. Schrader. 1998. A noteon dealing with poultry in demand analysis.Journal of Agricultural and Resource Economics

27 (October): 558–567.Eales, J., and L. J. Unnevehr. 1993. Simultaneity

and structural change in U.S. meat demand.American Journal of Agricultural Economics

75 (May): 259–268.Ehrlich, P. R., A. H. Ehrlich, and G. C. Daily. 1993.

Food security, population, and environment.Population and Development Review 19(1): 1–32.

Eichenwald, K. 2001. Redesigning nature: Hard les-sons learned; biotechnology food: From the labto a debacle. New York Times, 25 January, finaledition.

Ekboir, J., and G. Parellada. 2000. Continuous com-plex innovation processes: Public-private inter-actions and technology policy. Paper presentedat a meeting of the International Associationof Agricultural Economics, 1988. Populationgrowth, infrastructure, and real incomes inNorthern India. Population, food, and rural devel-

opment, ed. R. D. Lee, W. B. Arthur, A. C. Kel-ley, G. Rodgers, and T. N Srinivasan. Oxford,England: Clarendon Press.

Evenson, Robert E. 1988. Population growth, infra-structure, and real incomes in North India. In Pop-

ulation, food and rural development. Oxford: Claren-don Press; New York: Oxford University Press.

Evenson, R. E., C. E. Pray, and M. W. Rosegrant.1999. Agricultural research and productivity growth

in India. Research Report 109. Washington, D.C.:International Food Policy Research Institute.

Faini, R. 1992. Infrastructure, relative prices andagricultural adjustment. In Open economies:

Structural adjustment and agriculture, ed. I. Goldinand L. A. Winters. Cambridge, U.K.: CambridgeUniversity Press.

Fan, S., P. Hazell, and S. Thorat. 1999. Linkages

between government spending, growth, and poverty

in rural India. Research Report 110. Washing-ton, D.C.: International Food Policy ResearchInstitute.

Fan, S., and F. Tuan. 1998. Evolution of Chinese andOECD agricultural policy: Long-term lessons forChina. Paper prepared for the workshop on Agri-cultural Policies in China and OECD Countries:Review, Outlook, and Challenges, organized bythe OECD Directorate for Food, Agriculture andFisheries, 19–20 November, Paris.

Fan, S., L. Zhang, and X. Zhang. 2000. Growth andpoverty in rural China: The role of public invest-ments. Environment and Production Technol-ogy Discussion Paper 66. International FoodPolicy Research Institute, Washington, D.C.

FAO (Food and Agriculture Organization of theUnited Nations). 1994. Assessing the contribution

of high potential areas in developing countries to

improving food security on a sustainable basis. Pro-visional agenda, Committee on World FoodSecurity. Rome: FAO.

———. 2000a. FAOSTAT database. <http://faostat.fao.org>, Rome.

———. 2000b. Agriculture: Towards 2015/30.Technical Interim Report. Global PerspectiveStudies Unit, Rome.

Fedoroff, N. V., and J. E. Cohen. 1999. Introduc-tion to “Plants and population: Is there time?”Proceedings of the National Academy of Science

96(11).Finger, J. M., M. D. Ingco, and U. Reincke. 1996.

The Uruguay Round: Statistics on tariff concessions

given and received. Washington, D.C.: WorldBank.

Gao, X. M., E. J. Wailes, and G. L. Cramer. 1996.A two-stage rural household demand analysis:

202 REFERENCES

Microdata evidence from Jiangsu Province,China. American Journal of Agricultural Economics

78 (August): 604–613.Garrett, J. L. 1997. Challenges to the 2020 Vision for

Latin America: Food and agriculture since 1970. 2020Vision for Food, Agriculture, and the Environ-ment Discussion Paper 21. Washington, D.C.:International Food Policy Research Institute.

Goldman, R. H., and S. Block, eds. 1993. Proceed-

ings from the symposium on agricultural transfor-

mation in Africa. ARAP II Technical Report No.137. Cambridge, Mass.: Abt Associates.

Hayami, Y., and M. Kikuchi. 1983. New rice technol-

ogy, intra-rural migration, and institutional innova-

tion in the Philippines. Los Baños, Laguna, Philip-pines: International Rice Research Institute.

Hayami, Y., and J. P. Platteau. 1997. Resource endow-

ments and agricultural development: Africa vs. Asia.

Serie Recherche–No. 192. Namur, Belgium:Centre de Recherche en Économie du Développe-ment (CRED).

