ham. :2
EEE

a:

.5.
‘ 111.
:23: :3
. is).

, a.
«4"

. .x x749
as}... It

53.1....
35.31..
1.. l ,

\slilv:

{31. ..
.c..~v\t I
2:... .1121 3.. NH:
. ~ 21.1..tal
v17; ‘1.
s

v)»:
.-

 

1007.

 

LIBRARY
Michigan State
Universlty

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

’Jutl 62004

 
 

 

 

 

”9‘2 {)3 2%
9119235 9039

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/ClRC/DateDue.p65~p.15

ECONOMIC ASSESSMENT OF BEAN RESEARCH
IN THE DOMINICAN REPUBLIC

By

David Len Mather

A THESIS
Submitted to
Michigan State University
in partial fiJlfillment of the requirements
for the degree of
MASTER OF SCIENCE

Department of Agricultural Economics

2001

ABSTRACT

ECONOMIC ASSESSMENT OF BEAN RESEARCH
IN THE DOMINICAN REPUBLIC

By

David Len Mather

This study estimates the economic impact of two technologies developed by the
USAID Bean/Cowpea CRSP in the Dominican Republic: PC-SO, an improved red bean
variety, and the fallow period in the San Juan Valley (SJV). Subscctor analysis is used to
document the policy and institutional factors which have led to widespread adoption of ’
PC-SO, and assesses the future competitiveness of domestic production.

While bean production in the SJV is financially profitable, it is unviable under
economic valuation. While both the financial and economic ex post rate of return to PC-
50 adoption in the DR during the period 1984-2002 indicate that B/C CRSP research
attributed to screening, promotion, and multiplication of P050 was profitable. Although
bean is uneconomic in the SJV, PC-SO adoption decreased the magnitude of economic loss
that would have occurred without PC-SO, thus generating a positive incremental benefit.
However, the welfare benefits from PC-SO adopters have gone almost exclusively to
irrigated producers, while consumers have not benefitted from this technology. When the
SJV fallow period is considered an output of the CRSP, joint analysis of PC-SO and the
fallow period lead to the result that the CRSP research in the DR is not profitable in

economic terms.

Cepyright by
DAVID LEN MATHER
2001

ACKNOWLEDGMENTS

I express my sincere gratitude to Dr. Richard Bernsten, my major professor, for
the opportunity to work on the Bean/Cowpea CRSP, and for his attentive, supportive, and
critically constructive comments and suggestions during the thesis process. I also thank
the other members of my committee, Drs. Eric Crawford and Russell Freed, for their
timely and constructive comments. I especially thank Dr. Crawford and Dr. James
Oehmke for their help with the benefit-cost analysis.

This research would have been virtually impossible without the logistical and
research support of Dr. Eladio Amaud Santana, the DR CRSP PI at the Centro de
Investigacion Agricola del Suroeste (CIAS, Arroyo Loro). Special thanks to my team of
enumerators: Segundo Nova and Ana Mateo. Their long hours of work were essential to
the completion of this thesis.

I also thank the USAID/Bean Cowpea CRSP and the Department of Agricultural
Economics at MSU for financial support during my M.Sc. program, and logistical support
during my field work.

In the Dominican Republic, useful suggestions and comments were made by Dr.
Cesar Paniagua (PRODAS), Dr. Jesus de los Santos (JAD), and Dr. Ruben Nunez (IICA).
I thank Wagner Mendez for help in collecting secondary data in Santo Domingo, and for
his hospitality.

I cherish the time I have spent with so many fi'iends in the MSU community.

Special thanks for support and fiiendship to Brady Deaton, Matt Schaefer, Jim Stems,

iv

Allen Kannapel at EMMSU, and the soccer team.
I am especially gratefiil for the support of my family. I would not be here if it were
- not for the sacrifice, love, and inspiration provided by my mother, father, and brother.

I dedicate this work and my life to Jesus Christ, my Lord and Saviour.

TABLE OF CONTENTS

 

 

.LIST OF TABLES ................................................... xii
LIST OF FIGURES ................................................... xv
CHAPTER ONE: INTRODUCTION ...................................... 1

1.1 Problem Statement ........................................... 1
1.2 Research Objectives .......................................... 3
1.3 Hypotheses ................................................. 4
1.4 Organization of Thesis ........................................ 5
CHAPTER 2: LITERATURE REVIEW AND METHODS ...................... 7
2.1 Introduction ................................................ 7
2.2 Literature Review ............................................ 7
2.2.1 Benefit-Cost Analysis .................................. 7

2.2.1.1 Technical Innovations and Resource Productivig ...... 7

2.2.1.2 Benefit-Cost Methodolggi_e_§ ...................... 8

2.2.1.3 Appropriate Benefit-Cost Methodology ............ 13

2.2.1.4 Relevance of Benefit-Cost Literature .............. 15

2.2.2 Institutional Analysis .................................. 16

2.2.3 On the Rapid Appraisal/Subscctor Method of Analysis ........ 17

2.3 Instruments and Methods of Analysis ............................ 18
2.3.1 Rapid Appraisal ..................................... 18

2.3.2 Record-Keeping data ................................. 19

2.3.3 Farm-level Adoption and Production Survey ................ 20

2.3.3.1 Survey Obiectives ............................. 20

2.3.3.2 Population and Sampling Method ................. 20

2.3.4 Varietal Trials ....................................... 23

2.3.5 Related Research .................................... 23

2.3.6 Questionnaire design .................................. 24

2.4 Chapter Summary ........................................... 24
CHAPTER THREE: THE DOMINICAN REPUBLIC ........................ 26
3.1 Physical and Socioeconomic Characteristics ....................... 26
3.1.1 Geography and Agroecology ........................... 26

3.1.2 Population Grth ................................... 26

3.2 The Economy .............................................. 27
3.2.1 Macroeconomy ...................................... 27

3.2.2 Income Distribution .................................. 29

3.2.3 Poverty ............................................ 29

3.2.4 Structure of the Economy .............................. 29

3.3 Agricultural Sector .......................................... 31

3.3.1 Agricultural Price and Trade Policy ....................... 31
3.3.2 Land Use and Ownership .............................. 33
3.3.3 Domestic Food and Cash Crop Production in 1997 ........... 34
3.3.4 Food Security ...................................... 35
3.4 Conclusion ................................................ 36

CHAPTER FOUR. AN OVERVIEW OF THE DOMINICAN BEAN SUBSECTOR. 37

 

 

 

4.1 Introduction ............................................... 37
4.2 Demand Analysis ............................................ 37
4.2.1 Beans in the Dominican Diet ............................ 37

4.2.2 Consumer Preferences ................................ 38

4.2.3 Domestic Consumption ................................ 39

4.2.4 Imports ............................................ 40

4.3 Production Analysis ......................................... 41
4.3.1 Bean Classes and Varieties ............................. 41

4.3.2 Bean Production Location and Seasons ................... 42

4.3.3 Recent Changes in Red Bean Area and Yields ............... 43

4.3.4 Recent Changes in the Seasonality of Red Bean Production . . . . 45

4.3.5 Land Tenure and Farm Size ............................ 47

4.3.6 Constraints to Increasing Productivity ..................... 48

4.3.6.1 Bean Disease Prevalence, Freguengy, and Severity . . . . 48

4.3.6.2 Other Production Constraipt_s ..... . ............... 49

4.3.7 Agricultural Organization Membership .................... 49

4.4 Bean Price Analysis . . .i ....................................... 50
4.4.1 Bean Price Seasonality ................................ 50

4.4.2 Trends in Real Prices ............................... 53

4. 4. 3 Trends in Relative Prices ............................... 54

4. 5 Government Policies Affecting the Bean Subscctor .................. 55
4.5.1 Introduction ........................................ 55

4.5.1.1 Agricultural Sector ............................ 55

4.5.1.2 Import Restrictions ............................ 56

4.5.1.3 GATT Obligations ............................ 56

4.5.2 INESPRE .......................................... 58

4.5.2.1 Institutional Background of INESPRE ............ 58

4.5.2.2 INESPRE’s domestic bean purchases .............. 60

4.5.2.3 INESPRE’s Imports ........................... 63

4.5.2.4 INESPRE Direct Sales to Consumers ............ 64

4.5.2.5 The Future of INESPRE ....................... 66

4.5.3 SEA Seed Multiplication and Distribution Program . . . 66

4.5.4 Banco Agricola ...................................... 68

4.5.5 PROSEMA ......................................... 69

4.5.6 INDRHI ........................................... 70

vii

4.5.7 CVMA Input Sales ................................... 70

 

4.6 The Bean Marketing System ................................... 70.
4.6.1 Marketing Channels .............. ' .................... 70

4.6.2 Farm-level Bean Storage and Sales ....................... 73

4.6.3 Alternative Marketing Institutions ........................ 74

4.7 Research and Extension to Improve Bean Varieties and Management

Practices .................................................. 75
4.7.1 Bean Breeding and Foundation Seed Production ............. 75

4.8 Chapter Summary ...................................... . . . . . 76
CHAPTER FIVE: BEAN PRODUCTION IN THE SAN JUAN VALLEY ......... 79
5.1 Introduction ............................................... 79
5.2 Data ..................................................... 81
5.3 Human Capital ............................................. 82
5.3.1 Introduction ........................................ 82

5.3.2 Producer Associations . . . . . . . ......................... 83

5.3.3 Socioeconomic Characteristics of Producers ................ 84

5.3.4 Extension Advice Received ............................. 85

5.3.5 CRSP Extension of Technical Knowledge ................. 88

5.4 Bean Cropping Patterns and Land Use ........................... 89
5.4.1 Farm Size .......................................... 89

5.4.2 Land Tenure ........................................ 90

5.4.3 Irrigation .......................................... 93

5.4.4 Cropping Systems .................................... 95

5.5 Land Preparation and Planting Dates ............................. 96
5.5.1 Land Preparation ............................... ' ..... 96

5.5.2 Fallow Period and Planting Dates ........................ 97

5.5.3 Planting Equipment .................................. 100

5.5.4 Farmer Opinions on Planting Date Period ................. 100

5.6 Use of Improved Bean Varieties ............................... 102
5.6.1 Improved Bean Varieties and Yields ..................... 102

5.6.1.1 History of the Development of PC-SO ............. 102

5.6.1.2 PC-SO .................................... 103

5.6.1.3 JB-l78 .................................... 105

5.6.1.4 Other Newly-released Varieties ................. 107

5.6.2 Distribution Channels of Bean Seeds ..................... 108

5.6.3 Seed Cost ......................................... 108

5.6.4 Seed Density ....................................... 108

5.7 Chemical Inputs ........................................... 109
5.7.1 Fertilizer .......................................... 109

5.7.2 Insecticide ........................................ 110

5.7.3 Fungicide ......................................... 112

5.7.4 Herbicides ........................................ 1 13

viii

5.7.5 Total Purchased Input Costs ........................... 113

5.7.5.1 Purchased Input Cost Shares ................... 113

5.8 Machine and Labor use ...................................... 114
5.8.1 Introduction ....................................... 114

5.8.2 Land Preparation .................................. 118

5.8.3 Planting .......................................... 119

5.8.4 Application of Insecticide ............................. 119

5.8.5 Manual Weeding, Irrigation, and Harvest ................. 119

5.9 Other Costs ............................................... 120
5.9.1 Credit ............................................ 120

5.9.2 Irrigation Charge ................................... 123

5.9.3 Internal Transport ................................... 123

5.10 Total Costs of Production ................................... 124
5.10.1 Cost Shares ...................................... 124

5.10.2 Decreasing the Costs of Production .................... 124
5.10.2.1 Input Application Rates ....................... 126

5.10.2.2 Manual Weeding vs. Herbicides ................. 126

5.10.3 Government Policy and the Costs of Production ........... 127
5.10.3.1 Input Prices ............................... 127

5.10.3.2 Wage Rates in the San Juan Valley .............. 128

5.10.3.3 Financing, Fertilizer Prices, and Implications ...... 128

5.11 Production Problems ....................................... 129
5.12 Marketing ............................................... 130
5.12.1 Marketing Channels ................................ 130

5.12.2 SEA Seed Program ................................. 131

5.12.3 INESPRE ........................................ 131

5.12.4 Private Traders . . .g ................................. 132

5.12.5 Average Farmer Prices by Farm Size .................... 133

5.12.6 Implications of SBA and INESPRE Bean Purchases ........ 133

5.13 Profitability of Bean Production .............................. 135
5.13.1 Financial Profitability ............................... 135

5.13.2 Economic Profitability .............................. 136

5.13.3 Alternative Crops in the SJV .......................... 138

5.14 Sensitivity Analysis ........................................ 138
5.14.1 Introduction ...................................... 138

5.14.2 Changes in Bean Yields and Prices ..................... 139

5.14.3 Sensitivity to Changes in Input Costs ................... 142

5.14.4 Farmers’ Sensitivity to Price Declines ................... 144

5.16 Alternative Cost of Production Figures ......................... 146
5.16.1 San Juan Valley ................................... 146

5.16.2 Rainfed Bean Production ............................ 146

‘ 5.17 Chapter Summary ......................................... 147

ix

CHAPTER SIX: ECONOMIC IMPACT OF BEAN RESEARCH IN THE D.R ..... 148

 

 

6.1 Introduction .............................................. 148
6.2 Assumptions of the PC-SO Analysis ............................. 150
6.2.1 Farm-level Benefits from the Adoption of P050 ........... 150
6.2.2 Incremental Farm-level Production Costs ................. 151
6.2.3 Incremental Farm-level Input Use ...................... 151
6.2.4 Economic Surplus Approach ........................... 152
6.3 Financial Benefit Stream ..................................... 155
6.3.1 Adoption Rates and Time-line .......................... 155
6.3.1.1 Introduction ................................ 155
6.3.1.2 San Juan Valley ............................. 156
6.3.1.3 Qibao and Higpey Valleys ...................... 159
. 6.3.1.4 Other Regions .............................. 161
6.3.2 Red Bean Area by Region ............................. 161
6.3.3 Red Bean Yield by Region ............................ 161
6.3.3.1 Yield Difl‘erentJiafi ........................... 161
6.3.3.2 Incremental Yields ........................... 162
6.3.3.3 Incremental Production ....................... 163
6.3.4 Red Bean Financial Prices ............................. 164
6.3.5 CPI, Exchange Rate, and Discount Rate .................. 164
6.4 Financial Costs Stream ‘ ...................................... 165
6.4.1 CRSP Research Costs ................................ 165
6.4.2 SEA Costs ........................................ 166
6.5 Financial Rate of Return to PC-SO .............................. 166
6.6 Economic Benefit Stream ......................... 167
6.6.1 Economic Import Parity Prices ......................... 167
6.6.2 Market Price Premium for Domestic Red Bean ............. 168
6.7.3 Seed Subsidy ...................................... 171
6.7.4 Irrigation and Land Preparation Subsidies ................. 173
6.8 Economic Analysis of PC-50 .................................. 173
6.8.1 Economic Rate of Return ............................. 173
6.8.2 Sensitivity Analysis .................................. 174
6.8.3 Distribution of Benefits ............................... 175
6.8.3.1 Introduction ................................ 175
6.8.3.2 Consumers ................................. 176
6.8.3.3 Producers .................................. 177
6.9 Financial and Economic Analysis of The Fallow Period .............. 179
6.9.1 Introduction ....................................... 179
6.9.2 Intuitive Results .................................... 185
6.9.3 Aggregate Financial and Economic Benefits of the Fallow Period
................................................ 186
6.10 Financial and Economic Analysis of PC-50 and the Fallow Period ..... 188

6.10.1 Combining the Two Technologies in Benefit-Cost Analysis . . . 188

6.10.2 Research Costs of the Fallow Period .................... 189

6.10.3 Financial Return to PC-SO and the Fallow Period .......... 189

6.10.4 Economic Return to PC-SO and the Fallow Period ......... 190

6.10.5 Paradoxical Result ................................. 190

6.10.6 Distribution of Benefits .............................. 191

6.11 Chapter Summary ......................................... 191

6.11.1 Aggregate Results .................................. 191

6.11.2 Distributional Results ............................... 192

CHAPTER SEVEN: SUMMARY OF FINDINGS .......................... 194
7.1 Introduction .............................................. 194

7.2 Summary of Findings ........................................ 194

7.2.1 Subscctor Analysis .................................. 194

7.2.2 Profitability Analysis ................................. 195

7.2.3 Benefit-Cost Analysis of PC-SO ........................ 195

7.2.4 Benefit-Cost Analysis of PC-SO and the Fallow Period ....... 196

7.2.5 Distribution of Benefits from Technological Change ......... 197

7.3 Prospects for Future Economic Impact of the CRSP in the DR ........ 197

7.3.1 General Outlook .................................... 197

7.3.2 Research for Rainfed Producers ........................ 198

7.3.3 Extension of Technologies ............................ 199

7.4 Limitations of This Study .................................... 200

7.5 Recommended SOcioeconomic Research ......................... 201
APPENDIX A: Validity of the Sample Frame .............................. 202
APPENDIX B: Percentage of Regional Bean Area by Technology Level ......... 204
APPENDD( C: SEA Seed Sales . , ....................................... 205
REFERENCES ..................................................... 207

LIST OF TABLES

_ Table 3.1 Major Food and Export Crops, DR 1997 .......................... 34
Table 4.1 Per Capita Bean Production, Consumption, and Imports in the DR ....... 41
Table 4.2 Regional Area, Production, and Yield of Red Beans in the DR .......... 44

Table 4.3 Prevalence, Frequency, and Severity of Bean Diseases in the DR 1992 . . . . 48

Table 5.1 Socioeconomic Characteristics of Producers by Bean Area, SJV, 1998 . . . . 85
Table 5.2 Farmers’ Sources of Technical Assistance, SJV, 1998 ................. 86
Table 5.3 Sponsors of Bean Production Courses and Field Days, SJV, 1998 ........ 88
Table 5.4 Characteristics of Sampled Bean Producers by Bean Area, SJV, 1998 ..... 90
Table 5.5 Land Tenure of Sampled Bean Producers, SJV, 1998 ................. 91
Table 5.6 Land Tenure of Sampled Bean Producers by Bean Area, SJV, 1998 ...... 92
Table 5.7 Summer Cropping by Bean Farmers by Bean Area, SJV, 1998 .......... 95
Table 5.8 Farmers’ Recommended Starting Dates for the Winter Bean Planting Period,
SJV, 1998 ................................................. 101
Table 5.9 Mean Farmer Yields of PC-50 in the San Juan Valley, 1993-97 ........ 104
Table 5.10 Average Quantity and Cost of Purchased Input Use Per Hectare by Type of
Farmer, Winter 1997-98, SJV ................................. 111
Table 5.11 Equipment and Labor Use per Hectare, SJV, 1998, (SUS/ha) ......... 116
Table 5.12 Hours per Hectare and Wage Rates for Bean Production Operations by Farm
Size, San Juan Valley, DR 1998 ............................... 117
Table 5.13 Credit for Bean Production, SJV, 1997-98 ....................... 121

Table 5.14 Financial Costs of Production, Red Beans, Winter Season, 1997-98, SJV,
DR .................................................... 125

xii

Table 5.15 Principal Bean Production Problems, SJV, 1997-98 ................. 130

Table 5.16 Financial and Economic Profitability of Bean Production in the SJV, DR
1997-98 .................................................. 137

Table 5.17 Financial and Economic Profitability of Selected Annual Crops, DR
1998 .................................................... 138

Table 5.18 Sensitivity Analysis of Net Retums/ha, Red Beans, SJV 1997-98: Average
Farmer ................................................... 141

Table 5.19 Sensitivity Analysis of Net Returns/ha, Red Beans, SJV 1997-98: Average
Small Farmer .............................................. 141

Table 5.20 Sensitivity Analysis of Net Retums/ha, Red Beans, SJV 1997-98: Average
Large Farmer .............................................. 142

Table 5.21 Sensitivity Analysis of Enterprise Profit/ha, Red Beans, SJV 1997-98:
Average Farmer . . . . ........................................ 143

Table 5.22 Farmer Reactions to Declines in Hypothetical Farmgate Bean Prices, SJV,
1998 .................................................... 145

Table 6.1 Economic Surplus Equation Definitions and Formulas ............... 153

Table 6.2 Producer Adoption of Major Bean Varieties and Types, SJV, 1997-98 . . . 158

Table 6.3 Real CRSP Research Costs .................................. 166
Table 6.4 Economic Import Parity Price, Red Beans, DR .................... 168
Table 6.5 Real Economic Cost Components ............................. 170
Table 6.6 Subsidy Calculations ........................................ 171

Table 6.7 Sensitivity Analysis of the Economic ROR to PC-SO Adoption in the DR
1984-02 .................................................. 174

Table 6.8 The SJV Fallow Period in With- and Without-Fallow Scenarios,
1990-02 .................................................. 179

Table 6.9 Financial and Economic Returns to Bean Production, SJV: With Fallow . . 182

xiii

Table 6.10 Financial and Economic Returns to Bean Production, SJV: Without

Fallow ................................................. 182

, Table 6.11 Financial and Economic Returns to Pigeon Pea Production, SJV ....... 183

Table 6.12 Financial and Economic Returns to Maize Production, SJV ........... 183
Table 6.13 With— and Without-F allow Scenarios Assuming Switch to Alternative Crops,

1990-02 ................................................. 184

Table 6.14 Financial Analysis of the SJV Fallow Period ...................... 186

Table 6.15 Economic Analysis of the SJV Fallow Period ..................... 186

xiv

LIST OF FIGURES

, Figure 4.1 DR Red Bean Supply 1981-98 .................................. 46
Figure 4.2 DR Red Bean Price Index ..................................... 51
Figure 4.3 Red Bean Retail Price Index .................................... 52
Figure 4.4 DR Red Bean Prices .................................... V ..... 54
Figure 4.5 Relative Bean Prices ......................................... 55
Figure 5.1 Purchased Input Cost Shares .................................. 114
Figure 6.1 PC-SO Adoption: San Juan Valley .............................. 158
Figure 6.2 PC-SO Adoption: Higuey/Cibao & Other Regions .................. 160

CHAPTER ONE: INTRODUCTION
1.1 Problem Statement

Beans and rice are two of the most important staple foods in the Dominican
Republic. Rice is the principle source of calories, and dry beans are the principle source
of vegetable protein for the majority of Dominicans, especially the poor. Together, the
traditional staples - rice and beans -- account for 35% of calories and 27% of protein in
the Dominican diet (Rogers, 1987). In addition to being an important staple food crop for
consumers, beans are also an important cash crop for both commercial and subsistence
farmers throughout the country. The financial importance of this crop to commercial
farmers, combined with high farmer participation in farm commodity organizations and
heavy government involvement in the direct marketing of beans in the DR since the mid-
1970s, makes beans one of the most political crops in the country. In addition to the
importance of beans to both consumers (as a staple food) and to producers (as a cash
crop), consumption of beans is an integral part of the country’s culture. The most
common midday meal for the majority of Dominicans is arroz can habichuelas (rice and
beans).

Compared to many developing countries, Dominican bean yields are relatively
high, averaging 765 kg/ha over the last twenty years (SEA, 1998). However, during this
period, yields have ranged from 565 to 1,021 kg/ha due to erratic growing conditions.
Agronomic constraints to increasing yields include heat stress and the widespread threat of
yield-reducing diseases, including bean golden mosaic virus (BGMV), common blight,

rust, and web blight.

In light of the nutritional, economic, political, and cultural importance of beans,
and the realization that the country’s producers were not keeping pace with consumer
demand, in 1981, the Dominican Ministry of Agriculture (SEA) joined together with the
USAID-funded Bean Cowpea Collaborative Research Support Project (CRSP) to
undertake research designed to increase bean production and stabilize bean yields in the
DR Major research objectives of this initiative have included developing new higher-
yielding, diseasetolerant varieties and identifying improved management practices to
reduce disease incidence.

This collaboration on varietal improvement led to the release of “PC-50” in 1989,
a red melted variety chosen from among 250 selections of the widely grown local land
race “Pompador Checa.” P050 is high-yielding, has uniform seed size and maturity,
good cooking attributes, and is tolerant to rust and to both common and web blight. Since
1990, the Seed Department of SEA (the Dominican Ministry of Agriculture) has
contracted farmers to grow “seed of good quality” for subsequent sale to commercial
producers.

In addition to developing improved seed, the project identified “fallowing” as a
strategy for reducing the incidence of BGMV in the San Juan Valley (SJV), the leading
bean-producing area in the country. Based on this research, in 1991, the project
recommended that SJV farmers incorporate a “fallow” period for beans into their cropping
pattern fi'om August to November, follbwed by a planting period fiom 15 November to 10
January. The Southwest regional government made this fallow and planting period

recommendation into law for the 1991/92 season. As a result of subsequent research on

optimal planting dates in the San Juan Valley, in 1993, the Southwest regional government
adjusted the mandated planting period to 5 November to 15 December. Anecdotal
evidence suggests that the fallowing and optimal planting strategy have contributed to a
much lower level of BGMV than seen in the 1989-91 period, and has reduced farmers’
dependence (and expenditures) on insecticides used to control the white fly, the vector of
BGMV.

The widespread adoption of both PC-SO and the fallow period in the SJV suggests
that the Bean/Cowpea CRSP research program has had a major impact on bean
production in the DR. While previous studies have documented the widespread adoption
of this technology package (I-Ieikes 1993, Arnaud 1997), the role of institutional
arrangements and the policy environment that contributed to its widespread adoption has
not been documented, nor has an analysis of the economic rate of return to CRSP bean
research in the DR been conducted.

1.2 Research Objectives

The general objective of this study is to assess the impact of PC-SO and the SJV
fallow period and to describe the policy factors that facilitated PC-SO adoption. The
policy fi'amework that delineates the opportunities and constraints facing all bean
subsector agents also defines the distribution of the impact benefits among them. This
study thus documents the policy fiamework as an integral part of the impact assessment.

The specific objectives of this research are to:

1) Assess the financial and economic costs of bean production in the SJV;

2) Determine the incremental benefits associated with farmer adoption of PC-SO; the

3)

4)

5)

6)

adoption of the SJV fallow period; and the economic rate of return (ROR) to
CRSP investments;

Assess the sensitivity of this ROR to key assumptions of bean area, yield, and
fann-level bean price;

Identify the policy and institutional factors that account for widespread adoption of
PC-50 and associated management practices in the SJV and other principal
production areas, including: the roles of the CRSP, the govermnent, and farmer
associations in promoting seed variety development, production, distribution, and
adoption; CRSP field days and other extension efl‘orts; and government price
supports;

Describe the bean subsector, focusing on the seed and grain price negotiation
process, input/output marketing channels, and price differentials between domestic
and imported beans and between beans grown for seed and consumption.
Document farm-level constraints to increased production and profitability; farmers’
perceptions of the advantages/disadvantages of P050 and the planting dates; and

farmer responses to various hypothetical bean price scenarios.

1.3 Hypotheses

This study tests the hypothesis that investments in agricultural research have

positive returns. Specifically, it will determine if the economic rate of return (ROR) to

investment in bean research in the DR is greater than the opportunity cost of capital.

Many ROR studies in Latin America report a high rate of return to agricultural research

(Daniels er al., 1992), and previous studies indicate that project-developed technologies

have been widely adopted, have improved bean yields, and have brought BGMV under
control in the San Juan Valley, the primary bean grain and seed-producing region in the
DR. While CRSP documents report that CRSP bean research in the Dominican Republic
has been a resounding success, this is not clear from an economic perspective. This
guarded pessimism is due to the high production costs of domestic beans relative to
imports, and the high domestic retail bean prices (relative to other countries and to protein
substitutes) that result from import restrictions.

Subscctor analysis will describe how policy and institutional arrangements facilitate
the adoption of agricultural technologies and thus the economic rate of return. It will also
describe the distribution of the benefits of technology adoption, which are hypothesized to
be significant for larger, resource-abundant farmers, small for smaller, resource-poor
farmers, and potentially negligible for consumers. Because domestic production rarely
meets demand, and because imports are restricted based upon annual projections of
domestic production, real domestic retail bean prices have not fallen, meaning that benefits
from increased yields are captured almost entirely by producers.

1.4 Organization of Thesis

Thisthesis is divided into seven chapters. Chapter 2 briefly reviews the rate of
return literature and presents the research design and methodology. Chapter 3 gives an
overview of macroeconomic and demographic conditions in the DR including a general
overview of the agricultural sector. Chapter 4 presents an overview of the Dominican
bean subsector. Chapter 5 presents a detailed view of bean production in the San Juan

Valley. Chapter 6 reports results of the rate of return analysis and the distribution of the

project impact, given the institutional framework described in the subsector analysis.

Chapter 7 summarizes the findings of the study and draws policy implications.

CHAPTER TWO: LITERATURE REVIEW AND METHODS
2.1 Introduction

This chapter first presents literature relevant to determining the economic impact
of bean research in the DR and documenting the institutions and policies which account
for the magnitude and distribution of this impact. The choice of benefit-cost methodology
is then discussed, followed by a discussion of the data collection methodologies and
instruments.

2.2 Literature Review
2.2.1 Benefit-Cost Analysis
2.2.1.1 WI!

The challenge of increasing agricultural production requires continuous investment
in the development and extension of productive agricultural technology. Since the early
1960s, developing countries, assisted by foreign donors, have invested resources to
strengthen their agricultural research systems to generate technologies to increase
agricultural productivity. Agricultural economists have long supported this strategy,
arguing that technical innovations in agricultural production drive the development of the
agricultural sector, whose growth is viewed as a precondition to the sustainable and
broad-based growth of the general economy (Mellor, 1966). Because such investment is
costly and competes for scare public and donor resources, socioeconomic impact studies
of such investments can help policymakers to assess the value of these past investments
and prioritize alternative firture research investments to enable improved allocation of

scarce resources to meet their economic, political, and social interests. Numerous studies

that have assessed the impact of research on agricultural productivity growth, particularly
in the US, Latin America, and Asia, indicate a high rate of return to investment in
agricultural research (Daniels et a1, 1992).

2.2.1.2 Benefit-Egg; Methpgplogies

Most published impact assessments attempt to quantify benefits as economic or
financial returns to expenditures for agricultural research. Three extensive reviews of such
impact methodologies include Schuh and Tollini (1979), Norton and Davis (1981), and
Echeverria (1990). Impact methodologies can be grouped into two main categories: ex
post and ex ante approaches. The ex post approach is used to determine the impact of
past research expenditures while the ex ante approach is typically used to estimate
expected rates of return for proposed projects.

Echeverria (1990) groups quantitative benefit-cost methods into two general
classes, the economic surplus approach (consumer-producer surplus, benefit-cost, and
index number methods) and the econometric approach (production, profit and supply
functions and their derivations). Economic surplus approaches “estimate returns on
investment (an average rate of return) by measuring the change in consumer and producer
surplus fiom a shift to the right in the supply curve due to technological change (ibrd
1990).” Econometric approaches “treat research as a variable and allow a marginal rate of
return on investment to be calculated (ibid, 1990).” The average rate of return over some
time period is considered appropriate for ex post analysis as an indicator of the financial
and economic profitability of past investments. The marginal rate of return is considered

more appropriate for situations in which research expenditures related to the technology in

question are ongoing.

Several different indicators or measures of the return of the investment may be
employed. The internal rate of return (IRR) is once such measure, defined as follows:
The IRR is the interest rate that equates the net present value of cash flows to zero, as

calculated by the equation:

" Br-Cr
=0
131(1-1-1')’

 

where “B,” and “C,” are the values of the benefit and cost streams in each time period

fiom t = 1 to n, and “r" is the interest rate that solves the equality (the IRR).

The IRR is the “rate of return on capital outstanding per period while it is invested
in the project (Gittinger, 1982);” in other words, the return to the money invested in the
project. For example, an IRR of 18 percent means that project returns are large enough to
cover all operating costs, pay back the principal on the capital invested in the project, and
return an average 18 percent annually for the use of the money in the meantime. A project
with an IRR that exceeds the average real market interest rate during the project life is
deemed “profitable.”

Economists make further distinctions when measuring profitability, primarily
between “financial” and “economic” valuation. The financial ROR is an IRR where
benefits and costs are valued in terms of market prices and domestic currency, unadjusted
for “distortions” in the prices of inputs or outputs in question. Thus, financial values are

the actual and unadjusted prices faced by producers and consumers over the time period in

question. The economic ROR is an IRR where the benefits and costs are valued in terms
of their economic opportunity costs. Economic opportunity costs are “financial” prices
adjusted to reflect the “economic values” of the inputs and outputs in question. For
example, the financial farmgate price of beans in the DR is the actual reported price per
unit at any given point in time. On the other hand, the economic farmgate price of a
metric ton of Dominican beans is the sum of the total cost (of production, marketing, and
transport) necessary to bring a metric ton of imported beans (assumed to be a perfect
substitute for the domestic product) to the same physical location in the DR at which the
farmgate 'price was recorded. This study will present rate of return indicators reflecting
both financial and economic valuation of benefits and costs.

Many of the pioneering efi’orts to measure the impact of agricultural technology
adoption assumed closed-economy scenarios and reported aggregate benefits to society or
benefits disaggregated between producers and consumers of the commodity in question.
More recent methodological developments have focused on incorporating distributional
issues, social costs, and the efi’ects of government trade and price policy regimes into the
impact analysis.

The consumer-producer surplus approach generally assumes that producers who
adopt new technologies enjoy higher productivity (lower production costs per unit of
output), while consumers enjoy the lower relative product prices that result from an
increased product supply. An advantage of this approach is that it is relatively flexible and
can be modified to account for the effect of trade and price policy on the distribution of

benefits between consumers and producers. This is only feasible when the technology in

10

question can be associated with a particular research program. The empirical information
required is knowledge of how much the technical change shifted the supply curve and
estimates of the elasticities of supply and demand for the commodity.

Consumer-producer surplus approaches need not limit impact analysis to the
aggregate benefit to society or to consumer and producers in the aggregate. In fact, the
efi’ect of technological change on the structure of input and output markets for the
commodity afl'ected by technological change can influence the distribution of income
within a society and the relative productivity of factors of production. For example, when
technology adoption leads to increased production of a staple commodity (and assuming
that increased production leads to lower retail prices of that commodity), low-income
groups benefit relatively more than high income groups fi'om lower output prices.
However, lower output prices may reduce the profitability of late or non-adopters of new
production technologies.

Ifthe technology package requires such complimentary inputs such as irrigation or
fertilizer, small rainfed producers lacking these resources would either not adopt the
technology or not enjoy the full benefits of its yield-augrnenting potential. Therefore,
lower product prices could therefore have a detrimental efi‘ect on producers not able to
adopt said cost-dirninishing or yield-augrnenting production technologies, assuming said
producers are net sellers of the commodity. With that in mind, Scobie and Posada (197 8)
note that “fi'equently it is the well-being of only the mral poor (both small farmer and
landless worker) that is the focus of attention. The presence of large concentrations of

urban poor who are potential beneficiaries of expanded production of basic food stufi’s is

11

sometimes neglected when castigating (new technologies)” Their influential study of the
impact of improved rice varieties in Colombia demonstrated that the benefits of lower
food prices for the urban poor in Colombia far outweighed the costs to rural rice
producers in the form of lower farmgate prices. Their work thus demonstrated the value
of investigating the distribution of both benefits and costs - not just the distribution
between consumers and producers, but the distribution between subgroups within
consumer and producers in the aggregate.

In their assessment of the impact of mechanical tomato harvester adoption in the
US, Schmitz and Seckler (1970) argued that an increase in productivity due to technical
change can lead to a release of resources fi'om the sector, depending on the demand
conditions for the product. If alternative employment possibilities are not available, the
resources released may be unemployed. They argue that in such cases, the income lost by
these unemployed resources has to be deducted from the benefits of the technical change
to determine the net benefits. Their research thus contends that adjustment costs faced by
displaced labor should not be assumed away. The work of Hirschman ( ) and
Williamson (1985) show that this reasoning also applies to capital; assets highly specific to
a given transaction that suddenly become unemployed may in fact face substantial
adjustment costs. 1

Other variations on more standard consumer-surplus approaches include Capalbo
and Antle’s (1989) fiamework to include social costs such as environmental damage and
human health risk in the impact analysis. Oehmke (198 8) demonstrated that the

interaction between successfirl research and other government interventions in the same

12

commodity market can increase the cost of those interventions, thus decreasing the return
to the research. For example, a government commodity procurement or price support
payments program may suddenly incur greater costs if domestic production is increased.
These incremental commodity program costs to the government would be considered
indirect economic costs to be included along with the research costs. Not including such
indirect economic costs in the benefit-cost analysis would therefore bias the IRR upward.
He also noted that many studies in the literature - often those reporting high returns -
assume that the only relevant government intervention was the research, and thus
neglected this bias that unambiguously increases the calculated rate of return. Various
practitioners suggest that high returns among ex ante evaluations may also be due to
selection bias, in that successfirl projects are more likely to be the focus of impact
evaluation (Echeverria, 1990).
2.2.1.3 Apprppriate Benefit-Cog Methgdology

The choice of impact assessment method employed depends on several factors
including the availability of data, the objectives of the research, and the timing of the
study. As noted above, the average rate of return (from economic surplus approaches) is
considered appropriate for ex post analysis as an indicator of the financial and economic
profitability of past investments, while the marginal rate of return (fiom econometric
approaches) is considered more appropriate for situations in which research expenditures
related to the technology in question are ongoing. The econometric approaches are not
appropriate for estimating the ROR to bean research in the DR because the expenditures

in question are not ongoing, and because these methods require highly accurate and

13

detailed time-series data on inputs, outputs, and research expenditures which are not
available. While CRSP-funded bean research in the DR is fact on-going, this analysis is
only focusing on the impact of a technology which developed between 1985 to 1990 and
subsequently released in 1990.

An economic surplus method is therefore more appropriate for this analysis
because all relevant expenditures of the project under review occurred in the past. The
objective of the analysis is limited to estimating the average rate of return over the sum of
these historical expenditures. While there are no known estimates of the bean supply and
demand elasticities in the DR the consumer-producer surplus method is nevertheless used,
and sensitivity analysis of assumed elasticities is presented. The benefit-cost method is
also presented here in the sense that one elasticity scenario within the consumer-producer
supply method sensitivity analysis is the same calculation as that of the benefit-cost
method (in‘ which elasticity of supply =1 and elasticity of demand=0). Therefore, this
study will calculate a cost stream that includes research costs of PC-SO fi'om 1985 to
1990, as well as economic costs of government input subsidies from 1990 to 2004. The
benefit stream is calculated as the adopted area (hectares) from 1989 to 2004 times the
incremental yield fi'om of adopted area (kg/ha) times the unit value of beans (U SSlkg).
These cost and benefit streams are used along with assumed supply and demand
elasticities to estimate a financial and an economic average internal rate of return to

CRSP-funded bean research in the DR related to the development of the variety P050.

14

2.2.1.4 Relgvgpe prenefit-ng Literature

Beginning in 1989, the variety PC-SO was adopted nationwide primarily by farmers
who were already using irrigation and fertilizer with the previous variety Pompadour
Checa. While increased fertilizer usage was not promoted as part of a technology package
including PC-SO, sensitivity analysis in Chapter 6 nevertheless includes a scenario in which
fertilizer usage is assumed to have increased with adoption, thus augmenting production
costs per hectare. Though bean yields in the DR have increased since 1989, bean area and
total production have declined. This fact, coupled with price stability stemming from the
government’s supply control policy (import restrictions), implies that producers have
captured all the gains to incremental yields. All economic surplus is thus assumed to have
accrued to producers. The issue of incremental environmental damage is not entertained
here, because costs of production per hectare is assumed to be the same in both the with-
and without-project scenarios, and because area planted has not increased. While areas of
the SJV have recently experienced soil degradation due to impaction and over-salinization,
these problems cannot be attributed to PC-SO adoption, as mechanical land preparation
and irrigation were employed in the SJV long before 1990, and bean area has not
increased.

The technology “package” embodied by PC-SO is assumed to be factor neutral
(with respect to labor versus' mechanization) yet scale-biased. There is nothing inherent in
PC-SO adoption that implies labor savings. In fact, increased yields could increase harvest
and post-harvest costs, 80 percent of which is labor (CRSP, 1998). Though it is true that

farmers with better (i.e. mechanized) land preparation and fertilization rates have higher

15

yields with PC-SO relative to resource-poor farmers, the larger farmers enjoyed higher
yields with Pompadour Checa. However, PC-SO is still scale-biased in the sense that it is
assumed here that incremental yields to rainfed producers are negligible. ‘ Oehmke’s
concern for the interaction between increased yields from research and the costs of
government intervention in the seed and grain markets is not relevant in the DR case.
Because the government’s bean seed and grain purchasing programs existed almost a
decade prior to PC-SO adoption (see Chapter 3), and because government purchases for
neither the bean seed multiplication nor the bean grain purchasing programs are tied in any
way to the size of a given harvest (to yields), it is assumed that the costs of government
intervention in seed and grain markets did not increase due to PC—SO adoption.