Herdt, R. W. 1998. Reflections on keeping Asia’sfood baskets full. American Journal of Agricultural

Economics 80(5): 969–972.Hobbs, P., and M. Morris. 1996. Meeting South Asia’s

future food requirements from rice wheat cropping

systems: Priority issue facing researchers in the post-

Green Revolution era. Natural Resources GroupWorking Paper 96-01. Mexico: Centro Inter-nacional de Mejoramiento de Maiz y Trigo(CIMMYT).

Hoffman, D. 1999. Asian livestock to the year 2000

and beyond. Working Paper Series 1/2. Bangkok:Food and Agricultural Organization of theUnited Nations, Regional Office for Asia and thePacific.

Hopper, G. R. 1999. Changing food production andquality of diet in India, 1947–1998. Population

and Development Review 25(3): 443–477.Huang, J., and H. Bouis. 1996. Structural changes in

the demand for food in Asia. 2020 Vision for Food,Agriculture, and the Environment DiscussionPaper 11. Washington, D.C.: International FoodPolicy Research Institute.

Ingco, M., and F. Ng. 1998. Distortionary effects of

state trading in agriculture: Issues for the next round

of multilateral trade negotiations. Washington,D.C.: World Bank.

International Road Federation, 1999. World road

statistics 1999. Geneva and Washington, D.C.

Jaeger, W. K. 1992. The effects of economic policies on

African agriculture. World Bank DiscussionPaper 147, Africa Technical Department Series.Washington, D.C.: World Bank.

Janssen, B. H. 1993. Integrated nutrient manage-ment: The use of organic and mineral fertiliz-ers. In The role of plant nutrients for sustainable

food crop production in Sub-Saharan Africa, ed. H.van Reuler and W. H. Prins. Leidschendam, theNetherlands: Dutch Association of FertilizerProducers (VKP).

Jones, W. I. 1995. The World Bank and irrigation.

Washington, D.C.: World Bank.Karshenas, M. 1999. Structural obstacles to economic

adjustment in the MENA region: The international

trade aspects. Working Paper Series 88. London:Department of Economics, School of Orientaland African Studies.

Kendall, H. W., and D. Pimentel 1994. Constraintson the expansion of the global food supply.Ambio 23(3):198–205.

Kherallah, M., C. Delgado, E. Gabre-Madhin, N.Minot, and M. Johnson. 2000. The road half trav-

eled: Agricultural market reforms in Sub-Saharan

Africa: A synthesis of research findings. Washing-ton, D.C.: International Food Policy ResearchInstitute.

Killick, T. 1994. Africa’s post-independence devel-opment experiences. In Economic development,

ed. E. Grilli and D. Salvatore. Handbook of Com-

parative Economic Policies, vol. 4. Westport,Conn., U.S.A: Greenwood Press.

Kinnucan, H. W., H. Xiao, C.-J. Hsia, and J. D. Jack-son. 1997. Effects of health information andgeneric advertising on U.S. meat demand. Amer-

ican Journal of Agricultural Economics 79 (Febru-ary): 13–23.

Koester, U., and S. Tangermann. 1990. The Euro-pean Union. In Agricultural protectionism in the

industrialized world, ed. F. H. Sanderson. Wash-ington, D.C.: Resources for the Future.

Lal, R. 1995. Erosion–crop productivity relation-ships for soil of Africa. Soil Science Society of Amer-

ica Journal 59(3): 661–667.Linneman, H., J. De Hoogh, M. A. Keyser, and H.

D. J. Van Heemst. 1979. Potential world foodproduction. In MOIRA. Model of International

Relations in Agriculture. Report of the Project Group

on Food for a Doubling World Population. Amster-dam: North-Holland.

REFERENCES 203

Luyten, J. C. 1995. Sustainable world food production

and environment. Report 37. Wageningen, theNetherlands: Agricultural Research Depart-ment, Research Institute for Agrobiology andSoil Fertility.

Macours, K., and J. F. M. Swinnen. 2000a. Causesof output decline in economic transition: Thecase of Central and Eastern European agricul-ture. Journal of Comparative Economics 28(1):172–207.

———. 2000b. Agrarian transition and productiv-ity patterns: Synthesis of experiences in East-ern Europe, the former Soviet Union, andEast Asia. Paper presented at a conference ofthe International Association of AgriculturalEconomics.