2.2.2 Institutional Analysis

In the absence of consumer-producer surplus techniques, qualitative institutional
analysis can help identify not only how gains are distributed among different groups in
society but also why a given research expenditure had a high, low, or negative rate of
return. This analysis is also relevant because it includes a review of economic policy,
which can largely determine whether or not new technologies are adopted at the farm
level, which farmers may adopt them, and whether or not consumers share in the
incremental benefits.

Schmid broadly defines institutions as “sets of ordered relationships among people

that define their rights, their exposure to the rights of others, their privileges, and their

 

‘ It is possible that rainfed farmers have benefitted from the improved uniformity of
PC-SO, assuming improved uniformity has resulted in receipt of higher grades for their
marketed beans (and thus higher prices paid by intermediaries).

16

responsibilities” (Schmid, 1987). In the DR institutions that heavily influenced the
adoption and profitability of CRSP-developed technologies included farmer organizations,
the government’s bean seed and grain marketing programs, and its trade and price
policies. “In effect, (the pelicy and institutional fiamework) can provide the means
whereby the biological and physical research program will be quite productive for the
society, or item cancel it out almost in its entirety” (Schuh and Tollini, 1979). Thus, the
Dominican government’s marketing and trade policies may therefore be as vital to the
continuing profitability of bean production in the DR as it was for the adoption of P050.

2.2.3 On the Rapid Appraisal/Subscctor Method of Analysis

The subsector approach has been used efi‘ectively as a tool to conduct research on
a country’s food system, including the United States as under the NC-117 project. The
subsector paradigm was first proposed by Shaffer (1973) as the study of “the vertical set
of economic activities in the production and distribution of a closely related set of
commodities.” This vertical set of activities is composed of horizontal levels such as input
provision, farm-level production, assembly, processing, storage, transportation,
wholesaling, retailing, and consumption. Each horizontal level, or industry, represents a
transformation of inputs to produce a commodity with increasing value of time, space, and
form. As noted by Byerlee (1993), the subsector approach generates information
especially useful to policy makers and scientists when “a commodity or a region is ‘
undergoing rapid changes due to demand and supply factors or policy changes.”

Because long term subsector research can be prohibitively costly and since the

research findings are typically demanded by policy makers within a short time fi'ame, rapid

l7

appraisal techniques have been merged with the subsector analysis paradigm (Holtzrnan,
1986). A rapid appraisal survey is characterized as a broad and preliminary overview of
the organization, operation, and performance of a food system or components thereof,
designed to identify key system constraints and opportunities (ibid, 1986). These
techniques thus enable a researcher to synthesize data collected fi'om secondary sources
and key informants to generate an overview of the historical and current status of the
subsector.
2.3 Instruments and Methods of Analysis

2.3.1 Rapid Appraisal

In implementing this study, a rapid appraisal was conducted during the initial two
weeks of field work in June 1998. The primary objective of this rapid reconnaissance was
to identify itnportant constraints to increasing profitability of bean production, as well as
factors that have accounted for the widespread adoption of the PC-SO bean variety in the
San Juan Valley. The rapid appraisal also provided a clearer understanding of the current
state of the bean subsector in the SJV which was required to design the questionnaire
design and identify an appropriate sampling strategy. The rapid appraisal included a
review of secondary documents, as well as interviews with key informants including CRSP
scientists, government officials, non-governmental researchers, and farmer associations.
Additional appraisal work was conducted at various intervals during and after
implementation of the farmer survey. This activity involved continued review of
secondary documents and interviews with key informants at various levels of the

subsector, including input suppliers, researchers, merchants and traders at rural and urban

18

markets and supermarkets, bean importers, and government officials from various
programs active in the bean subsector.

2.3.2 Record-Keeping Data

The .CRSP DR scientists implemented a cost of production record-keeping (RK)
study in the San Juan Valley (SJV) during the 1997-98 winter bean season. A total of 29
bean farmers were selected fiom among of the valley’s farmer associations. Dr. Eladio
Amaud-Santana, the CRSP-Principal Investigator in the DR and Ana Mateo, one of the
two enumerators, selected the sample. The sample farmers were visited once a week
throughout the bean season by two enumerators under the direction of Dr. Amaud-
Santana. This RK study was undertaken in order to generate accurate data to assess the
farm-level profitability of irrigated bean production in the DR and to compare levels of
input use among irrigated farmers of various resource endowments. The study was
repeated with a new sample of 30 farmers during the 1998-99 winter bean season,
although the results fiom this study are not used here.

While the RR farmer sample was not selected strictly at random, the distribution of
sample farmers among each of the valley’s bean farmer associations approximates the
sampling quotas used in the larger producer survey (discussed below) undertaken in July-
August 1998. Allocating the sample among the various farmer associations resulted in a
sample that was representative of the diversity of socioeconomic, educational, technical,

and resource endowment characteristics of the bean farmer population in the SJV.

l9

2.3.3 Farm-level Adoption and Production Survey
2.3 .3.1 W

A survey of 100 bean producers in the San Juan Valley was conducted fi'om early
July 1998 through early August 1993. The study focused primarily on collecting data
required to assess the impact of bean research in the SJV, a region that accounts for 40
percent of national production, 80 percent of seed production, and has been the area
targeted by the CRSP research. The survey was designed to collect retrospective data
required to estimate farmers’ bean yields (yields of PC-SO and varieties grown prior to
adopting PC-SO), inputs used, farmers’ opinions concerning the advantages and
disadvantages of P050 (compared to their previous variety), and their knowledge of the
sources of disease-problems common to the SJV and appropriate chemical and non-
chemical remedies to said problems.

Adoption research typically focuses on estimating the overall adoption level and
the characteristics that distinguish adopters from non-adopters. However, as non-
adopters in the SJV account for less than 5 percent of all irrigated producers, the survey
instead focused on Collecting data to document the historical pattern of adoption (1'. e. the
year farmers adopted PC-SO) and the physical, institutional, and household characteristics
that accounted for this pattern. I

2.3 -3 -2 WW1

The target pepulation for the study was all farmers in the San Juan Valley who

planted beans in the 1997-98 growing season. An estimated 95 percent of the estimated

population of 1,001 SJV farmers (see Appendix A-l) are members of farmer

20

organizations, and each organization is required by law to register information on the
number and addresses of its members.2 Thus, while the farmers included in these
membership lists approximates the target population, they include farmers who did not
grow beans.

A multistage sampling procedure was used to select a random sample of 100
farmers who planted beans in 1997-98. First, a membership list was obtained fi'om each
farmer organization, and a leader of each organization was asked to identify which of its
members grew beans in 1997-98’. Second, the number of bean farmers across
associations was summed to estimate the total population of bean farmers in the SJV.
Third, each organization’s share of the SJV farmer population was estimated to determine
the percent of respondents to be selected (the quota) from each organization’s
membership list. Fourth, 27 of the 29 RK farmers were then assigned to their respective
organizations. Fifth, fiom each organization’s membership list, a random sample of
farmers (excluding those who participated in the RK study) was selected to fill the quota
assigned to that organization. Finally, at the beginning of the interview with the selected
farmers, a screening question (“Did you plant beans in 1997-98?”) was asked to determine

if they planted beans during the previous growing season. If the farmer didn’t grow beans,

 

2 Please see Appendix A for a discussion of the validity of the sample fiame.

3 Firstly, a few associations were organized with respect to crops other than beans,
though some of the farmers in said organizations grow beans. Secondly, some farmers
within bean-dominated associations may choose not to grow beans a given year. Thirdly,
some association members may be traders who have never grown beans before but simply
use association membership in order to sell beans to the government (to obtain higher
prices).

21

he/ she was replaced with another randomly selected member of the same farmer
organization.

While bean farmers with relatively large bean plantings account for a relatively
small proportion of all bean farmers, they account for a large share of the bean area. '
Thus, in-order to insure that the sample included a sufiicient number of large farmers to
obtain reliable estimates of their characteristics, the quota for large farmers was set higher
than their actual perCentage in the population (i.e. their sampling rate was higher than that
for the small and medium size farmers). Similarly, the quota for small farmers was set at a
lower percentage than their actual percentage in the population (i.e. the sampling rate of
small farmers was lower than their incidence in the population). For example, although
land reform farmers (i.e. all small farmers) make up 27.5 percent of the estimated SJV
bean farmer population (N=l,001), the land reform quota was set at 19 farmers (19
percent of the sample, N=100). On the other hand, the sampling rate for the three
associations that contain many of the larger farmers in the valley was increased slightly in
order to ensure a sizeable number of large farmers in the sample. However, in the analysis
section (Chapter 5), these data were weighted to reflect the actual percent of farmers in
each bean farm size class (i.e. the weights enable accurate extrapolation of sub-sample
statistics to the population). The actual sampling weights and their respective calculations
are included in Appendix A-2. I

2.3.4 Varietal Trials

As part of the process of selecting PC-SO fi'om among various local, foreign, and

crossed materials, CRSP scientists carried out experimental and semi-commercial varietal

22

trials fi'om 1985 to 1990. This experimental yield data serves as an upper bound to the
incremental yields of P050 assumed in the benefit-cost analysis.

2.3.5 Related Research

In 1992, the Bean/Cowpea CRSP firnded a nationwide study to document the
adoption rate of PC-SO, compare the characteristics of adopters versus non-adopters, and
collect more general information about bean production and the constraints facing bean
producers in the DR (I-Ieikes, 1993). In addition to updating findings fi'om the 1992
research, the 1998 research (which includes the RK survey, the Producer Survey, and the
Rapid Appraisal) documents the institutional arrangements and government policies that
contributed to widespread adoption of PC-SO and provides more accurate estimates of the
costs of production and the benefits of adoption, which is used to assess the
competitiveness of Dominican beans and to calculate financial and economic rates of
return to CRSP investments in the DR. In addition to recall data from the Producer
survey, this study uses the RK data to estimate costs of production and yields.

In recent years, Dr. Amaud has implemented several SJV producer surveys which
collected data on farmers’ yields, input use, planting dates, and production constraints
(Amaud et a1, 1994, 1996, 1997). While the sample sizes of these surveys ranged fi'om 48
to 105 farmers and included information relevant to the planned benefit-cost analysis,
these samples were not randomly selected. Therefore, while this data was not used
directly for the benefit-cost or profitability analysis, it was nevertheless reviewed and used

to confirm/supplement data collected in 1998 via the RK and Producer surveys. ‘

23

2.3.6 Questionnaire design

A drafi bean producer survey for use in the SJV of the DR was designed by David
Mather, Bean/Cowpea CRSP Research Associate/MSU Agricultural Econonrics Graduate
Student, and Dr. Richard Bernsten, CRSP Economist, prior to travel to the DR. Upon
arrival in the SJV, Mr. Mather and Dr. Bernsten solicited advice from Dr. Amaud, ’
Segundo Nova (also of the CRSP-DR), and Ana Mateo (a SJV bean farmer who served as
an enumerator for both the 1998 RK and 1998 Producer surveys), who all helped revise
the draft into a pilot survey. Thus, Dr. Amaud, Mr. Nova, and Mrs. Mateo played a vital
role in revising the questionnaire as well as in formulating additional survey research
questions. In addition, rapid appraisal interviews during the first week in-country also
influenced the questionnaire design. Mr. Nova and Mrs. Mateo, the 1998 Producer survey
enumerators, and Mr. Mather, the survey director, pre-tested the pilot survey on small,
medium, and large farmers before completing the final questionnaire.
2.4 Chapter Summary

The impact analysis undertaken in this study uses benefit-cost, institutional, and
rapid appraisal/subsector methods in order to not only assess the aggregate economic
impact of bean research in the DR but to also describe the distribution of the benefits of
adopting technology developed by the CRSP project, as well as the policies and
institutions which underpin the incentives to adoption. Thus, this chapter reviews relevant
literature in these methodologies and describes the research instruments developed and
implemented in 1998 to collect the quantitative and qualitative data necessary to assess

and document the profitability of bean production in the DR the aggregate economic

24

impact of PC-SO adoption, the distribution of said benefits, and the institutions that

facilitated adoption and influenced the distribution of benefits.

25

CHAPTER THREE: THE DOMINICAN REPUBLIC
3.1 Physical and Socioeconomic Characteristics

3.1.1 Geography and Agroecology

The Dominican Republic is located in the Caribbean and together with Haiti forms
the. second largest Antillean island of Hispanola. The country occupies 48,442 square
kilometers (4.8 million hectares). The climate is subtropical and topographically quite
variable, consisting of fertile valleys, high and partly deforested and eroded mountains, and
desert-like plains (FAS, 1997).

The mean temperature varies between 22 and 28 degrees centigrade (72 and 82
degrees Fahrenheit), rarely exceeding 32 or falling below 15 degrees - except at the high
elevations. Precipitation patterns are very complex, due to the influence of the mountain
and prevailing winds fi'om the Caribbean and the Atlantic Ocean. In some regions, rainfall
is evenly distributed throughout the year, while in other regions two distinct rainy seasons
occur. The Northeast and East sections of the country receive the most rainfall (1500-
2750 mm/year) while the Southwest and Northwest are much drier (350-1000 min/year)
(JICA, 1998).

3.1.2 Population Growth

The Dominican Republic’s population was estimated at 8.1 million in 1997,
yielding a population density of 167 persons/km? Population growth in the DR has
increased at an average rate of 1.9 percent annually in the 19903, decreasing fiom 2.6
percent in the 19703 and 2.3 percent in the 1980s (IDB, 1998). Although population

growth rates have declined, two additional demographic factors pose a challenge for

26

Dominican food security. First, almost 50 percent of the population is less than 14 years
of age. Second, migration fi'om the countryside to the cities has dramatically increased the
percentage of the population living in urban areas from 30 percent in 1960 to 51 percent in
1980 to 63 percent in 1997. During the 19903, the growth in urban pepulation has
averaged 3 percent annually, while that of the rural areas has averaged 0.08 percent
during the same period. While population growth and urbanization are not inherently
detrimental to economic growth and poverty alleviation, in an environment of stagnate
domestic food production and unbalanced macroeconomic growth, these trends can lead
to decreased per capita food supplies and diminished nutritional status for the poorer
segments of ’society.
3.2 The Economy

3.2.1 Macroeconomy

Macroeconomic indicators of the performance of the Dominican economy have
been very positive over the last few years. Real GDP growth averaged 5.4 percent fi'om
1992-97, with growth at levels above 7 percent in 1996 and 1997 (Banco Central, 1997).
While inflation averaged 16.9 percent annually from 1991-95, it fell to 4 percent in 1996
and stood at 8.6 percent in 1997. For 1998 and 1999, inflation has been at or below 10
percent (EIU, 1997).

However, several structural weaknesses in the Dominican economy belie the
optimism of these figures. Firstly, the DR has recorded a trade deficit in every year since
1976, which has been financed by receipts fi'om tourism, remittances, and fi'ee zone

exports. In 1996, tourism contributed USS 1.7 billion, remittances $1.1 billion, and free

27

zone exports $520 million (EIU, 1997). Family remittances, at an estimated 6.5 percent of
GDP, are quite high compared to the 4 percent average for Central America, but remain
below the 12 percent level which the IDB considers to be a problem (IDB, 1997).

Though remittances may increase consumption, imports, and voluntary unemployment (at
the expenseof savings, domestic purchases and participation in labor markets), they may
also increase investment in real estate and family-run small businesses, thus contributing to
broadening the economic base (ibrd p.4 1997). '

Secondly, although inflation has been brought under control in recent years,
interest rates have remained high (around or above 25 percent). There are several
explanations for these high rates. First, the central bank has maintained a tight'monetary
policy to keep inflation in check, to attract foreign capital, and to keep domestic capital
within the country. Second, large government expenditures often crowd out private
investment because of the payment and financing of government projects. Using the
example of several recent construction projects, private contractors hired by the
government are usually paid long after services have been rendered. These contractors
must borrow fiom private banks to finance their operations, thus crowding out other
investments (Nunez, 1999). Finally, costs of intermediation are quite high. A new foreign
investment law, effective September 1997, opened up the banking sector to firrther foreign
participation, and may eventually lead to lower interest rates. Rates are clearly segmented
by sector, however, with more productive sectors enjoying lower rates than those

available for agriculture (Libre, 1998).

28

3.2.2 Income Distribution

The third structural weakness of the Dominican economy is its highly skewed
income distribution. In 1997, the poorest 20 percent of Dominicans received 4.5 percent
of national income and the wealthiest 20 percent received 60 percent (IDB, 1997).
Therefore, while increasing per capita income is encouraging, this type of aggregate
indicator hides growing income inequality. Income redistribution is minimal as no welfare
or social security system exists, and the so-called land redistribution program of the
government has granted very few titles to land reform recipients, keeping them dependent
on the state. In addition distribution of use rights to land alone has moved extremely
slowly.

3 .2.3 Poverty 5

The final structural weakness of the Dominican economy is the persistence of
poverty. According to standards constructed on the basis of a 1992 national income and
expenditure survey, more than one in four Dominicans lives in poverty, and almost one in
ten in extreme poverty (World Bank, 1995). Rural poverty is three times greater than in
urban areas and extreme poverty is twice as prevalent. A factor which exacerbates
poverty, especially rural poverty, is stagnant, growth of the agricultural sector.

3.2.4 Structure of the Economy

Manufacturing’ and Agriculture are the two largest economic sectors, contributing

17 percent and 12.7 percent respectively to GDP in 1997 (Banco Central, 1998), and

 

‘ Manufacturing consists primarily of the processing of sugar, rice, beer, and
tobacco.

29

accounting for 18.3 percent and 14.6 percent of formal sector employment. However,
the largest share of the work force is employed in Other Services (27.2 percent) and
Commerce (23 .4 percent). Growth in Manufacturing and Agriculture has lagged that of
the rest of the economy; from 1992-1997, agricultural GDP growth averaged 4.1 percent
while manufacturing averaged 3.1 percent. The fastest growing sectors during this period
were Hotels, Bar and Restaurants at 14.9 percent and Communications at 14.8 percent
(Banco Central, 1993).

The Dominican economy is dualistic in nature, wherein those sectors relatively
Open to foreign investment and with close ties to international markets have grown rapidly
’(fi'ee trade zones (FTZs), telecommunications, and tourism), while domestically-oriented
sectors have exhibited slow and often stagnate growth. To stimulate their expansion, the
government has provided the FT Zs total exemption from taxes for 15-20 years and fi'om
import duty on inputs. Thus this growth comes at the cost that direct revenue generation
for the government does not exist for 20 years and though FTZ’s, especially tourism,
generate foreign exchange and employment, these enterprises have few other linkages to
the domestic economy (IDB, 1997).

In 1996, 434 firms operated in 34 industrial FTZs, about one-half privately
financed and operated. Most of the enterprises in these areas are involved in assembly and
light manufacturing (textiles, shoes and leather goods, electronics, pharmaceuticals, and
cigar manufacturing). From 1983-93, FTZ output expanded at an average of 26 percent
per year and tourism grew by 17 percent annually (World Bank, 1995). In contrast, the

rest of the economy grew at an average 1.8 percent per year. The implications of this

30

growth for employment is clear: FTZ employment rose from 1,000 workers in 1970 to
165,000 in 1993, or 7 percent of the total labor force, with 70 percent of FTZ workers
female. .

However, this success is qualified. Furthermore, many FTZ products have
benefitted fi'om preferential access to the US. Market via the Caribbean Basin Initiative
and the Generalized System of Preferences, and to the European Market through the
Lorne Convention. FTZ growth in 1996 fell to 5 .9 percent, and productivity has fallen
among many FTZ firms, suggesting diminished competitiveness. As a result of recent
trade liberalization, government omcials and FTZ business leaders are concerned about
the potential impact of NAFTA on FTZ textile firms. However, the amount of trade that
may be displaced by Mexico’s new preferential access to the US textile, sugar, and other
markets is not yet clear, as Mexico does not yet export significant quantities of many of
the goods on which US tarifl‘s were eliminated (IBRD, 1995).

3.3 Agricultural Sector

3.3.1 Agricultural Price and Trade Policy

In September 1990, the government initiated a major economic reform program
that. included trade, tax, financial, and exchange rate reforms. As a result, foreign
exchange taxes were eliminated, the number of difl’erent tariff rates was reduced from 140
to 7 and the range of tarifi’s was reduced from 5 to 226 percent to Sito 40 percent, all
specific tarifl’s were converted to ad valorem rates, direct export taxes were removed, and

use of non-tariff barriers decreased (World Bank, 1995).

31

The DR joined the GATT in 1994, which requires the replacement of existing
licensing and quota systems with the tariff-rate-quota system. However, the
implementation of WTO commitments remains unfinished. Although the agreement bound
all agricultural imports to an across-the-board 40 percent maximum duty, the government
proposed a “technical ratification” in 1995 to the original tariff schedule, modifying initial
tarifi’ and quota levels for eight commodities deemed politically sensitive (rice, beans,
poultry, sugar, maize, onions, garlic, and milk). Perhaps more significant is the continued
use of import licenses by the government to control the timing and quantity of imports of
these sensitive commodities. The continuation of the import licensing regime maintains a
nontransparent license acquisition process that, according to numerous accounts,
generates enormous rents for government oficials and importers while increasing price
uncertainty over time for commodity wholesalers.

The government’s assistance to the agricultural sector since the mid-19703 has
primarily taken the form of input and output subsidies. For certain areas and crops, the
government provides a limited amount of subsidized inputs (seed, irrigation, credit, tractor
services, and land), as well as support prices for several commodities (some product is
purchased directly by the government at supported prices). The goal of this assistance is
to increase farm-level net income and commodity production, though long-term
productivity-enhancing activities such as agricultural research and extension receive
marginal funding. In June 1997, the government implemented several additional measures
to assist agricultural producers, including eliminating import taxes on agricultural inputs

(notably corn, soybean meal, fertilizers and pesticides) and abolishing the official exchange

32

rate (below open market) for traditional agricultural export products (sugar, coffee,
cocoa, and tobacco). For years, the differential between the oaicial and market exchange
rates had served as an indirect tax on cash crop exports (V aldes, 1997).

3.3.2 Land Use and Ownership

Of the country’s 48,000 square kilometers (4.8 million ha), approximately 1.35
million hectares are arable - about 28 percent of the country’s total area. Of this arable
land, approximately 29 percent is planted in permanent crops, 71 percent is planted in
annual crops, and 16 percent (222,000 has) is irrigated (FAOSTAT, 1999). Permanent
pasture occupies approximately 42 percent of total area, while forests and woodlands
occupy 12 percent. Since 1965, the growth of arable and permanent crop area has
averaged 1.5 percent annually, while the forest and woodland area has decreased during
the 1980s by an average of 2.9 percent annually. However, the permanent pasture area
has remained constant during the 1980s (World Bank, 1997).

While the average farm size is 7 hectares, land ownership is highly skewed.
According to the 1981 census, farms with less than 5 hectares represented 82 percent of
landed properties, but accounted for only 12 percent of cultivable land. In contrast, farms
with 50 or more hectares represented less than 2 percent of landed properties while
accounting for 55 percent of cultivable land. Furthermore, holdings with less than 5
hectares are characteristically marginal lands, of steep or rugged topography with no

irrigation (Heikes, 1995).

33

3.3.3 Domestic Food and Cash Crop Production in 1997

The eight food crops with the largest area harvested in the Dominican Republic
are: rice (97,400 has), beans (33,300 has), plantains (34,000 has), maize (28,700 has),
cassava (16,000 has), pigeon peas (15,500 has), sorghum (8,500 has), and sweet potatoes
(5,900 has). While maize and sorghum rank fourth and seventh, these two crops are
generally used for feeding animals and not for direct consumption. As measured by total
area planted, the five most important export crops are: sugar (194,000), cacao (141,000),

coffee (125,700), bananas (33,000), and tobacco (18,800 mt) (Table 3.1).

 

    

Table 3.1 M J0? Food and Export Crops, DR, 1997

 

 

 

 

 

 

 

Maize 28,783 -1.2
Cassava 16,321 -0.6
Pigeon Peas 15,500 -14.8
Plantains 34,000 -28. 1
Sorghum 8,541 -36.9

 

 

Sweet Potatoes

 

 

 

18

 

 

 

 

 

 

 

 

 

 

Sugar 194,986 3.2

Cacao 141,521 10.8

Coffee 125,786 0
Bananas 33,000 3.7
Tobacco 18,824 -1 1.9

 

Source: FAS, 1997

34

 

3.3.4 Food Security

1 Over the last thirty years, the agricultural sector has failed to keep pace with the
food needs of the country’s growing population. While daily per capita calorie supply
averaged 2,310 in 1995 (UNDP, 1999), thereby meeting the recommended supply of daily
calories, these data disguise the fact that because income distribution is highly skewed,
food insecurity was likely widespread among the rural and urban poor. Furthermore, this
requirement is met only through increased food imports. For example, imports of cereals,
primarily wheat for processing and maize for the poultry industry, has increased from 48
percent of domestic supply during the 1970s to 70 percent in the 19903. Wheat is not
produced in the DR due to climatic constraints, and although maize is grown,
approximately 95 percent of the DR’s maize consumption is imported. The ratio of bean
irnports to consumption has similarly risen from an average 17 percent annually in the
1970s to 32 percent in the 1990s. On the other hand, the ratio of imports to consumption
for rice has-fallen from an average of 14 percent annually in 1970s to 7.3 percent in the
1990s. This “success” might be tempered by the fact that rice receives the lion’s share of
government credit assistance and is grown almost entirely under irrigation.

Rice, beans, vegetables and chicken have been staple foods in most Dominican
households for over fifty years. Yet rice and bean production has not been able to keep up
with population growth, mirroring the performance of the agricultural sector in general,
which has consistently lagged most other sectors in the economy. Because retail prices of
rice and beans have been supported through import restrictions that limit domestic supply,

these prices have increased during recent years, relative to alternative carbohydrate

35

sources such as bread, pasta, and other staples like plantain, cassava, sweet potato and
dasheen. The result has been falling per capita consumption of rice and beans during the
'1 9905.

Food price inflation in general has been an average 3 percent greater than the
overall average CPI since 1970 (Banco Central, 1998). The consequence of the DR’s
increasing food import dependency is decreased foreign exchange availability, while the
result of government supply control policies is often food insecurity for the poor. This
situatiOn leads to malnutrition and severe economic hardship for the rural and urban poor
for whom such primary agricultural commodities constitute a large portion of their
household expenditures. Production of staple food crops and their marketing
environments remain a serious challenge to poverty alleviation in the DR, for food
insecurity in this context remains a threat regardless of glowing macroeconomic
performance indicators.

3.4. Conclusion

The DR has recently enjoyed very high economic growth rates and slowing
population growth, yet highly skewed income distribution and poverty persists.
Agricultural growth rates have lagged those of other sectors in the economy, contributing
to continued high mral poverty rates. This trend will not likely change until the '
government changes the nature of its assistance to the agricultural sector from
subsidization and protection to investment in the development and difi‘usion of new

technology and improved land distribution.

36

CHAPTER FOUR: AN OVERVIEW OF THE DOMINICAN BEAN SUBSECTOR
4.1 Introduction

This research is primarily concerned with using subsector analysis to help identify
factors that contributed to the adoption of CRSP-developed bean production technologies
and associated management practices, and to identify how the benefits of these yield-
augrnenting technologies are distributed among different groups in society. An
understanding of the policy and institutional framework is required to explain why the
research expenditure had a high, low, or negative financial or economic return, and the
level of dependence of this return (profitability at the aggregate and at the individual farm
level) on the continuation of the existing institutional and policy fiamework.
4.2 Demand Analysis

4.2.1 Beans in the Dominican Diet

Rice and beans are the major staple food crops in the Dominican Republic, with
beans the principal source of vegetable protein for the majority of Dominicans, especially
the poor. Two combinations of these staples are consumed daily by the majority of
Dominicans. Arroz con habichuelas refers to a separate portion of rice served with a
bowl of beans that has been cooked in its own sauce. The beans and sauce are then placed
over the rice and eaten. El moro refers to a dish comprised principally of rice, but
containing a few beans which are cooked with the rice and served as one dish. El more is

the typical form of beans consumed by lower income households (Heikes, 1.993).

37

4.2.2 Consumer Preferences

Although various market classes of beans are consumed in the DR, Dominican
consumers have historically preferred large red varieties — the domestically produced
Pompadour Checa in particular -— over black, white, cranberry and imported red varieties.
However, as importation of pinto beans increased in the 1990s due to falling domestic
production, consumers began to be increasingly exposed to this new market class.
Because Dominican consumers classify beans by much more than simply price and color
(as throughout the Caribbean and Central America), difi‘erences in bean size, texture,
fi'eshness, cooking time, and culinary traditions can result in significant price difi‘erentials
between market classes.

Conventional wisdom in the DR asserts that pintos are primarily purchased by
poorer Dominicans, who are attracted by their lower price (normally 10-20 percent lower,
depending on the season). Thus, thus assuming that the domestic red’s culinary attributes
are preferred to. those of the pinto, a price differential may exist whereby Dominicans will
pay a premium for the preferred attributes of the domestic red. Since SEA has only
recently begun collecting weekly and monthly retail prices for both domestic reds and
pintos (before they were treated as the same commodity by SBA and INESPRE), it is
difiicult to empirically determine price elasticities of demand for the two bean classes at
this time.

However, rapid appraisal interviews with bean traders in San Juan and wholesalers
and supermarket managers in Santo Domingo in July/August 1998 strongly suggested that

the dynamics of consumer preferences, regarding imported pinto beans vis a vis traditional

38

domestic red beans, may be changing. _In San Juan, several traders said that pintos were
actually preferred by some consumers because they were “fresher” and “cooked faster.”
This is surprising, considering that at that time (July 1998), pinto prices in the market were
a bit higher than those for domestic reds (US$1.82/kg for pinto versus US$1.37/kg for
domestic red). This price differential was likely a seasonal efi‘ect in that February and
March production was still depressing domestic red prices, whereas pinto imports in the
spring through mid-summer are typically light. Interviews with three supermarket
managers in Santo Domingo revealed that among middle- to upper-income consumers,
pintos were in fact becoming popular due to their perceived fieshness and faster cooking
time relative to domestic red beans. One manager observed that some consumers were
simply using tomato paste with pintos to reproduce the red sauce that comes fi'om
cooking the domestic red.

If consumer preferences are changing from domestic reds towards pintos, such a
change may play a very important role in determining the future of bean production in the
DR If pinto imports increase in the fixture, due to either falling domestic production or a
more liberal import policy, or if current pinto imports are no longer restricted to late
summer/fall entry, then whether or not domestic reds can command price differentials
based on consumer preference may significantly afi'ect the future of the Dominican bean
industry.

4.2.3 Domestic Consumption

Available data indicates that annual per capita Dominican bean consumption has

varied considerably over the last three decades. Annual per capita dry bean consumption

39

averaged 7.5 kg during the 1970s, 8.9 kg during the 19803, and 5.5 kg from 1990-98
(SEA, 1999; FAOSTAT, 1999). Thus, average per capita dry bean consumption in the
19905 has fallen 38 percent from that of the previous decade. However, it should be
noted that the large imports of 1997 and 1998 increased per capita bean consumption
from a low of 4.6 kg/capita in 1996, to 5.7 kg/capita in 1997, and to 7.2 kg/capita in 1998
(as production/capita fell in 1998 to a low of 2.6 kg/capita). While the decline in
consumption per capita from, 1990-96 can be partially explained by sustained economic
growth during this period, it would be hard to argue that the macroeconomic growth in
this period stimulated an income effect among enough consumers to have such a large
aggregate effect on bean consumption per capita. More likely, as a result of high bean
prices during this period due to supply control (import restrictions), consumers substituted
other protein sources for beans. However, now that bean supplies have increased due to
larger import volumes, consumption per capita is again increasing.

The Dominican Republic’s 1981 national food and nutrition plan established a
recommended level of red bean consumption at 27 grams/day. Comparing this
recommendation to actual daily consumption with actual consumption per capita for the
1981 to 1997 period indicates an average daily deficit of 30 percent fi'om 1981-89 and 43
percent from 1990-97 (SEA, 1998). Only in 1988 did average daily consumption meet the
recommended level.

4.2.4 Imports

As Table 4.1 demonstrates, Dominican bean production per capita has declined in

the 1990s, leading to increasing dependence on imports to meet domestic demand for dry

40

beans. Importsl increased dramatically in the late 19905, accounting for half of domestic
consumption, and as much as 67 percent of consumption in 1998.
arts in the DR

Table 4 1 Per Ca n ita Bean Production Consum ttion and Im

    

     

 

5.5

  
  

 

6,078 6,291 14,404

 

 

 

17% 10% 32%

 

' 1990-98 consumption and imports estimated using SEA/INESPRE data. Data for 1970-89 is
fiom FAO. Consumption is defined as food consumption, estimated to be 95% of total annual
production plus imports.

Source: SEA, 1999; FAOSTAT, 1999

 

 

4.3 Production Analysis

4.3.1 Bean Classes and Varieties

While red, black, and white beans are grown in the DR, 88 percent of the bean
area is planted to red molten bean varieties. The most common reds are strong and erect
shrub types, particularly Pompadour Checa, PC-50 (a selection of the Pompador Checa
land race), Constanza I, Jose Beta, and JB-178 (a cross between Jose Beta and C1308).

Black varieties account for 8.1 percent of bean area planted, and white varieties 3.6

percent.

 

' Bean imports are controlled by the Secretaria do Estado de Agricultura (SEA),
the Dominican Ministry of AgriCulture, which defines the quantity, market class, timing,
and importer for all bean imports. Importing agents include INESPRE, the government
food marketing parastatal, and various private sector importers who must acquire an
import license from SEA for each shipment (discussed in greater detail below).

41

4.3.2 Bean Production Location and Seasons

Planting and harvesting dates vary greatly within and between regions because the
Dominican Republic is an island where rainfall patterns are influenced by location (leeward
versus windward side) and topography. In addition, producers with access to irrigation
have additional flexibility with respect to planting dates. While beans are planted
practically year-round in the Dominican Republic, the three principal bean seasons are:
spring, fall and winter. Vegetative cycles vary between seventy-five and one hundred five
days, depending on bean variety and elevation.

The spring crop (March through June), which accounted for an average 17.6
percent of total annual red bean production (1994-97), is mostly grown in high altitude
areas (500 to 1300 meters above sea level). The fall crop (July through October), which
accounted for an average 23.5 percent of total red bean production during this period, is
generally grown between 400 and 500 meters above sea level. Finally, the winter bean
crop (November through early March), which accounted for 59 percent of total
production (SBA Breve, 1990 and SEA/DEA, 1998) during this period, is grown below
400 meters above sea level (in valleys). The timing of these seasons and their respective
shares of total red bean production have shifted considerably in recent years, due primarily
to an increase in irrigated area as a percentage of total bean area planted, changes in the
timing of bean production in the San Juan Valley (Southwest region), and an increase in
that region’s share of total production — which increased from 40 percent in the late 1980's

to over 50 percent in recent years.

42

Data for area planted to red beans and red bean production are reported by the
Ministry of Agriculture’s administrative zones (Table 4.2). From 1994-97, the most
important regions for red bean production were the Southwest (a four-year average of 52
percent), Northcentral (16 percent), North (8.3 percent, and East (7.8 percent). Notable
changes in regional bean production from the 1985-89 to the 1994-97 period occurred in
the South, which declined fiom period averages of 10.1 to 5.4 percent of national
production, and the Northwest, which declined'from 9.9 to 5.2 percent. Between these
periods, average production shares increased in the Southwest from 41 to 52 percent, and
in the East fiom 3.5 to 7.8 percent.

4.3.3 Recent Changes in Red Bean Area and Yields

Analysis of red bean production in the DR (Table 4.2) suggests several national
and regional trends. Comparing the period 1985-89 to that of 1990-94, total average
annual red bean. area2 fell 20 percent, production fell 11 percent, and yields increased 14
percent, flour a period average of 7 19 to 821 kg/ha. Comparing the period 1990-94 to
that of 1995-98, average annual red bean area fell 13 percent, production fell 14 percent,
and yields fell 3 percent, from 821 kg/ha to 796 kg/ha. Comparing the 1985-89 period to
that of 1995-98, yields increased 10.8 percent, while area declined 30 percent and
production declined 24 percent. While a simple explanation for this trend could be an
increase in the percentage of area under irrigation, the data shows otherwise.

During the 1985-89 period, irrigated land accounted for an average 35 percent of

total red bean area and 47 percent of total production (SEA Breve, 1990). From 1990-94,

 

2 Through this document, area refers to harvested area, unless otherwise noted.

43

 

83: £238. oh< o_Eo:8m ow oEoEotméQ <Mm coo. o>2m <mm .ooSom
.28 052: E 2 cocoa—coke! 38,—. .3228: E 2 82 30,—. _

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mob : ones. 2.23. m: 2%.: n2 :. on. «no S an on
R» . n: 2% owe oo- Iae owe ooe ooe
Se E me Now v.2 a: E _.o2 o2
m3 3 no so o2 v.2 new as 22
a: No S E 3 do o; oo on
«S S 2 8o 3 S E E S
ooo n2 ed omo v.2 2: one o.: n2
one 3 No on. 3 oo 5. no 3.
to on we mom 3 2. 2o 3 oo

EEC 52.8.55 82 fine noooeoea 8:. 935 8:885 82

22> 28> .x. 28> .x. 22> 28> 2. 28> axe 22> 282 .x. 28> 2.

o o a

a: o... e. Seen .5. .e 22> E... 522.5... 8E .55. .3. N... 2.5,

44

irrigated land accounted for an average 37 percent of total bean area planted and 49
percent of total bean production (red, black, and white varieties) (SEA, 1999). From
1995-1998, 'irrigatedland accounted for 33 percent of total bean area planted and 42
percent of production (SEA, 1999). Thus, while irrigated area did increase in the 1990-94
period, it declined back to below the average for the 1985-89 period, while yields
increased between these two periods by 10 percent.

Considering that the average difl'erential between irrigated and rainfed bean yields
was 66 percent during the 1985-89 period (989 kg/ha for irrigated vs. 595 kg/ha for
rainfed areas), increases in the average national red bean yield fi'om the 1985-89 to the
1990-94. periods are at least partially explained by an increasing share of total bean area
under irrigation (fi'om 35 to 37 percent). However, the yield improvement of 14 percent
between these two periods cannot simply be explained by an increased proportion of area
under irrigation, demonstrating that a change in technology (increased input use, increased
technical or allocative efiiciency, etc.) contributed to this increase. Annual data for area
planted, domestic production, and import levels are reported in Figure 4.1.

4.3.4 Recent Changes in the Seasonality of Red Bean Production

Until 1990, bean farmers in the San Juan Valley grew beans during two seasons:
one fiom August/September-November/December and then again from December/
January through March/April. Afier a devastating epidemic of bean golden mosaic virus
(see 4.4.6.1) in the SJV during 1989-90, the Bean/Cowpea CRSP project convinced the
regional government to restrict bean planting in the SJV to a period between 15

November-and 10 January and to strictly enforce a fallow period from August

45

 

Figure 4.1: DR Red Bean Supply 1981-98
(production, imports, area planted)

0|
0|

 

T)

 

 

'000M
3.
or

 

pgrts(
' OI

 

to
Or
a.
0
Area Planted ('OOO Ha)

Production & Im
at

 

 

Ware-WA“

C I

1 I

I 0

MWF,‘rt‘a‘.‘,.‘vf.j.‘.fgrunt-1.5“; _ Hyman-c». .:.'.'._.‘-'.‘.‘-'.‘.'. .x‘n‘ j.~_.‘.j.'.-.

. .

. o

G 0
Wuvm‘flr'c‘m-Y4fifi.3);.‘.3‘.'.;.:.;.;.;.;.;.;.;.;.;.;.;.;.; ;.:.;.:.;.;.;.;V;.‘...4. . :1 In. ‘

.

2 s

l :5. 2

l' ;:l é : E. :5

§§§ SE? 35‘ -
.5 f; ‘u '
-,_: 23 ': ' "
... ff: ... :: ... .. -

I as: a £52 2::
ft 5 =2: 2* it
:5: ?. t‘ :5: 1 g i =5 5:
3:: 1:5; i :‘5 . . t, '32 =:: :3 g
=- ‘3 S 3 ’ S 3 ‘t i? 5

183 85 87 89 9193 95 97
Production (MT). Imports (MT) -I- Area Planted I

5

 

 

 

 

 

through mid-November, during time which no hosts (beans, peppers, tomatoes, etc.) to
the whitefly could be planted anywhere in the valley. Thus, beginning with the 1991-92
season, farmers in the SJV went fi'o'm growing a fall (small) and a winter (large) crop to
growing one large winter crop. A ’second planting period change in the SJV occurred in
1994-95, when the CRSP project presented research to the local government
demonstrating the benefits of an earlier planting period, 5 November - 15 December. The
regional government enacted this modified planting period into law for the SJV, which
remains in force to the present.