McDougall, R. A., A. Elbehri, and T. P. Truong.1998. Global trade assistance and protection: The

GTAP 4 data base. Lafayette, Ind., U.S.A.: PurdueUniversity, Center for Global Trade Analysis.

McDowell, H. 1991. Implications of EC 1992 forU.S. agricultural commodity trade. In EC 1992

implications for world food and agricultural trade,

ed. D. R. Kelch. Staff Report No. AGES 9133.Washington, D.C.: Agriculture and TradeAnalysis Division, Economic Research Service,U.S. Department of Agriculture.

McGuirk, A., and Y. Mundlak. 1991. Incentives and

constraints in the transformation of Punjab agricul-

ture. Research Report 87. Washington, D.C.:International Food Policy Research Institute.

Mundlak, Y., D. F. Larson, and R. Butzer. 1997. The

determinants of agricultural production: A cross-

country analysis. World Bank Discussion Paper1827. Washington, D.C.: World Bank.

Murgai, R. 1999. The Green Revolution and the pro-

ductivity paradox: Evidence from the Indian Punjab.

Policy Research Working Paper 2232. Wash-ington, D.C.: World Bank.

Nordblom, T. L., and F. Shomo. 1995. Food and feed

prospects to 2020 in the West Asia/North Africa

region. ICARDA Social Science Papers 2. Syria:International Center for Agricultural Researchin the Dry Areas.

North, D. C. 1994. Constraints on institutionalinnovation. In Agriculture, environment, and

health: Sustainable development in the 21st century,

ed. V. W. Ruttan. Minneapolis, Minn., U.S.A.:University of Minnesota Press.

OECD (Organisation for Economic Co-operation

and Development). 1999. Producer and con-sumer support estimates: OECD database 1986–1998. CD-ROM.

Oldeman, L. R. 1998. Soil degradation: A threat to

food security? Report 98/01. Wageningen, theNetherlands: International Soil Reference andInformation Centre.

Oldeman, L. R., R. T. A. Hakkeling, and W. G. Som-broek. 1991. World map of the status of human-

induced soil degradation: An explanatory note.

Wageningen, the Netherlands: International SoilReference and Information Centre, Nairobi:United Nations Environment Programme.

Oram, P., P. Hazell, T. Nordblom, N. Chaherli, andT. Ngaido. 1998. International conference on agri-

cultural growth, sustainable resource management

and poverty alleviation in the low rainfall areas of

West Asia and North Africa. Environment andProduction Technology Workshop SummaryPaper No. 6. Washington, D.C.: InternationalFood Policy Research Institute.

Ostrom, E. 1990. Governing the commons: The evolu-

tion of institutions for collective action. Cambridge,U.K. and New York: Cambridge UniversityPress.

Ouraga-Djoussou, L. H., and M. Bokanga. 1998.Cassava and wheat consumption in Africa: Newopportunities for cassava in the 21st century. InRoot crops and poverty alleviation, ed. M.O. Ako-roda and I. J. Ekanayake. Proceedings of theSixth Triennial Symposium of the InternationalSociety for Tropical Root Crops (ISTRC)–Africa Branch, 22–28 October, 1995, Lilongwe,Malawi. Lilongwe, Malawi: ISTRC.

Pardey, P. G., J. M. Alston, J. E. Christian, and S.Fan. 1996. Hidden harvest: U.S. benefits from inter-

national research aid. Washington, D.C.: Inter-national Food Policy Research Institute.

Paroda, R. S. 1998. Priorities and opportunities ofrice production and consumption in India forself-sufficiency. In Sustainability of rice in the

global food system, ed. N. G. Dowling, S. M.Greenfield, and K. S. Fischer. Los Baños, Philip-pines: Pacific Basin Study Center, InternationalRice Research Institute.

Pingali, P. L., and H. P. Binswanger. 1988. Popu-lation density and farming systems: The chang-ing locus of innovations and technical change.In Population, food, and rural development, ed. R.D. Lee, W. B. Arthur, A. C. Kelley, G. Rodgers,

204 REFERENCES

and T. N. Srinivasan. Oxford, U.K.: ClarendonPress.

Pingali, P. L., and P. W. Heisey. 1999. Cereal crop

productivity in developing countries: Past trends and

future prospects. Mexico: Centro Internacional deMejoramiento de Maiz y Trigo (CIMMYT).

Pingali, P. L., M. Hossain, and R. V. Gerpacio.1997.Asian rice bowls: The returning crisis? New York:CAB International.