These changes in the bean planting schedule in the San Juan Valley have resulted in

a decrease in the fall season’s share of annual national production and a substantial

46

increase in the winter season’s share. From 1985-89 to 1994-97, the October through
December season’s production share fell fiom an average of 23 to 14 percent, and the
January and February season’s share increased from 20 to 48 percent. In contrast, the
March and April season’s share, which used to include the end of the winter SJV harvest,
fell fiom 39 to 15.8 percent. This change in production seasonality has several
implications. Declining domestic bean production requires a greater level of been imports.
However, as an. increasing share of domestic production occurs during January and
February, the optimal timing of bean imports is late summer and fall, when bean prices
begin to rise (as discussed below). Furthermore, if beans are imported during the first six
months of the year, imports will further depress prices for the Southwest region
producers, who are notably politically organized and active.

4.3.5 Land Tenure and Farm Size

According to the 1992 CRSP-funded nationwide survey of bean producers
(Heikes, 1993), 61 percent of total bean area is owned, 16 percent is under the Agrarian
Reform program, 13 percent is rented, and 9 percent is sharecropped (Heikes, 1993). As
the last national agricultural census was in 1981, it is difiicult to estimate with certainty the
farm size of bean producers, although data fiom the 1998 producer survey in the San Juan
Valley indicates that land distribution among bean farmers is similar to that of national
agricultural land distribution averages (even since 1981; see Chapter 3.3.2), and also that
smaller commercial producers are moving out of bean production and into other crops
(Chapter 5.4. 1). However, it is clear fi'om national level data that irrigated land, as a share

of total production, is increasing, and that been yields have increased (unsurprisingly), as

47

noted above (Chapter 4.3 .3).
4.3 .6 Constraints to Increasing Productivity
4.5 -6.1 WW

- The four principal bean diseases in the DR are bean golden mosaic virus (BGMV;
Mosaico Dorado), rust (roya), web blight (Mustia Hilachosa), and common blight
(Bactefiosis Connm). In the 1992 nationwide survey, producers were shown pictures of
these four diseases and asked ifthey had experienced any of them during the last five
years. They were also asked to judge the severity of each with respect to yield reduction.
The results (Table 4.4), demonstrated the prevalence, frequency, and severity of these four
diseases for bean production across the DR in 1992. From the perspective of these

farmers, rust and BGMV were the most prevalent, frequent, and severe bean diseases’.

 

Table 4. 3 Prevalence, Frequency, and Severity of Bean Diseases III the DR, 1992

 

 

 

 

 

 

 

 

Bean Golden Mosaic Virus 73 2.6 73 '
Rust 77 . n.a. 70
Web Blight 63 ' 2.1 68
Common Blight . 38 1.3 53

 

 

Source: 1992 CRSP Nationwide Producer Survey; N=219, table from Heikes, 1993.
percentage of producers who experienced the disease at least once between 1988-

1992

2 number of years out of the last five that the producer experienced the disease.

3 percentage of producers stating that the presence of the respective disease

significantly reduces yield.

 

 

 

 

3 Results from similar questions posed in the 1998 Producer Survey in the San
Juan Valley are discussed in Chapter 5.

48

4.3.6.2 Other Production Constraints

When asked to rank their major production constraint, producers in 1992 cited
insuficient rain (40 percent), insects (19 percent), too much rain (17 percent), disease
(13), and other problems (11 percent). These results help to explain the recent decline in
rainfed bean area as a share of total area, which has declined steadily over the last decade.
Confirming these results, the government attributes low rainfall in the past few years as the
main factor in decreased bean area (SEA,1998).

4.3.7 Agricultural Organization Membership

- Nearly two-thirds (63 percent) of respondents to the 1992 nationwide survey were
members of an agricultural organization (Heikes, 1993). Since many organizations are
primarily social clubs, respondents were asked'if their given organizations were actually
beneficial to the production of beans. Forty-seven percent said that their organization was
“beneficial,” with the benefits being better access to seeds (45 percent), credit (30
percent), and overall improved influence (25 percent). Other producers cited benefits such
as better access to technical assistance, information, chemical inputs, and land preparation
services/equipment. These responses are very similar to those of San Juan Valley farmers
(Chapter 5.3.1), who rely on their association for quality seed, credit, and negotiation
power with INESPRE regarding INESPRE’s purchase price each season (as well as

access to INESPRE’s purchasing system in the first place).

49

4.4 Bean Price Analysis

4.4.1 Bean Price Seasonality

Approximately 59 percent of bean production is fi'om the winter season, with a
corresponding February harvest period, 17.5 percent from the spring season with a June
harvest period, and 23.5 percent from the fall season with an October harvest period
(Section 4.3.2). Bean prices at the farmgate,‘ Wholesale, and retail levels are afl‘ected by
the timing and quantity of these harvests, suggesting that many producers sell their
surpluses soon after harvest.

The seasonality of wholesale and retail bean prices (1993-97) is demonstrated in
Figure 4.2. Although the average January wholesale price is 5 percent above the annual
average price, wholesale prices begin a sharp decline after the February winter season
harvest, falling to 15 percent below the in April. They gradually climb back to slightly
below the retail price in April, and then they gradually increase to slightly below the
average by July, as the June spring season harvest arrives. At this point, the domestic
production fiom the winter season has reached the retail level (or has been consumed).
Thus, prices increase, dramatically and peak in October at 15 percent above annual average
price. Following the October (fall season) harvest, prices fall back to the annual average
by'mid-December. The maximum spread between minimum (April) and maximum

(October) wholesale prices averages 30 percent.

 

‘ Monthly farmgate prices were only available for 1997. The seasonality of
farmgate prices is thus implied by wholesale price seasonality.

50

 

Figure 4.2: DR Red Bean Price Index
1993-97

120

 

l
U

 

2100..
3 Zn‘

 

 

“L- / ..

\\ [271/ 33:...

0
0|
1

l

8

 

 

I

Annual Mean

(0
0|

 

 

\\,
\/

0 i i i i : : : 4 i i : :
Jan FebMarAprMayJun Jul AugSepOctNovDec
Month

 

8

 

1

Index
a:
OI

 

l
I

 

 

 

 

 

 

' The retail price index follows a similar cycle, although price increases and declines
are of smaller magnitude. The spread between minimum (April) and maximum (October)
retail prices is 19 percent. Comparison of the monthly retail real price index average fiom
the 1985-90 period to that of the 1993297 period (Figure 4.3) clearly indicates that

monthly retail price fluctuations have diminished considerably since the 1935-90 period.

51

 

 

130

Annual Mean=100
8 a '8

Index
8

Figure 4.3 Red Bean Retail Price Index
1985-90 vs.1993-97

 

 

:: , /\

 

 

 

 

Jén FSb Mar Apr May Jt'tn Jul Aug st'ap dot Nov Dec

 

‘ [-3. 1993-97 + 1985-90

 

 

The coefficient of variation in monthly retail price from the 1993-97 period (8.2

percent) is in fact halfthat of the 1985-90 period (16.6 percent). Noting that this index
was constructed with deflated monthly prices, one possible explanation for the decreased
monthly retail price fluctuation is that increased bean imports in the 19908 (due to
declining domestic production) have .enabled wholesalers and retailers access to cheaper
beans in the second halfof the year, when domestic production is low. Decreased
fluctuation from the annual mean during the first half of the year is likely due to the change
in planting dates in the SJV (as discussed above), resulting in one large harvest in

February, as opposed to a small one in December and another in March.

52

 

 

4.4.2 Trends in Real Prices

As shown in Figure 4.4, annual average farmgate red bean prices (real) over the
1984-97 period have been quite variable, although the trend since 1994 has been
downward. This decline in the real bean price since 1994 helps to explain why the bean

area has declined substantially since that year, reaching its lowest level coming in 1997.

 

Figure 4.4: DR Red Bean Prices
Real Annual Average (1 984-1997)

1.80

 

 

 

1.60 _ 3 2 m I
1.40 _ \
T/r/AJNRV/"Tifi

951.00

 

 

 

0.80

0., .. /K\V/’°*/
N

 

 

I.

 

 

L
I

 

0.40 3 4 3 3 3 3 3 3 3 3 3 3 3
84 86 88 90 92 94 96
Year

l-n— Retail -e- Wholesale-e— Farrngate

 

 

 

 

 

Real annual retail and wholesale prices have declined since 1994, although per capita
consumption of beans has continued to decline throughout the 1990s. It should be noted
that although national average real bean prices have been falling in recent years, the prices

received by San Juan Valley farmers are on average higher than the national average,

53

helping to explain why the bean area in this region has not fallen as much as it has in other
areas. These higher prices are the result of bean purchases by SBA and INESPRE, as
discussed in the following section. '
. 4.4.3 Trends in Relative Prices

Relative prices are alien used in subsector analysis as a measure of the efficiency
of domestic production at a given level of the subsector. Figure 4.5 shows real farmgate
prices for Dominican red beans, compared to relative international parity prices (the
financial import parity price (US pintos), and the economic import parity price (US
pintos))’. It is clear from this table that the DR is a highly inefficient producer of beans,
compared to US pintos. These findings are further amplified at the wholesale and retail
levels, due to the oligopsonistic/ oligopolistic wholesale bean markets for domestic and
imported beans. Even in 1998-99, when DR real bean prices reached their lowest levels in
decades, wholesale real red bean prices in the DR were still 20 to 25 percent higher than
red bean prices across Central America (CORECA, 2000), and DR retail red bean prices

were 30 to 35 percent higher than those of Central America.

 

’ Chapter 6 includes a discussion of the calculation of financial and economic ,
import parity prices (IPP) used in this study. Financial prices include all taxes and tarifi‘s,
while economic prices exclude them. The financial IPP in this table demonstrates what US
pintos could sell for in San Juan using US price plus transport, shipping, insurance, port
fees, tarifi‘s, taxes, Dominican transport 'to San Juan, etc.

54

 

Figure 4.5: Relative Bean Prices
Real Annual Average (1984-1999)

 

 

 

 

 

 

 

 

 

l j l I
I

aaeaewmaeeaameee
Year
[-I- DR Farmgatee- Financial IPP-w Economic 1P}?

 

 

 

 

 

 

 

4.5 Government Policies Afl'ecting the Bean Subscctor

4.5.1 Introduction

4.5.1.1 Wm

Several Dominican government policies directly and indirectly affect the bean
subsector. In the bean grain market, INESPRE, the government food marketing
parastatal, purchases an annually varying amount of domestic and imported beans and
resells them through its consumer food subsidy program. With respect to inputs for farm-
level bean production, SEA provides partial funding for bean research, subsidizes the

multiplication and distribution of bean seeds, and provides technical assistance through

55

extension services. Various government agencies also provide subsidized inputs as credit
(Banco Agricola), land preparation services (PROSEMA), irrigation (INDRHI), technical
assistance (SBA), and some chemical inputs (CVMA sales centers). SEA also oversees
lNESPRE’s regulation of the quantity and timing of bean imports, and the Agrarian
Reform Institute (IAD) guarantees use rights to agrarian reform land. In addition, in
1990, the government passed legislation prohibiting the distribution of inputs to been
producers intending to plant on steep sloping lands.
4.5.1.2 MW

Perhaps the primary policy that shapes the opportunity and constraint sets of
subsector agents is the government’s import licensing regime which limits the timing and
quantity of bean imports. Because the DR is not self-sufficient in beans, this means that
SEA is able to support the annual bean prices by adjusting the quantity and timing of been
imports each year according to domestic production. The average nominal protection rate
(NPR) for red beans in the DR over the 1985-89 period was 58, and 116 over the 1990-94
period (V aldes, 1995). It is therefore clear that consumers in the DR pay more for their
beans than they would in the absence of import restrictions — even considering that
domestic red beans may very well command sorne price differential over imported pintos,
given preferred characteristics of the domestic red.

4.5.1.3 QAZIIIZ leigatigns

Although the Dominican Government signed the GATT treaty in 1994, soon

thereafter it submitted a “technical ratification” (TR), whereby the government claimed

exemption of certain commodities (rice, beans, sugar, corn, garlic, onion, milk, and

56

poultry) fi'om their original GATT agreement tarifi' schedules due to the political and
economic importance of these commodities. After some negotiation, the US. agreed with
this TR Under the TR, the DR must allow the importation of up to 12,000 metric tons
(MT) of beans a year at a within-quota tarifi‘ of 25 percent. For imports above 12,000
MT, the DR may impose an out-of-quota tarifi‘ of 99 percent. The initial quota will be
increased over 10 years (from 1995-2004) to a final level of 18,000 MT, and the over-
quota tarifl‘will be reduced to 89 percent (FAS, 1998).

The effect of the import levels required by the GATT on the domestic production
of beans is unclear, as the implementation of the agreement remains uncertain. Although
the TR tariff rates and quotas are not in dispute (at least between the US and the DR), it is
clear that the non-transparent distribution of import licenses for bean imports, as currently
practiced, is against the letter and spirit of the GATT. While open tenders for the sale of
these licenses (the timing and quantity determined by SEA) are the official policy, these do
not occur in practice. However, in the short term, the import quota level itself should not
be a large threat to domestic bean growers because the DR imported nearly 25,000 MT of
beans in 1997 (all at the within quota tarifl‘ of 25 percent) and has approached or exceeded
that level almost every year since 1987.

What may be more problematic for domestic growers is if consumers — especially
middle or upper-income consumers - become accustomed to and begin to prefer imported
pintos. If this were to happen, as some supermarket managers and open market traders
indicated is occurring to some degree, pintos may gain enough market share year round to

dampen demand for domestic reds following the large winter harvest, thereby lowering

57

farmgate prices for domestic reds.

There are two additional scenarios in which pintos, without a change in consumer
preferences‘, could result in lower consumer demand for domestic reds (and thus lower
farmgate prices). First, if bean importers are allowed unrestricted entry of imports with
respect to seasonality, spring farmgate prices may fall if imports arrive then (currently very
little is imported in the spring, although this is difficult to measure given thepolitically
sensitive nature of these data). Second, if import licenses are distributed through a
transparent tender system, a portion of the rents that are now reportedly captured by
government oflicials will be passed on to consumers in the form of lower prices. Thus,
lower-priced pintos would surely decrease demand for domestic reds. Modeling the
potential efl‘ects of various combinations of import levels and timings is not attempted
here, though this is necessary to more accurately predict the potential efi‘ects of these
changes on domestic prices and production.

4.5.2 INESPRE

4.5.2.1 Ingitutignal Background of INESPRE

Contemporary price controls and the legal and institutional structure needed to
implement intervention in prices, are a legacy fi'om the Trujillo period (1930-61) (Greene
and Roe, 1991). By the end of this era, his family and associates controlled a wide variety
of manufacturing and industrial firms (Bell, 1981). After Trujillo’s death in the mid-

19605, many of these firms became part of the Dominican Corporation of State-Owned

 

‘ Conventional wisdom holds that consumers prefer the color and culinary
characteristics of domestic reds vis a vis imported pintos, although poorer Dominicans
purchase pintos because they are usually cheaper.

58

Enterprises (CORDE), such as the state-owned sugar enterprise (CEA), which evolved
from the process of nationalization of this conglomerate of private firms (Greene and Roe,
1991). .

Control of food prices was first implemented in 1957 by the Bank of Agricultural
and Industrial Credit, which was authorized to set rice prices for producers, wholesalers,
and retailers. The 1969 reorganization of the public sector enterprises in agriculture led to
the creation of the Instituto de Esmdilizacion de Precios (INESPRE). INESPRE initially
took over the government’s rice marketing and price control activities and later became
the principal parastatal for implementing the marketing and price policies for other foods
such as dry beans, sugar, milk, eggs, garlic, onions, chicken, wheat and corn flour,
soybean oil, pasta, and bread.

Implementation of these policies led to INESPRE involvement in activities such as
commodity procurement (domestic and/or international), transportation, storage, and sales
at the to wholesalers and retailers, as well as direct sales to consumers through food
subsidy programs. The basis of the government’s price control activities for many crops is
SEA’s import licensing regime, by which INESPRE maintains a measure of control over
domestic commodity prices. As the Secretary of Agriculture (the head of SEA) is also the
director of INESPRE, SEA effectively controls INESPRE’s activities.

Since its inception, INESPRE’s level of involvement in these various levels of the
marketing system has fluctuated by commodity, by season, and by year — depending upon
several factors including the financial situation of the government as a whole and of

INESPRE individually, the government’s expenditure priorities in a political economy

59

context (the proximity of elections), etc. As its title suggests, the stated goal of INESPRE
is to stabilize prices of politically sensitive commodities for both producers and
consumers. The idea is to offer producers “fair” and “negotiated” prices for their product
as well as import protection fi'om typically lower world prices, while protecting consumers
from seasonal price increases through timely public or private sector imports, as regulated
by SEA’s irnport licensing regime. However, a closer inspection of INESPRE-s marketing
activities in the bean subsector demonstrates a significant difl‘erence between publicly
stated goals and actual performance outcomes of INESPRE involvement in commodity
markets. .
4.5.2.2 INESPRE} dgmgic bean pgrgflses

INESPRE’s authorization to participate. in bean marketing was issued in August
1975, by Decree # 1194. INESPRE participates in the Dominican bean subsector by
engaging in the procurement of domestic and imported beans, their transportation and
storage, and the eventual sale of beans to wholesalers, retailers, and directly to the public
through a consumer food subsidy program. 3

INESPRE purchases several varieties of domestic red beans for eventual resale to
consumers, including Pompadour Checa, PC-SO, Jose Beta, JB-178, and Yacomelo.
Beans are typically only purchased by INESPRE in San Juan (Southwest Region) and
Santiago (North Central), areas that fi'om 1993-97 accounted for an average of 51 and 15
percent of total annual bean production, respectively. INESPRE’s purchasing activities in
San Juan demonstrate how the agency’s commodity support activities actually serve to

create the very price and income risk it is paradoxically intended to remove (from the

60

market).

I Although in some years INESPRE buys a significant amount of the production of
these two regions (37 percent of been production in the SJV in 1998) at a price-
consistently higher than that ofi‘ered by private traders", INESPRE’s price is not
announced before or during the season, nor does the government even commit to a
procurement quantity and/or timing 'prior to planting. Even more significantly, when a
producer sells beans to INESPRE, he is paid anywhere from two weeks to five months
later. Therefore, INESPRE’s domestic procurement activity is not the predictable and
transparent type of process usually associated with government attempts to stimulate
supply response. Furthermore, INESPRE’s procurement activity changes from year-to-
year, seemingly driven by the current financial and political situation of the government.

As data from the 1998 Producer Survey in the SJV shows (CRSP, 1998),
INESPRE’s purchases do not actually result in much higher prices, the program primarily
benefits larger farmers, and the supply response is questionable - given that SJV bean area
has. not declined as has the area in other regions, while DR bean production on aggregate
has declined in the last decade. The implication of these findings is that the activities of
INESPRE are too small, too unpredictable, and poorly targeted if supply response were
the true goal of the endeavor.

For example, although INESPRE purchased beans at $1.57/kg (RD$1,03 5/qq) in

1998, it did not pay growers in firll until an average 13.8 weeks after the sale (CRSP,

 

7 The difi‘erential between the INESPRE and trader price was 15-22 percent in the
past three years, although it averaged about 13 percent higher since 1992.

61

1998). This delayed payment resulted in financial losses to farmers for interest payments
on outstandingloans, as well as for the loss of the interest on the difference between a
farmer’s total sale revenue less his total outstanding loans, valued over the waiting period.
When lNESPRE’s “purchase price” was adjusted to account for these real costs to
farmers', the INESPRE adjusted price for the sample was actually $1.47/kg (RDS964/qq),
which represents a seven percent decrease in the “sale price” ofl‘ered by INESPRE. Thus,
instead of enjoying a price 18 percent higher than that of average trader price of $1.29/kg
(RDS849/qq), farmers who sold to INESPRE in 1998 received an adjusted price 12
percent higher than the trader price. However, because larger farmers have access to
much lower interest rates, their adjusted prices are higher than those of small farmers (see
Chapter 5.11.4). Many of the smallest farmers in the SJV don’t sell to INESPRE because
they need cash at harvest to pay off their loans, which are usually made at high interest
rates. Therefore, the case of INESPRE producer price support activities mirrors that of
US. crop price supports in that the benefits 'of such programs typically accrue to the'
largest farmers (those most able to remain profitable without support), while claiming to
protect the most vulnerable. These figures and INESPRE’s purchasing and payment

activities are discussed in more detail in Chapter 5.1.3 and 5.1.5.

 

' This calculation was made using information collected in the 1998 Prbducer
Survey on each farmer’s sale volumes and sources, average interest rates on outstanding
loans, and how long the farmer waited for payment. It is assumed that outstanding loans
are not resolved by each farmer until receipt of INESPRE payment. Ifthey were resolved
before this time, the farmer would have had to borrow the money fi'om somewhere
anyway, or transfer it fiom an alternative productive use. Farmers with loans fi'om Banco
Agricola were not charged interest on said loans once their beans were received by either
INESPRE or SEA.

62

4.5.2.3 INESPRE} Impgfls

INESPRE was the only official importer of beans until 1988, when the private
sector was also allowed to import beans, although only after obtaining an import license
fi'om SEA that restricts the timing, quantity, and market class of beans to be imported.
However, the role of INESPRE as a bean importer itselfhas diminished considerably over
the past decade, due to the organization’s continuing financial dimculties (coming fi'om
the'subsidies absorbed in the operation of its domestic purchase and sales programs).
From 1988 to 1991, INESPRE itself accounted for 95 percent or more of been imports;
but during 1992-94, that share fell to between 33 to 67 percent, and then bottomed out
from 1995-98 to between 0 to 10 percent.

According to every non-governmental source contacted, even a former customs
oficial, the import licensing system is used by omcials to extract rents both at SEA
(distribution of import pemrits) and at customs (enforcement of permits). Though SEA
claims that import permits are sold through an open bidding auction, the reality is that this
never occurs. Thus, while there is surely “bidding” for these permits, the value of the bid
is captured by an official — not by consumers in the form of price savings on imported
goods. In addition, SEA’s official policy is to only import beans in the fall, When domestic
production is lowest. However, permits are still granted in the spring, thereby depressing
prices at a time when the country produces the most beans. One additional problem is that
while the permits specify a specific market class, timing, and quantity of the shipment
permitted to enter the DR, customs officials at times do not enforce one or more of these

conditions. Thus, the market class, timing, or quantity of what actually enters the country

63

does not match the permit. The implication of this system is that imported beans are sold
at higher prices than would othemise be (given the “cost” of the permit captured by an
oficial), and that rents are extracted in this manner fi'om consumers by government
ofiicials.

Rents are not only captured by oficials who sell import licenses, but also by
importers who enjoy a markup of approximately 20 to 40 percent above the wholesale
import parity price of US Colorado Pintos shipped to Santo Domingo”. Thus,
conservatively assuming a 20 percent markup (about US$7.50/cwt) existed throughout
1998, during which time the DR imported 913,000 cwt of pintos, government omcials and
importers shared rents of approximately US$6.85 million. By 1999, a few of the more
powerful SJV farmer organizations (not all of them) had negotiated with the government
and importers and managed to share in these rents”.

45.24 W
Since 1979, INESPRE has implemented some version of a subsidized, targeted

basic foods program. While these programs are renamed every 2-4 years, the program

 

’ Calculation based upon monthly data fi'om August 1998 to July 1999 from
USDA AMS (US Pintos) and SEA (wholesale pinto prices in Santo Domingo), employing
the import parity price calculations presented in Chapter 6.

1° According to several key informants, in 1999, the government negotiated a deal
in which several wholesalers would be given access to import licenses (in the fall) under
the condition that_they purchase beans from select SJV farmer associations at higher-than-
rnarket prices (in the previous spring). In this way, a few of the SJV farmer associations
share in the rents associated with the import licensing regime. Reportedly, in 1998, a
group of wholesalers purchased SJV beans under a similar arrangement only to later be
denied access to import licenses as promised by the government. These events firrther
highlight the politicized, erratic, and destabilizing nature of government involvement in
bean marketing.

64

remains essentially the same. Under the current edition, Programa de Alimentos para
Todos (program of food for all), direct sales are made from the backs of yellow program
trucks or fi'om program stores, which are located in urban areas and regional centers. In
previous years, large numbers of program stores - financed by the government and
stocked with INESPRE-purchased agricultural commodities — were scattered throughout
urban and rural areas to “protect the consumer from middlemen.” In the past, INESPRE
has also sold domestically purchased or imported beans to wholesalers and retailers at
given prices, and has then regulated those sellers to ensure their margins were not
excessive.

According to INESPRE data, during the period fi'om 1983-1991, INESPRE’s
direct and indirect sales to consumers accounted for an average of 24 percent of total
domestic consumption (INESPRE, 1998). This was clearly the height of INESPRE’s
marketing activities, when domestic purchases, imports, and sales were at their highest
levels. Eventual financial dimculties forced INESPRE to curtail their degree of activity.
Thus, their direct and indirect sales from 1992-98 averaged 6.8 percent of total domestic
consumption, with four of these years registering shares around 1 percent. It is not
possible to accurately calculate the prices charged by INESPRE to wholesalers, retailers,
and the general public as available data only reports aggregate quantities sold and the
value received by year. Although data on aggregate value received and quantity sold for
INESPRE sales is available, it is not possible to estimate the subsidy to consumers per
year because information on how much was sold at the wholesale and retail levels is not

available.

65

4.5.2.5 :1th thgre gfINESPRB
While INESPRE was supposedly within the sights of the Fernandez

administration’s reform agenda when he took oflice, the institution actually thrived in his
tenure. By 1998, it was back to its pre-l990 levels in purchasing, importation, and sales.
How long the government will be able to continue underwriting the losses incurred by
these programs is unclear. However, it does seem clear that the most powerful lobby
against reform will come fi'om bean producer associations, given their low political
mobilization costs vis a vis consumers. Importers who have access to import permits
would also be expected to fight reforms.

4.5.3 SEA Seed Multiplication and Distribution Program

SEA has supported a seed multiplication and distribution program for many years
because of the importance of disease-free seed for insuring high national bean yields, and
due. to the susceptibility of stored bean seeds to heat, moisture, and insects, especially
under storage by small to medium size farmers. In recent years, the multiplication
program has taken a form close to the following: the CRSP develops, tests, and selects
varieties, multiplies the basic material, and gives this to SEA; SEA guarantees
approximately 50 pre-approved growers in the San Juan Valley a premium price to
multiply the seed and purchases their seed at harvest"; SEA cleans, treats, and stores the

seed, which at this point is worth close to 8RD 1,400-1,450/qq (Dept. Semillas/SEA,

 

“ SEA purchases seed from contract farmers at about 35 percent above market
prices and 14 percent above the INESPRE price, based on 1998 CSRP Producer Survey
data. SJV contract growers are responsible for close to 80 percent of bean seed
production for the SEA Seed Multiplication Program.

66

1998); SEA sells the seed the next season via CVMA outlets located in many parts of the
country for 3RD 3900/qq. While not “certified,” the seed is considered semilla de buena
calidad (seed of good quality). Recently, the CRSP has given basic seed directly to farm
associations in the SJV for their seed banks (in exchange for use of a member’s land for
basic seed multiplication).

3 The calculated price subsidy of this seed is approximately US$0.36/lb of seed”.
Using data fi'om the SEA Seed Department on total seed distributed per year, combined
with the subsidy per pound of seed, the economic cost of this program is included in the
cost-benefit analysis in Chapter 6. Also using data from the SEA Seed Department on
seed sales by region, and dividing these amounts by an assumed seed rate of 15 lb/ta,
estimated area planted to SEA-direct seed was calculated for each year. This calculation
per region, combined with SEA data on annual area planted by region, shows that each
year, approximately 30 percent of been area is planted with seed purchased that year from
SEA (according to SEA data). As approximately 33 percent of SJV farmers purchase
seed direct fi'om SEA in a given year (Arnaud, 1997), this percentage figure for the
country as a whole seems reasonable.

Problems in recent years with the quality, consistency, and timing of SEA’s seed
sales have led several SJV farmer associations to set up seed banks and store their own
seed. The Seed Department in San Juan allows associations to store seed in special cold-

storage rooms for free; the only condition is that seed deposited at the department may not

 

‘2 The subsidy is calculated as RDSl400/qq economic cost of the seed, less the
RDS900/qq paid by farmers, thus leaving a RD$500/qq subsidy, or US$0.36/1b’. '

67

be removed until November. Thus, association members who lose their crop one year can
often buy seed from their association’s seed bank.

Private sector bean seed suppliers have a brief history in the DR Semillas
Surenas, Inc. was one of three companies that tried to enter the bean seed market in the
early 1990s. They faced the challenge of competing with government subsidized seed
distribution, as well as government subsidized fi'ee cold-storage (in SJV). Semillas
Surenas was primarily a rice seed supplier that ventured into beans when SEA said they
would leave the bean seed market to the private sector. While the company made little or
no profit on bean seed, Semilla Surenas was able to sell bean seed primarily to their rice
seed customers as added business. However, SEA re-entered the market again a year
later, and the three companies summarily exited the bean seed business (Semillas Surenas
went out of business entirely).

4.5.4 Banco Agricola 3

Banco Agricola, the state agricultural development bank, offers subsidized loans
3 for crops and livestock production at 18 percent annual interest. They have charged an
average rate of 19 percent over the five-year period fiom 1993-97 (SEA, 1998), while
market interest rates averaged 26 percent over the same time period - implying a subsidy
of 6-7 percent. This subsidy is perhaps larger when considering that loans for- bean
production alone - apart from other business the farmer may have with the bank — are
often made at a rate higher than the average commercial market rate, reflecting the risk of
been production (Libre, 1998). Loans to bean producers in 1997 accounted for 2,421 ha

of bean area planted (7 percent of total bean area-planted). By comparison, loans to rice

68

growers accounted for 27,723 ha of the area planted (26 percent of total rice area
planted). Of Banco Agricola’s loan portfolio, 3.7 percent went to beans (which account
for 3.7 percent of the total value of all DR crops produced in 1997), while 56 percent
went to rice, which accounts for 25 percent of the total value of all DR crops for 1997
(SEA, 1998). As discussed in more detail in Chapter 5.10, Banco Agricola targets small
to medium-size farmers, most of whom are not able to secure loans fi'om commercial
sources. Thus, Banco Agricola promotes developmental objectives while promoting
domestic production of crops such as rice and beans, which employ many rural farmers
and laborers and provide domestic sources of carbohydrates and protein for consumers.

4.5.5 PROSEMA

PROSEMA, a government agency, provides subsidized land preparation for
various crops across the country. According to; CRSP data, PROSEMA charges roughly
one-half that of private land preparation services, at least in the San Juan Valley.
However, as is noted in the following chapter, farmers who uSe PROSEMA ofien report
problems with delayed service, poor land preparation, or both. According to PROSEMA
data, 37 percent of the land planted to beans in 1997 received some land preparation
service fi'om PROSEMA (SEA, 1998). Roughly two-thirds (68 percent) of these services
were utilized by farmers during September-December, coinciding with the planting periods
in the San Juan and Cibao Valleys. However, the 1998 producer survey showed that
within the San Juan Valley, only 25 percent of farmers contracted with PROSEMA

(CRSP, 1998).

69

4.5.6 INDRHI

INDRIH, the Instituto nacional de Recursos Hidraulicos, is the government
agency that builds and maintains much of the irrigation infi’astructure in the DR. User
rates charged to recipients are subsidized. In the San Juan Valley (SJV), this rate is
US$13.16 per hectare per year (RDSl2/ta), although the on-going JICA/SEA Yaque del
Sur Integrated Rural Development project estimates that a fee of approximately US$31.00
per hectare per year would be required for the SJV Irrigation District to recover all
operating costs (JICA, 1998). The IDB is currently working in the SJV with INDRHI
(through the PRODAS project) to try and improve the infrastructure, as well as the
eficiency of water distribution and management in the valley.

4.5.7 CVMA Input Sales

The government also subsidizes various chemical inputs purchased by farmers at
CVMA distribution centers. HoweVer, while data on amount of product(s) sold is not
available, various sources indicate that CVMA sales represent a very small proportion of
been farmers’ input purchases.
4.6 The Bean Marketing System

4.6.1 Marketing Channels

Most domestic bean growers sell their produce to commercial truckers that come
to the growers’ field at or soon after harvest. However, producers also sell to local
middlemen (suppliers of local credit and/or chemical inputs), INESPRE, SEA, and bean
processing agro-industries. In addition, producers with access to transportation sell

directly to wholesalers or retailers (Heikes, 1993).

70

Commercial truckers are usually based in urban areas, traveling throughout the
country collecting beans either directly from the producers’ bean fields or fi'om the
producers’ homes or storage facilities. In turn, these commercial truckers sell their
purchases to professional traders, or directly to wholesalers or retailers (Heikes, 1993).

Local middlemen usually operate on a smaller scale within a given region. Often
they provide growers with financing or other inputs in exchange for exclusive rights to the
production or a percentage of the production. These regional middlemen usually collect
the. beans directly from the producers’ fields, sometimes storing the beans before selling
them to wholesalers. Some of these regional middlemen also sell to commercial truckers
(Heikes, 1993).

As discussed at length above, the government parastatal IMESPRE is also involved
in bean marketing, although it typically only purchases in the San Juan and Cibao Valleys.
Likewise, the government seed multiplication program run by SEA provides a marketing
option for selected growers (50 or so) in the San Juan Valley and a few in other regions.

The role of farmer associations in the San Juan Valley in facilitating bean
purchases by INESPRE was discussed above in Section 4.5.2.2. While these farmer
associations have other activities common to cooperatives, with respect to output
marketing, these organizations only facilitate sales to INESPRE, not to private buyers or
agribusinesses. One farmer organization in the SJV is investigating export opportunities in
Venezuela and Cuba, and is in the process of moving towards becoming a cooperative,

which may eventually lead to cooperative output marketing to private sector agents.

71

Agra-industries receive their bean supplies directly from producers, as well as
fiom middlemen. In turn, packaged bean products are either distributed to wholesalers or
retailers (Heikes, 1993).

Imported beans mostly come into the country through the private sector, but are
also procured by INESPRE. Private sector bean imports are initiated by bean processing
agro-industries or importers, who are either wholesalers or retailers themselves, or who in
turn sell to wholesalers and retailers. Beans imported by INESPRE are either sold by
INESPRE directly to consumers or are sometimes channeled through wholesalers,
retailers, and agro-industries (Heikes, 1993).

Wholesalers usually have large storage facilities and many actually own and
operate commercial trucking services. These wholesalers store beans, then distribute and
sell to retailers throughout the country (Heikes, 1993).

Retail outlets vary, depending on location. In rural areas, a retailer is typically a
small store located in one room of a house, or a small structure just outside the house
(colmados). However, beans are also retailed in open air central markets (mercado
centrales) in rural areas. In urban areas, although colmados and mercado centrales exist,
beans are mostly sold at mid-sized grocery stores (mercados), and many large, modern,
luxurious, and even air-conditioned supermarkets (supermercados), which ofi‘er a
complete assortment of the world’s food products and consumer goods. Packaged or
processed beans are found in some urban mercados, but primarily in supermercados.
Supermercados in Santo Domingo have bagged domestic beans that are differentiated by

size and quality (i. e. for the Pompadour bean, retail classes include original, select, and

72

super select). INESPRE sells beans directly to consumers (at slightly subsidized prices)
fi'om the backs of trucks or from small vending centers (Heikes, 1993).

. - While the Dominican Republic does not oficially export beans to Haiti, many
commercial truckers from throughout the country, and regional middlemen along the
Haitian border, “unomcially” sell or barter beans to Haitian “importers” (Heikes, 1993).
Various sources also alleged that in recent years, beans donated to Haiti have ofien ended
up entering Dominican markets.

' 4.6.2 Farm-level Bean Storage and Sales

Nearly all respondents (96 percent) to the 1992 nationwide survey said they stored
beans for home consumption. Most stored them in sacks while some use steel or plastic
drums. One-half of all respondents stored beans which they later sell, and 41 percent
stored beans for use as seed (Heikes, 1993). Most of the respondents to the 1992
nationwide survey participated in bean markets, with 88 percent selling some or all of their
bean production. The remaining twelve percent of the respondents grew beans for
subsistence use only. Using SJV growers from the 1998 producer survey as a proxy for
growers in other regions, while these growers may store some beans for seed the
following year and/or for home consumption, those who sell beans typically do so within a
few weeks to a month after harvest. This is not surprising, given that the high costs of
bean production mean thatproducers need cash to pay 011‘ input loans. 3

3While SEA buys between 1,800 to 2,700 metric tons of beans each year for its
seed program, and while INESPRE (in years of operation) may buy up to 3,000 metric

tons of beans, these purchases are made almost entirely in the San Juan or the Cibao

73

Valleys. The implication is that while growers in those valleys may have the option to sell
some or all of their beans to SEA and/or INESPRE”, growers in the rest of the country
sell to commercial truckers and middlemen or local wholesale and retail markets at lower
prices.

4.6.3 Alternative Marketing Institutions

The Junta Agroempresarial Dominicana (J AD),-a agricultural consulting firm based
in Santo Domingo that is often hired by the government, agribusinesses, and even by some
farmer associations, has tried in recent years to move the bean marketing system towards a
more open-market orientation, at least within the domestic market. They bemoan
INESPRE.’s domestic and import purchasing processes as meficient and corrupt
institutions that create more harm than good by insulating bean producers from
competitive forces and allowing part of the benefits of low-priced imported pintos to be
captured by government oficials and importers in the non-transparent licensing process.

To address the uncertainty created year in and year out by INESPRE’s non-
guaranteed purchase quantities and prices, not to mention their late payments, in 1998,
JAD instituted the Bolsa Empresarial, an open trading market to facilitate market
transactions and (hopefully) contracting between producers and private sector buyers of
beans. Although several San Juan Valley farmer associations approved of the initiative in
spirit, in practice they nevertheless sold again to DIESPRE in the spring of 1998. Thus,

the Bolsa was not used to coordinate bean marketing in 1998.

 

‘3 In the San Juan Valley in 1998, 26 percent of growers sold to SEA, 64 percent
to INESPRE, and 28 percent to middlemen (CRSP 1998 Producer Survey and Chapter 5).

74

To address the uncertainty and corruption evident in INESPRE’s import licensing
process, JAD recommended an open permit system in which permits would be sold in
transparent auction to the highest bidder. JAD argues that SEA could use the income
fi'om the sale of licenses to firnd the seed program, bean research, etc. In other words, any
difi‘erence between the world price and the domestic price (minus marketing costs) would
be spent by SEA on productive investments, rather than ending up in the pockets of the
buyers and sellers of import licenses.

4.7. Research and Extension to Improve Bean Varieties and Management Practices

4.37.1 . Bean Breeding and Foundation Seed Production

In the 19605, the first government bean seed processing facility was established in
San Cristobal. In 1971, the SEA’s Department of Seeds was officially established and
given its mandateto regulate the production of all seeds in the country. In 1976, the Seed
Multiplication Program. was created through the small-scale agricultural producer project
funded by a loan from USAID. Part of this loan was used for the construction of
infiestructure, including a storage facility in Juma, a processing and drying facility, and an
experiment station and scientific laboratory (CESDA, located in San Cristobal) (Heikes,
1993)., 3

In 1979, through PIDAGRO 111 (an integrated agricultural development project
that was financed by USAID and the IDB), new and additional equipment was purchased
for CESDA in San Cristobal, and a processing plant and experiment station were

constructed in San Juan de la Maguana (Heikes, 1993).

75

The USAID Bean/Cowpea CRSP began collaborative bean research support to
SEA in 1984, with research centered at CESDA in San Cristobal and CIAS in San Juan/
Arroyo Loro. To date, CRSP support has included funds for research, training,
conference and project travel, and technicians’ salaries; SEA provided researcher salaries,
buildings, land, and vehicles. By the mid-19905, the CRSP-trained DR stafi‘ included two
Ph.D.’s (Dr. Eladio Amaud at Nebraska and Dr. Graciela Godoy at Wisconsin) and a
MSc. (Segundo Nova at Oregon State).