Pingali, P. L., and M. W. Rosegrant. 1998. Supply-ing wheat for Asia’s increasing westernizeddiets. Paper presented at the annual meeting ofthe American Agricultural Economics Associa-tion, 2–5 August, Salt Lake City, Utah.

Pinstrup-Andersen, P., and M. J. Cohen. 2000.Biotechnology and the CGIAR. Paper preparedfor the International Conference on SustainableAgriculture in the Next Millenium—The Impactof Modern Agricultural Biotechnology onDeveloping Countries, sponsored by Friends ofthe Earth Europe, Oxfam, Solidarity Belgium,and the Dag Hammarskjold Foundation, heldin Brussels, Belgium, 28–31 May, 2000.

Plucknett, D. L. 1995. Prospects of meeting futurefood needs through new technology. In Popula-

tion and food in the early twenty-first century: Meet-

ing future food demand of an increasing population,

ed. N. Islam. Washington, D.C.: InternationalFood Policy Research Institute.

Pritchett, L. A. 1996. Population growth, factor accu-

mulation, and productivity. Policy ResearchWorking Paper 1567. Washington, D.C.: WorldBank.

Ranong, V. N. 1999. The financial crisis and thelivestock sector in Thailand. Thailand Devel-opment Research Institute, Bangkok. In Pro-

ceedings of the workshop on the implications of the

Asian economic crisis for the livestock industry held

in Bangkok, Thailand, 6–9 July, 1999. Rome: Foodand Agriculture Organization of the UnitedNations.

Reca, L., and E. Díaz-Bonilla. 1997. Changes in Latin

American agricultural markets. Trade and Macro-economics Division Discussion Paper 24. Wash-ington, D.C.: International Food PolicyResearch Institute.

Rosegrant, M., and X. Cai. 2000. Water scarcity andfood security: Alternative futures for the 21stcentury. Paper presented at the Stockholm

Water Symposium 2000, 14–17 August, Stock-holm, Sweden.

Rosegrant, M., and P. Hazell. 2000. Transforming

the rural Asian economy: The unfinished revolution.

New York: Oxford University Press.Rosegrant, M. W., and N. D. Perez. 1997. Water

resources development in Africa: A review and syn-

thesis of issues, potentials, and strategies for the

future. Environment and Production Technol-ogy Discussion Paper 28. Washington, D.C.:International Food Policy Research Institute.

Rosegrant, M. W., and P. L. Pingali. 1994. Policyand technology for rice productivity growthin Asia. Journal of International Development

6(6): 665.Rosegrant, M. W., M. A. Sombilla, R. V. Gerpacio,

and C. Ringler. 1997. Global food markets andU.S. exports in the 21st century. Revisions of apaper presented at the Illinois World Food andSustainable Agricultural Program ConferenceMeeting the Demand for Food in the 21st Cen-tury: Challenges and Opportunities for IllinoisAgriculture, 28 May, Urbana-Champaign.

Rosegrant, M. W., and M. Svendsen, 1993. Asianfood production in the 1990s: Irrigation invest-ment and management policy. Food Policy 18(1):13–32.

Rutherford, A. S. 1999. Meat and milk self-sufficiency in Asia: Forecast trends and impli-cations. Agricultural Economics 21(1): 21–39.

Ruttan, V. W., ed. 1994. Agriculture, environment,

and health: Sustainable development in the 21st cen-

tury. Minneapolis, Minn.: University of Min-nesota Press.

———. 1996. Population growth, environmentalchange, and technical innovation: Implicationsfor sustainable growth in agricultural produc-tion. In The impact of population growth on well-

being in developing countries, ed. D. A. Ahlburg,A. C. Kelley, and K. O. Oppenheimer. Berlin andNew York: Springer.

Ruttan, V. W., and Y. Hayami. 1984. Towards atheory of induced innovation. Journal of Devel-

opment Studies 20(4): 203–223.Sanchez, P. A., K. D. Shepherd, M. J. Soule, F. M.

Place, A. U. Mokwunye, R. J. Buresh, F. R. Kwe-siga, A. N. Izac, C. G. Ndiritu, and P. L. Woomer.1997. Soil fertility replenishment in Africa: Aninvestment in natural resource capital. In Replen-

ishing soil fertility in Africa, ed. R. J. Buresh and P.

REFERENCES 205

A. Sanchez. SSSA Special Publication. Madison,Wisc.: Soil Science Society of America (SSSA) andAmerican Society of Agronomy.