CRSP bean research in the DR has included varietal development (Chapter 5.6.1),
research on optimal planting dates for the SJV (Chapter 5.5.4), and research on cultural
practices, including all aspects of been production, fiom land preparation to post-harvest
and seed storage. The philosophy of the CRSP staff has been to combat the disease and
pest problems facing SJV farmers with both varietal resistance and improved cultural
practices. While CRSP varieties have been difi‘used across the country through the SEA
Seed Multiplication Program, CRSP-recommended cultural practices have been difi‘used
primarily in the SJV through field days and production courses (Chapter 5.3.5) as well as
instructional pharnplets (cited in CRSP Annual Reports). 3
4.8 Chapter Summary

This subsector analysis demonstrates that the high adoption rate of PC-SO is due to
a combination of better field results (yields and uniformity; Chapter 5), continued/
increased government support to bean farmers in the form of negotiated prices above
those of the market, the multiplication and distribution of subsidized PC-SO seeds, and

continuing import protection which maintains high farm-level bean prices. Since 1984,

76

CRSP research has successfully developed yield-augmenting technologies and disease
prevention practices. However, past and existing government policies towards beans have
created a situation in which production and marketing risk has increased, resulting in
declining area planted and hence decreased national production. Future domestic bean
production (and future CRSP impact) hinges primarily on the nature of future government
activities in bean purchasing and marketing, the government’s policies toward import
levels and import administration, and consumer preferences for domestic beans 3with
respect to imported pintos.

In summary, the policy and institutional fi'amework, as defined by the Dominican
government, has promoted the rights of producers to produce beans at the expense of
consumers’ rights to' cheaper beans. In the process, over the past few years, the erratic
and poorly defined policies of the government marketing board, with respect to both
domestic support prices for beans and the timing and quantity of been imports, has created
a large amount of uncertainty in the production and marketing of beans. This uncertainty
is simultaneously creating substantial risk for bean farmers and traders, while providing
avenues for the extraction of rents by select farm and trader groups at the expense of other
farmers, traders, and consumers of beans in general. Existing government policies toward
the bean subsector - coherent or not - are clearly promoting substantial inefficiencies at
both the production and marketing levels.

The past, present, and future aggregate rate of return to bean research and farm-
level profitability in the DR is highly dependent upon the policy environment that shapes

the opportunities and incentives facing Dominican producers and marketing agents. At

77

present, that policy environment attempts to favor producers over consumers of beans.
Yet in the long run, the policies work to the benefit of the few at the expense of the many
producers who grow ever more dependent upon the policy fi'arnework to grow beans for
commercial. purposes. Smaller and medium size producers are both more vulnerable to
lower-priced imports and less able to adjust to price changes either through cost reduction
or substitution of alternative crops. Therefore, they will eventually bear most of the costs

of any move away from the current policy framework.

78

CHAPTER FIVE: BEAN PRODUCTION IN THE SAN JUAN VALLEY
5.1 Introduction

This chapter presents results fi'om two surveys completed by the CRSP in 1998
which focused on technical and financial aspects of bean production in the SJV. Results
from these surveys include a descriptive analysis of the process, costs, and profitability of
bean'production in the San Juan Valley (SJV), both on average and by farm size. The
chapter assesses the profitability of SJV bean production and its sensitivity to changes in
bean yield and price. Given the uncertainty of future government policy regarding input
and output subsidies and the historically high level of government support to bean farmers,
the sensitivity analysis demonstrates the farm-level efl‘ect of various yield and/or price
changes on financial and economic profitability. In addition, the analysis demonstrates
which producers will bear the adjustment costs that inevitably will arise fiom any firture
changes in government policy towards the bean subsector.

In the previous chapter, subsector analysis was employed to explain how benefits
of technological change in bean production are distributed between consumers and
producers. By contrast, this chapter explains how benefits that accrue to producers are
distributed among various types of producers and highlights areas of leverage in which
changes in government policy and/or farm management practices could most easily
increase the profitability of bean production through cost reduction.

Over the last five years, the SJV has produced close to one-half of the country’s
total annual bean production. The valley has traditionally led the country in bean

production and is an important producer of other grains such as rice and maize. In fact,

79

the valley is afi‘ectionately known in the DR as e! granero del sur (the granary of the
south) and the red beans produced there are preferred by a most Dominicans‘. The valley
is almost completely irrigated and is not surprisingly a highly commercial farming
community. Nevertheless, resource distribution in the valley mirrors that of the
agricultural sector across the DR and of Latin America in general; land ownership is highly
skewed, as are other critical factors of production such as human capital and access to
technical lmowledge.’ 3

Using farm size (bean area cultivated) as a proxy for a farmer’s general resource
constraint (including land, human capital, technical knowledge, credit, etc), average asset
levels and profitability of large farmers are compared with that of small farmers. This
comparison is used to highlight how hypothetical variation in prices and yields farmers
would afi‘ect the profitability of farmers with difl‘erent asset levels. Realistic scenarios of
increases in price variation could come fi'om either reduced government bean purchases in
the valley or fiom increased import levels of beans, while variation in yields could result
fi'om existing or future biological pressures and constraints on bean production in the
valley. Finally, the impact of the CRSP on bean production in the valley, as well as the
potential for future impact, is also explored within the context of the aggregate and

subgroup farmer levels.

 

‘ There are some consumers who prefer other types of red beans grown in the
DR such as Jose Beta (red with white speckles) or Constanza (a large red been grown
in the Constanza valley (Mather, 1999). However, the vast majority of consumers in
Santo Domingo prefer the Pompadour Checa variety grown in the SJV.

80

5.2 Data

Results reported in this study utilize two primary data sources. First, during the
1997-98 winter season (N ovember-February), farm record keeping (“RK survey”) data
were collected by CRSP DR stafi‘, supervised by Project PI Dr. Eladio Amaud-Santana.
These data were collected on the primary bean parcel cultivated by 29 irrigated SJV
farmers who grew beans as a monoculture. Data were collected on farmer’s
socioeconomic/household characteristics, bean area that 'season, tenure type, total
production fi'om that parcel, credit amount and sources, and the cost, amount, and type of
chemical inputs, labor and equipment used for each operation from land preparation to
harvest/threshing. Second, retrospective data were collected fi'om the same 29 RK
farmers plus 71 additional randomly selected SJV farmers (for a total of 100) through the
1998 CRSP producer survey (“Producer Survey”)2. The Producer Survey was used to
augment the RK information related to total bean area and yields, land tenure, and credit,
and to add new information regarding production constraints, sales, farmers’ technical
knowledge, and farmers’ perceptions of varieties, mandated planting dates, and the firture
of bean production in the SJV. Whereas the RK survey collected precise input/output
production data from 29 representative farmers, the producer survey collected

retrospective data fiom 100 randomly selected farmers (including producer bean area and

 

2 In this chapter, “RK Survey” will refer to the 1998 Record-Keeping Survey
managed by Eladio Amaud Santana during 1997-98 winter bean season, and
“Producer Survey” will refer to the 1998 CRSP Producer Survey managed by David
Mather in July-August 1998. Dr. Amaud also managed producer surveys in the SJV
following the 1995-96 and 1996-97 winter bean seasons; when statistics are drawn
from these earlier surveys, the reader will be duly notified.

81

yields from the previous 5 years).

This chapter reports estimates of average input use, costs of production, and
profitability for the total RK sample of farms (parcels), and then compares the
performance of small versus large farms. As all farmers in the RK sample (and almost all
farmers in the valley) use improved varieties and fertilizer, variation in yields and
profitability across farmers is due to factors such as level of fertilizer use, technical
knowledge, irrigation quality, land and land preparation resources, scale of bean
operations, access to credit, and price received for output. While some small farmers’
yields are as high as those of large farmers, larger farmers were generally more profitable
due to higher output prices received and lower production costs per hectare.

5.3 Human Capital

5.3.1 Introduction

Farmers difi‘er greatly in terms of their ability to manage their fixed and variable
resources to maximize profit (allocative emciency). Technical knowledge of the bean
production process'is essential for application of fixed and variable resources in
combinations that maximize financial returns per hectare. Thus, technical knowledge
enables a farmer to make optimal input and output choices given his opportunity set.
technical knowledge or human capital can be proxied by various indicators, including
years of experience producing beans, education level achieved, extension advice received,
and field days/Courses attended.

By contrast, socioeconomic status and political afl'mity helps to shape a given

farmer’s opportunity set with respect to their access to low priced, high quality, timely

82‘

inputs, as well as to higher output prices. For example, SJV farmers from the Productores
Agricolas, Inc. or dSOIEPRO producer associations will very likely have contact with
other association members that have university training in the agricultural sciences. Thus,
farmers in these associations likely enjoy access to the human capital resources of an
agronomist fiiend, which are unavailable to a farmer in another association. Although
some SJV producer associations are quite socioeconomically diverse, most associations
attract members with common socioeconomic characteristics or political orientations.
Producer association membership can bring significant benefits to the farmer, such as
differential access to inputs, credit, storage, and marketing opportunities.

5.3.2 Producer Associations

Producer associations can solve collective action problems for farmer groups such
as cost-sharing of high exclusion cost goods (lobbying/negotiation with INESPRE or
SEA), as well as cost-sharing of more easily recuperable costs stemming fiom large fixed
investments (tractors, threshing machines, etc.). By its collective nature, an association
can also facilitate access toother inputs by reducing transaction costs such as when an
association purchases inputs or negotiates Output market prices for its members. In the
SJV, the benefits of association membership are seen in improvements in seed access,
timeliness, and quality via association-managed seed banks, their access to credit fi'om

Banco Agricola’, and their ability to sell grain to INESPRE (discussed below).

 

5’ In most cases, credit from Banco Agricola is facilitated by the association for
each member individually; in other cases, credit is associative or shared.

83

Over 95 percent of been growers in the SJV belong to a producer association (see
Appendix A). This high participation rate facilitates political mobilization and enhances
the producers’ bargaining power with respect to INESPRE during negotiationsto set the
parastatal’s annual purchase price. From the perspective of technological dissemination
and adoption, these associations appear to have played a significant role in the high rate of
adoption of new varieties (PC-50) and associated management practices (mandatory
planting dates) that have been developed and promoted by the CRSP in the DR.

5.3.3 Socioeconomic Characteristics of Producers

Using education as a proxy for human capital, Table 5.1 demonstrates that the
average valley producer completed primary school, while the average small producer has
less than halfthe education (4 years) of an average large producer (8.9 years). Therefore,
farmers acr055 the valley have different aptitudes for the adoption of technologies such as
improved management practices, which require human capital as a complement.

With respect to gender, of the 100 respondents selected randomly fi'om farmer
association lists, only one was female. However, while women in the valley are not
involved in farmer associations‘ nor in most aspects of bean production in general, they
are the principal providers of labor in drying sorting cleaning, and bagging harvested
beans. More significantly, they most often are the household member who purchases

beans from local markets and prepares the household meals. Therefore, if we consider the

 

‘ While it is possible that a household male could merely serve as the
“negotiator” of the harvest (via the association) while a woman manages the bean
production, this is a very atypical situation, according to key informants - one of
whom is one of the few female bean farmers in the valley.

84

process of bean production as encompassing all activities that create utility fiom the farm
level to the consumer’s plate, then the women’s role in this process is primarily at end of

the production chain - at both the retail and food preparation levels‘.

100 42 32

48 51 45
6.3 4.0 7.2
6.4 6.8 6.0

  

' Defined as the number of dependents currently living within the house.
Source: 1998 CRSP Producer

 

5.3.4 Extension Advice Received

Given the importance of technical knowledge as a determinant of both technology
adoption (Feder, 1985), yields, and net returns to bean production, various survey
questions were used to investigate the sources of technical assistance sought and received
by bean producers in 1998. First, respondents were asked whom they generally contacted

first for been production advice. Second, they were asked whom they had actually

 

’ Preliminary results fi'om a consumer survey in the DR (Mather, . 1999) show
that consumer preferences for beans are based on a combination of various product
attributes, including freshness (taste) as well as cooking time (fresher beans cook
faster), for which womens’ demand as an product attribute would be expected to
increase as household income increases. Therefore, the opportunity sets of women
can have a profound efi‘ect on the demand for bean attributes, such as fi'eshness and
cooking time, and thus indirectly the profitability of bean production at the farmlevel.

85

contacted in the last year for advice and for what problem(5). Third, they were asked how
many times during the past year they were visited by the government (SEA) extension
agent assigned to them. Finally, the respondents were asked about their attendance at
bean prodl'rction courses or field days during the past year, as well as who sponsored or
led the courses/field days.‘ The results of the first two questions are presented in Table

5.2.

0
13

‘Ag Extension component of PRODAS (IDB project)
Source: 1998 CRSP Producer

 

 

‘ Technical assistance in the SJV is supplied by SEA (the government
extension service), the CRSP, input dealers, FDD (Fundacion Dominicana de
Desarrollo), EYCA (the agricultural extension component of PRODAS, a large IDB
irrigation project in the SJV), and IAD (the agrarian reform institute).

86

3 , From the results in Table 5.2, about one-half the producers generally first contact
their SEA ExtensiOn agent for technical assistance, while 13 percent report that they don’t
askanyone for advice. When the respondents were then asked who they actually
contacted first in 1998, twenty-seven percent reported contacting SEA, 11 percent
contacted "the CRSP, and 46 percent did not contact anyone. Of those respondents who
sought technical assistance in 1998, 43 percent were interested in information regarding
fungicide applications for rust, melds, spots, antracnosis, and other diseases; 30 percent
asked about BGMV, 21 percent asked about early defoliation, 19 percent asked about
insecticide applications, 9 percent asked about crop development, and 2 percent asked
about irrigation and planting, respectively.

3 When asked; about the number of extension contacts made in 1998 by SEA
extension agents, 42 percent of the respondents reported not receiving any SEA extension
visits during the past year, 13 thirteen percent reported'being visited once, 20 percent
were visited twice, and 27 percent received three or more visits. Key informant interviews
indicated that SEA extension quality and consistency depends greatly on the agent in
question. In general, low extension salaries and scarce transportation firnds were often
cited as a primary constraint to better extension.

In addition to extension visits, SJV farmers also gain technical knowledge by
attending field days and/or production courses offered by many of the same sources of
technical assistance. In 1998, forty-eight percent of the respondents attended a field day,
and 39 percent attended a production course. The field day and course sponsors are

reported in Table 5.3.

87

 

Table 5. 3 Sponsors of Bean Production Courses and Field Days, SJV,1998

 

 

 

 

 

 

 

 

 

 

" Responses of 48% of respondents who attended a field day in 1998

" Responses of 39% of respondents who attended a production course in 1998

° Federacion Dominicana de Desarrollo ‘ Ag Extension component of PRODAS (IDB
project) Source: 1998 CRSP Producer Survey

 

 

 

While it is difficult to determine the sponsor of the events indicated by the
response “farmer associations,” it is very likely that these are CRSP-organized events
(with certainty, this applies to the field days). If these events are attributable to the CRSP,
then CRSP-organized field days reached 57 percent of the valley’s farmers (87 percent of
the 66 percent of the total sample who attended a field day in 1997-98), and CRSP-
organized production courses reached 9 percent of the valley’s farmers (22 percent of the
39 percent who attended courses in 1997-98). By contrast, 17 percent of valley farmers
attended production courses sponsored by input dealers (44 percent of the 39 percent who
attended production courses in 1998), which suggests that input dealers have considerable
influence regarding farmers’ input use in the SJV.

5.3.5 CRSP Extension of Technical Knowledge

From the statistics presented above, it is apparent that more than half of the

valley’s farmers in the SJV have attended field days or courses led by the CRSP, and 11

88

percent report that they contacted the CRSP first for technical advice in 1998. In
addition, of the farmers who report going first to their association for technical assistance,
it is very probable that the information received there originated fiom the CRSP. In recent
years, the CRSP scientists in the DR have clearly helped to compensate for the DR’s
inefiiectiv’e public agricultural extension system by extending new bean production
technologies vis pamphlets, courses, and field days organized for (and sometimes by)
farmer associations and womens’ producer groups. This efl‘ort by the CRSP is aided
significantly by farmer associations, which help lower the transaction costs of information
dissemination from the research station to farmers by providing the CRSP a forum fi'om
which to reach many members at the same time. And because those members are apt to
share the bonds of association, it is likely that CRSP recommendations have been firrther
disseminated among members who did not actually attend a field day or course.
5.4 Bean Cropping Patterns and Land Use

. ‘ 5.4.1 Farm Size

Bean farmers in the SJV are-relatively large, compared to other countries in the

Central America/Caribbean region. The average total bean area per farmer in the SJV is
8.2 ha (CRSP, 1998). However, this mean value in the SJV belies great variance in total
bean area, both within the DR and within the SJV. Within the 1998 CRSP producer
survey of 100 farmers, total bean area in the SJV varied fi'om 0.4 to 100 ha. This study
classifies farmers into three categories: small (< 2.5 ha), medium (2.5 to 6.3 ha) and large

(> 6.3 ha).

89

While aggregate bean area in the San Juan Valley has decreased since 1993, it
appears that farm size may have increased. According to the 1998 CRSP Producer
Survey, during this five-year period, the share of small farmers dropped fiom 45 to 36
percent, while the share of medium farmers inoreased from 27 to 33 percent, and the share
of large farmers increased from 27 to 30 percent’.

Table 5.4 Characteristics of Sam led Bean Producers b Bean Area SJV 1998

 

 

 

 

 

 

 

 

Source: 1998 CRSP Producer Survey

 

5.4.2 Land Tenure

Five types of tenure exist in the SJV, as demonstrated in Table 5.5. Nearly one-
halfof the respondents are owner operators (48 percent), while 41 percent sharecrop, 38
percent receive land fi'om the state’s agricultural land reform program, and 11 percent rent
land. However, farmers typically cultivate beans under more than one form of tenure.
The land share portion of Table 5.5 demonstrates that land distribution in the SJV is highly

skewed. For example, farmers with land from the land reform program (38 percent of the

 

7 However, it is possible that this reflects selection bias in that only current
bean farmers are respondents. Thus, farmers who were small or medium in 1993 may
have expanded their operations in the past five years and become medium and large
producers respectively.

90

respondents) cultivate only 6 percent of the bean area in the sample. In addition, while
one-half of respondents own some of the bean area they cultivate, in fact, most of the land
‘ in the sample is sharecropped (50 percent), while 37 percent is owner-operated, 6 percent
is rented, 6 percent is associated with the state agricultural land reform program, and 1
percent is loaned (free of charge). For each tenure class the average percentage of land
within that class operated by the farmer varies considerably (Table 5.5). For example,
land owners (48 percent of the sample) own an average of 65 percent of their total bean
area, while sharecroppers (41 percent) sharecrop an average of 75 percent of their total

bean area.

48 37
41 50
11 6
38 6
3 1

' Figures do not sum to 100 because many farmers have bean area in more than one tenure class. Land
Share is calculated as the percentage of a farmer’s area in a given tenure class, weighted by the ratio of
the farmer’s total area divided by the total sample area. Thus, Land Share sums to 100.

Source: 1998 CRSP Producer

 

When Land Tenure statistics are firrther analyzed by farm size (Table 5.6), it is not
surprising that only 23 percent of small farmers own land and that 67 percent of small
farmers cultivate land reform land. However, while 65 percent of large farmers own land,

they only own an average of 47 percent of their total bean area. This is because 81

91

percent of large farmers sharecrop some land, with their average sharecropped land
accounting for 75 percent of their total bean area. Similar results are found for medium
farmers, although those who do own land tend to not sharecrop as much as the large
farmers. These results suggest that land ownership in the SJV rs highly concentrated, and
secondly, that small farmers are not likely to be able to expand their operations due to
resource constraints. This second result rs not surprising, given the high production costs
of beans relative to other crops, including high labor costs and mechanical land

preparation.

Table 5.6 Land Tenure of Sam led Bean Producers b Bean Area SJV 1998

 

 

 

 

 

 

 

 

' This should be interpreted as “0f farmers who own land (48%), land owned accounts for an average
of65% of their bean area” The same interpretation applies for the tenure type in the three subsequent
rows.

Source: 1998 CRSP Producer Survey

 

Within the sharecropping tenure class are three different arrangements that difi‘er in
terms of the sharing of benefits, costs, and risk between the landowner and sharecropper.

Eighty-five percent of sharecroppers farm a la quinta, 5 percent farm a la cuarta, and 10

92

percent farm a la media. Under a la quinta, by far the most common sharecropping
arrangement, the sharecropper pays all bean production costs and gives the landowner 20
percent of the harvest'. Thus, sharecroppers under this system bear almost all the risk of
bean production. Among sharecroppers, an average of 75 percent of their bean area was
sharecropped, with only 16 percent owned. .