Sanderson, F. H., and R. Mehra. 1990. Trends andbaseline projections in global food trade. In Agri-

cultural protectionism in the industrialized world,

ed. F. H. Sanderson. Washington, D.C.:Resources for the Future.

Sayed, H. 1996. Investing in junior secondary educa-

tion in Indonesia: Rationale and public costs.

Indonesia Discussion Paper Series. Washington,D.C.: World Bank.

Scherr, S. J. 1999. Soil degradation: A threat to devel-

oping-country food security by 2020. 2020 Visionfor Food, Agriculture, and the EnvironmentDiscussion Paper 27. Washington, D.C.: Inter-national Food Policy Research Institute.

Scherr, S., and S. Yadav, 1996. Land degradation in

the developing world: Implications for food, agricul-

ture, and the environment to 2020. 2020 Vision forFood, Agriculture, and the Environment Dis-cussion Paper 14. Washington, D.C.: Interna-tional Food Policy Research Institute.

Scott, G. J., M. W. Rosegrant, and C. Ringler. 2000.Roots and tubers for the 21st century: Trends, pro-

jections, and policy options. 2020 Vision for Food,Agriculture, and the Environment DiscussionPaper 31. Washington, D.C.: International FoodPolicy Research Institute.

Seppala, P. 1997. Food marketing reconsidered: An

assessment of the liberalization of food marketing in

Sub-Saharan Africa. Research for Action No. 34.Helsinki, Finland: United Nations University,World Institute for Development EconomicsResearch (UNU/WIDER).

Simon, J. L. 1981. The ultimate resource. Princeton,N.J.: Princeton University Press.

Smil, V. 1998. Food, energy, and the environment:Implications for Asia’s rice agriculture. In Sus-

tainability of rice in the global food system, ed.N. G. Dowling et al. Los Baños, Philippines:Pacific Basin Study Center, International RiceResearch Institute.

———. 2000. Feeding the world: A challenge for the

twenty-first century. Cambridge, Mass.: MIT Press.Smith, L., and L. Haddad. 2000. Explaining child mal-

nutrition in developing countries: A cross-country

analysis. Research Report 111. Washington, D.C.:International Food Policy Research Institute.

Soedjana, T. D. 1999. Impact and implications of theeconomic crisis on livestock production, market-ing and consumption in Indonesia. In Proceedings

of the Workshop on the Implications of the Asian Eco-

nomic Crisis for the Livestock Industry, held in

Bangkok, Thailand, 6–9 July, 1999. Rome: Food andAgriculture Organization of the United Nations.

Spencer, D. S. C. 1994. Infrastructure and technology

constraints to agicultural development in the humid

and subhumid tropics of Africa. Environment andProduction Technology Division DiscussionPaper 3. Washington, D.C.: International FoodPolicy Research Institute.

Srinivasan, T. R. 1988. Population growth andfood: An assessment of issues, models, and pro-jections. In Population, food, and rural develop-

ment, ed. R. D. Lee, W. B. Arthur, A. C. Kelley,G. Rodgers, and T. N Srinivasan. Oxford, U.K.:Clarendon Press.

Steinfeld, H. 1998. Livestock production in the Asiaand Pacific region: Current status, issues andtrends. In World animal review. Rome: Food andAgriculture Organization of the United Nations.

Steinfeld, H., and A. Kamakawa. 1999. The indus-trialization of livestock production in the lightof the Asian economic crisis. In Proceedings of the

Workshop on the Implications of the Asian Economic

Crisis for the Livestock Industry, held in Bangkok,

Thailand, 6–9 July, 1999. Rome: Food and Agri-culture Organization of the United Nations.

Stoorvogel, J. J., E. M. A. Smaling, and B. H Jansen.1993. Calculating soil nutrient balances in Africaat different scales, I—Supra-national scale. Fer-

tilizer Research 35: 227–335.Tabor, R. T., W. Janssen, and H. Bruneau, ed. 1998.

Financing agricultural research: A sourcebook. TheHague, the Netherlands: International Servicefor National Agricultural Research.

Taimeh, A. 1999. Environmental threats to the lowrainfall areas and their causes. In Agricultural

growth, sustainable resource management and

poverty alleviation in the low rainfall areas of West

Asia and North Africa, ed. N. Chaherli. Proceed-ings of the international conference, held 2–6September, 1997, in Amman, Jordan. Feldafing,Germany: Deutsche Stiftung für internationaleEntwicklung (DSE).

Trigo, E. J. 1995. Agriculture, technological change,

and the environment in Latin America: A 2020 per-

spective. 2020 Vision for Food, Agriculture, and

206 REFERENCES

the Environment Discussion Paper 9. Wash-ington, D.C.: International Food PolicyResearch Institute.

UN (United Nations). 1998. World urbanization

prospects: The 1999 revision. New York: UnitedNations.

UNCTAD (United Nations Conference on Tradeand Development). Various years. Handbook of

international trade and development statistics.

Geneva.———. 1999. World population prospects: The 1998

revision. New York: United Nations.UNESCO (United Nations Educational, Scientific

and Cultural Organization ). 1998. UNESCO’sworld education indicators. <http://www.unesco.org/education/information/wer/ htmlENG/tablesmenu.htm.>.

———. 2000. UNESCOSTAT database. <http://unescostat.unesco.org/>.

USDA (United States Department of Agriculture).2000. Agricultural outlook. May 2000. Washing-ton, D.C.

USAID (United States Agency for InternationalDevelopment). 1988. Urbanization in the devel-

oping countries. Interim Report to Congress.Washington, D.C.: USAID.

Valdes, R. 1996. Surveillance of agricultural price and

trade policy in Latin America during major policy

reforms. World Bank Discussion Paper 349.Washington, D.C.: World Bank.

Valdés, A., and B. Schaeffer. 1995a. Surveillance of

agricultural price and trade policies: A handbook for

Argentina. World Bank Technical Paper No. 294.Washington, D.C.: World Bank.

———. 1995b. Surveillance of agricultural price and

trade policies: A handbook for Chile. World BankTechnical Paper No. 291. Washington, D.C.:World Bank.

———. 1995c. Surveillance of agricultural price and

trade policies: A handbook for Colombia. WorldBank Technical Paper No. 268. Washington,D.C.: World Bank.

———. 1995d. Surveillance of agricultural price and

trade policies: A handbook for Dominican Republic.

World Bank Technical Paper No. 267. Wash-ington, D.C.: World Bank.

Versteeg, M. N., P. Adegbola, and V. Koudokpon.1993. Investigación participatoria en laRepública de Benín: El método investigación-desarrollo (con un ejemplo de caso sobre

mucuna) In Gorras y sombreros: Caminos hacia la

colaboración entre técnicos y campesinos, ed. D.Buckles. Memoria del Taller Sobre Mos méto-dos participativos de investigación y extensiónaplicados a las tecnologías Basadas en AbonosVerdes. Mexico, DF: Centro Internacional deMejorimiento de Maiz y Trigo (CIMMYT).

Wang, Q., F. Fuller, D. Hayes, and C. Halbrendt.1998. Chinese consumer demand for animalproducts and implications for U.S. pork andpoultry exports. Journal of Agricultural and

Applied Economics 30 (July): 127–140.Ward, R.W., and C. Lambert. 1993. Generic pro-

motion of beef: Measuring the impact of theU.S. beef checkoff. Journal of Agricultural Eco-

nomics (U.K.) 44(3): 456–465.WHO (World Health Organization). 1997. WHO

global database on child growth and malnutrition.

Programme of Nutrition. WHO DocumentWHO/NUT/97.4. Geneva: World HealthOrganization.

Wolter, H.W., and C. M. Burt. 1997. Concepts ofmodernization. Modernization of irrigation

schemes: Past experiences and future options. Pro-ceedings of the expert consultation, Bangkok,26–29 November, 1996. Regional Office for Asiaand the Pacific, Land and Water DevelopmentDivision, Food and Agriculture Organization ofthe United Nations (FAO), Bangkok, Thailand.

World Bank. 1997. World development indicatorson CD-Rom. World Bank, Washington, D.C.

———. 1998. Commodity price outlook. August 1998.Washington D.C.

———. 2000a. Can Africa claim the 21st century?

Washington, D.C.: World Bank.———. 2000b. The 2000 world development indi-

cators on CD-ROM. World Bank, Washington,D.C.

———. Various years. World development report.

New York: Oxford University Press for theWorld Bank.

WRI (World Resources Institute). 1996. World

resources 1996–97. New York: Oxford UniversityPress.

WSSCC (Water Supply and Sanitation Collabora-tive Council). 2000. Vision 21: A shared vision for

hygiene, sanitation and water supply and a frame-

work for action. Water Supply and SanitationCollaborative Council. Geneva, Switzerland:World Health Organization.

Documents

Global Food Projections to 2020 - International Food Policy