Thirty-eightpercent of the respondents were afiliated with the government’s land
reform program (IAD), whereby they received use rights (but not ownership) to farm 1.5
to 2.5 ha of irrigated land (25 to 40 tarea). Bean area under IAD accounts for about 6 to
7 percent of the total bean area in the SJV (JICA, 1998 and SEA, 1999). Although 71
percent of farmers affiliated with IAD cultivate only land fiom IAD, there is no restriction
preventing them from owning, sharecropping, or renting additional plots. In fact, 29
percent of IAD farmers retained use rights to IAD land yet own or sharecrop other land
for bean production. Of those using IAD land, an average of 86 percent of their bean area
is IAD land, with 7 percent sharecropped and 4 percent owned. Thus, for the majority of
IAD farmers, the land they receive from IAD represents nearly all of the land that they
cultivate. -

5.4.3 Irrigation

Although all bean farmers in the valley cultivate under irrigation, irrigation

eficiency (including both the timing and quantity of the nearest canal as well as the

 

' A la cudrta is similar to a la quinta except the tenant gives the landowner 25
percent of the harvest. A la media allows for some cost-sharing as the landowner pays
for land preparation, the sharecropper bears the costs of planting, and both share the
remaining costs and revenues half and half.

93

farmer’s water management) are also important yield performance factors. This survey
focused only on the farmers’ opinion of his/her water constraints by asking respondents
about the suficiency of their irrigation on each of their bean fields. Twenty-one percent
of respondents report “not having sufficient water”, with this constraint affecting an
average of 90 percent of their total bean area. Forty-seven percent report “usually having
suficient water”, with this condition accounting for an average of 88 percent of their bean
area. Finally, forty-four percent report “always having sufiicient water”, with this
condition accounting for an average of 84 percent of their area’. Analysis of water
sufficiency of fields by farm size and by tenure of each field indicated that there are no
significant difi‘erences in the frequency of water constraints by farm size or tenure.

These responses imply that 19 percent of the valley’s bean area does not receive
srificient water (21 percent times 90 percent), 42 percent usually receives sufficient water,
and 37 percent always receives sufficient water. Therefore, it is clear that although all
farmers have access to irrigation, there are substantial differences in the timing and quality

of access to that vital resource”.

 

’ These three figures do not sum to 100 because farmers were asked about
water sufliciency on each of their various bean fields. Because many farmers have
multiple fields that varying in irrigation quality, many farmers may have a certain
percentage of their bean area under two or three of the constraint categories.

‘° The principal objective of the on-going PRODAS project (Inter-American
Development Bank) is to improve irrigation efficiency and water distribution in the
valley by strengthening infrastructural and managerial capacity.

94

5.4.4 Cropping Systems

In the San Juan Valley, beans are grown almost exclusively in the winter season in
rotation with a variety of annual summer crops. During the winter 1997-98 season, only 9
percent of the respondents cultivated a crop in addition to beans. These additional crops
were sweet potatoes (5 percent of respondents), cassava, peppers, rice, and pigeon peas (1
percent each). Summer annual crops cultivated by bean farmers are typically maize, rice,
sweet potatoes, pigeon peas, and cassava (Table 5.7). In general, larger farmers in the
SJV grow rice, sweet potatoes or pigeon peas in the summer. While small farmers grow

primarily maize in the summer, some also plant rice, pigeon peas, or cassava.

 

Source: 1998 CRSP Producer

95

5.5 Land Preparation and Planting Dates
5.5.1 Land Preparation

‘ In the SJV, land preparation in carried out fi'om July through November. While
some farmers begin land preparation in July, most start in August or September. Farmers
complete fiom three to eight land preparation operations, although the average is five
(RK, 1997). Farmers vary in the number of operations, due in part to variance in both
technical knowledge of land preparation and access to tractor services. Sixty percent of
farmers used only a tractor for their various land preparation operations, while 37 percent
used both a tractor and animal traction, and 3 percent use only animal traction. Almost all
(88 percent) of the tractor users rent these services from either the government
(PROSEMA), their farmer association, or private contractors. A few farmers (7 percent)
owned their own tractors, and 2 percent borrowed a tractor from a family member or
fiiend. Farmers who rented utilized several different sources, including private companies
(29 percent), farmer associations (27 percent), PROSEMA (24 percent), individuals (18
percent) or IAD (2 percent). Although PROSEMA’s tractor service rates were subsidized
(costing about 40 percent less than market rates), survey and rapid appraisal interviews
indicated that PROSEMA’s tractor services are unreliable in terms of punctuality and
quality. For this reason, 23 percent of respondents using tractors chose a mix of rental

services fi'om two or more sources“.

 

". Several farmers who rented tractor services fi'om PROSEMA reported that
the service was of such poor quality that they either paid the PROSEMA operator to
redo the same operation or they simply hired a private operator to do so.

96

While most farmers were satisfied with the tractor services that they hired, 11
percent of respondents said that tractor services were not available when needed, and 15
percent reported that they were not satisfied with the quality of the service. Respondents
were most dissatisfied with' the PROSEMA service, which accounted for 46 percent of the
complaints about poor service availability and 46 percent of poor quality complaints
(although PROSEMA accounted for only 28 percent of tractor rentals). Association
tractors accounted for 36 percent of reported availability problems but only 15 percent of
quality problems. By contrast, private tractor operators accounted for 10 percent of
availability problems and 36 percent of the quality complaints. By far, the most common
quality problem was simply poor depth of cut (89 percent of the 15 percent who
complained of poor quality).

5.5.2 Fallow Period and Planting Dates

Since the occurrence of a serious outbreak of BGMV in the SJV during the 1989-
90 bean season, planting dates have been a very serious economic and political issue".
Research carried out by CRSP scientists showed that the valley’s whitefly vector increases
in the late fall (August through October) when there are ample host crops, primarily beans,
peppers, tomatoes, and melons. Thus, in an efi‘ort to reduce whitefly populations, in 1990,
CRSP researchers recommended that the regional government impose both a fallow

period and a planting date period. CRSP scientists reasoned that the fallow period would

 

‘2 According to several sources, the whitefly population in the Southwest
region exploded in the late 19803 due to the promotion of melon production in Azua,
(an area to the southeast of the San Juan Valley) by a USAID export promotion
project. ‘

97

eliminate the host crops on which the whitefly reproduced in the months prior to the main
bean season, thus suppressing the whitefly vector until at least late December, afier which
most farmers’ beans would no longer be vulnerable to the whitefly". In 1991, the
government implemented this advice by initiating a fallow period which prohibited the
cultivation of host crops in the valley fi'om 1 August to 15 November.

Concurrently, the government implemented a planting period which was intended
to work in tandem with the fallow period; beans could only be planted in the SJV between
15 November and 10 January. This planting period ensured first that no farmer would
plant prior to 15 November - and thus risk serving as a breeding ground for the whitefly
that could then spread across the valley - and that no farmer would find themselves with
tire-flowering beans in mid-January, by which time the whitefly population was expected
to be rejuvenated. As a result of these regulations, farmers were restricted to growing a
single bean crop“. Therefore, the benefit of maintaining control of BGMV in the valley
for the large winter season came at the cost to some farmers, who lost the opportunity to
cultivate two crops of beans; one in the smaller fall season, and one in the primary winter
season.

These decrees and their subsequent enforcement (violators’ fields were plowed

under almost immediately by the army) were vital to the success of efforts to control the

 

‘3 The whitefly, which may carry and transmit BGMV, can only transmit f
BGMV to a given bean plant up until it flowers.

“. Prior to the initiation of the fallow period, some farmers grew two crops of
beans fi'om August-March, although the majority only grew beans in the winter season
(N overnber/December - February/March).

98

whitefly vector. The decrees were deemed necessary due to the high information costs of
discerning the origin of whiteflies in the valley and the potentially disastrous externalitites
involved — farmers who grow a host crop during fallow period enable the whitefly to
multiply and to potentially infect other farmers’ fields with BGMV. Although some
fanners violated these regulations in the first few years, most farmers have generally
followedthe fallow period and planting dates - in large measure due to the ability of
farmer associations to both disseminate information about the whitefly and to enforce
compliance by its members and those of other associations. Clearly, the economies of
scale embodied in the farmer associations have helped to lower the'transaction costs of
both information dissemination concerning the regulations (the value and rational behind
the decrees) and enforcement, of these regulations.

‘ Because it appeared that farmers who planted earlier in the planting period tended
to avoid serious BGMV attacks, in 1993 -94, CRSP researchers carried out experiments to
isolate the effect of planting date on yield. When these trials confirmed that the optimal
planting period opening and closing dates should be earlier, the CRSP researchers advised
the regional government to move the planting date period to 5 November to 15 December.
Prior tohthe 1994-95 season, the government acted on this advice, moving the planting
date period back'as recommended by the CRSP. In 1994-95, as a result of both the new
planting dates and CRSP-recommended cultural practice, crop losses in the valley due to

BGMV declined to only 10 percent (CRSP, 1994).

99

5.5.3 Planting Equipment
, While most farmers used a tractor for land preparation, planting is still carried out

using traditional technologies. Nlnety percent of respondents planted with cabaIIo y
sembrqdora (horse and planter), one percent planted with a tractor and planter, and 5
percent with a donkey and planter.

5.5.4 Farmer Opinions on Planting Date Period

Given a continuing political debate in the valley over the timing of the start of the
planting period, respondents were asked if they felt that the beginning of the planting
period should be modified (to a date earlier or later than 5 November), what date would
be better, and why. Forty-eight percent reported that they wanted to keep the starting
date as it is. Some of those who wish to maintain the current starting date commented
that the date should not be changed until there is more time to study the results.

. However, fifty-two percent of the respondents said they would prefer to modify
the current 5 November planting date. Of these respondents who desired a change in the
start of the planting period, 64 percent prefer an earlier starting date (before 5 November),
while 36 percent prefer a date after (Table 5.8). It should be noted that 69.6 percent of
those who suggest modifying the starting date suggested an optimal date within 6 days of
the current starting date. Therefore, the current starting date of 5 November does not
appear to be significantly out of line with the opinion of even those who prefer to modify

it.

100

Table 5. 8 Farmers’ Recommended Starting Dates for the Winter Bean Planting

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Period, SJV, 1998
2 3.8% 10 to 20 October
9 17.3% 25 to 28 October
5 9.6% 30 October
17 33.0% 1 or 2 November
' 12 27.0% 7 to 10 November
5 ' 9.6% 12 to 15 November
‘ 1 1.0% 5 December
' Percentage of respondents, out of the 52 who prefer an earlier/later starting date for the bean planting
E”:Iglaot‘esarereportedinunequalincrementsoftimeinordertodemonstratethatthereliponsesrivere
clumped around a few dates.
Source: 1998 CRSP producer survey

 

Of the 33 respondents preferring it earlier starting date, 46 percent said that
planting earlier would enable farmers to further reduce insect and disease problems, 33
percent simply said that “planting earlier is better,” 12 percent said that planting earlier
would enable them to take better advantage of the late October rains, and the remaining 9
percent suggested that the date should be earlier to “try something new.” Of the 19
respondents preferring a later starting date, 61 percent simply said that “planting later is
better,” 25 percent said that planting early results in more disease problems, 16 percent

cited the need to avoid heavy rains, and the remaining 16 percent said that it’s too hot in

early November to plant.

101

5.6 Use of Improved Bean Varieties

5.6.1 Improved Bean Varieties and Yields

5.6.1.1 Histgg of the Dmlgpment QfPC-SO

In 1983, the Loyola Polytechnic Institute began efi‘orts to collect local germplasm
when they acquired 250 lines of Pompadour Checa fi'om commercial producers and
subsequently grew out and characterized the lines. In 1984, the CRSP began collaborative
research with SEA in the DR, with Ing. Freddy Saladin serving as PI. From these 250
lines, CRSP researchers identified the 18 most promising lines. Of these, PC-50 and PC-
157 were selected for yield testing throughout the country, to be compared with two lines
fi'om CIAT (BAT-.1412 and DOR-198), as well as a popular Pompadour Checa line
(Guerrero, 1987). From 1984-85 to 1987-88, screening continued, with 10-15 trials
conduCted each winter season. The primary selection criteria were grain type (color, size,
growth habit), number of seeds/pod and pods/plant, and resistance to Rust, Common
Blight, and Web Blight. Promising best selections were eventually tested on-farm (semi-
commercial tests). I

While the CIAT lines were higher yielding than PC-50 and PC-157, the CIAT
materials hada tough seed coat and were too dark red in color. In 1986, PC-50 was
identified as the best selection and registered. In the same year, the CRSP supplied
approximately 930 kg of PC-50 foundation seed to the SEA Department of Seeds to be
used in the SEA Seed Multiplication Program. In 1997, SEA multiplied PC-50 by
incorporating PC-SO into its Seed Multiplication Program, which contracts farmers in the

Aura, San Juan, and I-Iiguey valleys to' multiply the seed. SEA first sold P050 to farmers

102

across the DR for the 1988-89 season. By the 1989-90 season, PC-50 accounted for
roughly 50 percent of SEA’s bean seed sales.
5.6.1.2 PQ-_5_Q
In experimental and semi-commercial trials conducted by CRSP researchers

between 1984-1989, PC-50 yielded an average of 1,585 kg/ha, 45 percent more than the
Pompadour Checa commercial selections grown by SJV farmers at the time. When tested
in six regions to compare the various attributes of P050 versus Pompadour Checa (and
other red varieties), PC-50 averaged a 30 percent higher yield than Pompadour Checa. In
addition, PC-50 showed tolerance" to rust and common web blight, heat, drought, and
low fertility soil. Finally, due to its early and uniform maturity, PC-50 grain was expected
to have a higher commercial value. ~ .

i The experimental results fi'om 1984-88 were largely confirmed by the 1998 CRSP
Producer Survey: '76 percent of the respondents reported obtaining higher yields with PC-
50 (compared to Pompadour Checa), 13 percent reported obtaining the same yield, and 10

percent reported obtaining lower yields. Of the farmers reporting a higher yield fi'om PC-

 

” “Resistance” means that a particular bean variety is completely resistant to
the disease and no disease will develop on that variety. “Tolerance”.implies that the
disease will attack the variety, but the variety will show less disease development or
less yield loss than a susceptible variety. “Highly tolerant” is an arbitrary term
implying a greater level of tolerance (i. e. less disease or less losses to disease with that
variety). In the case of rust on beans, there are varieties that are completely resistant,
but if a new race or strain of rust appears that variety may prove to be susceptible and
many such cases have been documented. Varieties like PC-50 have a more
generalized form of resistance that does not prevent disease development, but
dramatically reduce the incidence compared with a susceptible variety. These varieties
are referred to as tolerant or highly tolerant, depending on the viewpoint of the
researcher, breeder, agronomist, or pathologist (Kelly, 1999).

103

50, 72 percent said that their yield increased by 30 to 60 percent“.
1 PC-SO yields reported by the 1998 CRSP Producer survey respondents from the
1993-94 through 1997-98 winter seasons are shown in Table 5.3. Farmers’ yields over
the 1993/94 - 1997/98 period averaged 1,107 kg/ha, compared to an area-weighted mean
yield of 1,204 kg/ha during the same period. It is not surprising that the areaeweighted
yield is higher, given the asymmetric land distribution' in the valley and the fact that larger
farmers consistently have the highest yields.
1993-97

-50 in San Juan

 

   

1,082 1,126 1,536
1,173 1,223 , 1,507
1,166 ' 1,262 , 1,224
1,023 1,265 1,267
1,091 ' 1,143 , 1,166-
1,107 1,204 949 1,340

Source: 1998 CRSP Producer Survey (total N=100; small farmer N=3 7; large farmer

 

"‘ It should be noted that the yield increase (30-60 percent) reported by
respondents is subject to rounding errors. Because farmers in the DR estimate their
yields in terms of quintals/tarea, which typically range from 0.5 and 2.5, asking how '
much a farmer’s yield increased (in terms of quintals/tarea) exposes the response to
rounding error. In addition, while farmers were asked about their yield history back to
1993, by 1998 almost all of the farmers in the SJV had already adopted PC-SO. Thus,
it was impossible to collect accurate data on farmers’ Pompadour Checa yields.

104

When asked about varieties used prior to PC-50, respondents reported planting
Pompadour Checa (77 percent), Jose Beta (9 percent), Negra (6 percent), Brasilena (2
percent), Constanza 1, Indiana, and Blanca (1 percent each). Respondents were then
asked to identify advantages or disadvantages associated with PC-50, compared to their
previous variety. Ninety-one percent of the respondents saw advantages, 52 percent saw
disadvantages, and 8.7 percent saw no advantages or disadvantages of PC-50. Of those
who saw advantages, 76 percent reported that it “yielded more,” 49 percent. said that it
was “more resistant,” 17 percent said that it commanded a “better price/market,” 8
percent reported “better uniformity,” and 5 percent said it has a “shorter growing cycle.”

A Of those who saw disadvantages to planting PC-50, 18 percent said it is “less
resistant,” 18 percent said it has a “bad price/market,” 5 percent said it has “degenerated,”
5 percent. said it requires “more inputs,” and 4 percent said it “yields less.”

- 5.6.1.3 m 1

While PC-50 is the most commonly-grown variety in the SJV, a newly released
variety - JB-.l78 - is rapidly gaining popularity. JB-178 is a cross between cxrsos and
Jose Beta (a local red variety). In experimental and semi-commercial trials by CRSP
researchers between 199341995, JB-178 yielded 1,446 kg/ha. In six regional trials
designed to compare various attributes of JB-178 versus PC-SO, JB-l78 averaged 3
percent higher yield than PC-50. However, the advantages of JB-l78 are embodied not in
higher yield but in high resistance to rust, tolerance to drought and low fertility soil, early
maturity and uniformity, and — perhaps most important for growers - seed color and

culinary qualities (faster cooking) from Jose Beta that command a higher commercial

105

value (approximately 5 percent higher market retail value." The CRSP’s planting and
fertilizer recommendations for JB-l78 and PC-SO are identical. JB-178 has the same cycle
of 75-80 days, although it flowers a little later than PC-50 (30-33 days as compared with
28-30). Although both JB-178 and PC-50 are susceptible to BGMV, JB-l78 is resistant
to rust while PC-50 is only tolerant.

JB-l78 was first tested semi-commercially by SJV farmers in 1995, and was then
multiplied in 1996 by some SJV growers through SEA’s seed multiplication program.
Thevariety was omcially registered and released in 1996, and was first sold by SEA to
farmers in 1997. According to the Producer Survey, 41 percent of SJV farmers have tried
JB-l78, and 34 percent grew JB-178 in the 1997/98.

1 Respondents who had planted JB- 1 7 8 were asked to identify its advantages and
disadvantages. Nmety-three percent of these farmers reported advantages, including: high
yield (45 percent), good market (23 percent), more resistant (13 percent), good culinary
qualities and taste (12 percent), and good color (2 percent). Sixty-eight percent of those
who have tried JB-178 saw disadvantages, including: less resistant (44 percent), low yield
(25 percent), bad uniformity (6 percent), uses more inputs (6 percent), bad color (2
percent), and a longer cycle (2 percent).

In 1997/98, farmers’ yields of JB-178 averaged 1,261 kg/ha, compared to an area-
weighted mean yield of 1,332 kg/ha during the same period. As with PC-SO, area-

weighted yields are higher because larger farmers usually obtain higher yields. While IB-

 

" SEA/INESPRE pay the same for JB-178 or P050 seed/grain, although the
former is usually 5 percent more expensive in the SJV market.

106

178's yield performance in 1997 is better than that of PC-50, it should be noted that JB-
178 was only sold by SEA to all valley farmers in 1997/98. Thus, few farmers outside of
the seed multiplication program have grown this variety for more than one year, and those
non-seed program farmers who adopted in 1997/98 are very likely higher-resource farmers
who learned about JB-l78 via contact with the CRSP. In fact, the mean yield in 1997 of
non-SEA seed program farmers was 1,178 kg/ha, compared with 1,371 kg/ha obtained by
the seed program farmers the same year. Nevertheless, JB-178's yield performance,
combined with strong market demand for the Jose Beta culinary attributes, makes it an
attractive and valuable addition to the mix of Dominican red bean varieties available to
farmers. .
5.6.1.4 W
. ' . The CRSP developed and tested the white variety Anacaona, a multiple cross of

(2b-5-1/2 x Nep-2/Black Turtle Soup) x BON 355, which was crossed at Michigan State
University in 1984. Anacaona is a high-yielding, stable white variety with tolerance to
rust, web blight, common blight, heat, and drought. It was registered in 1992-93 and
included in the SEA seed multiplication program. However, it is not Widely grown
because white varieties are not very popular among Dominican consumers, who strongly
prefer reds. .

Several other varieties have been developed and tested by the CRSP in the 1990s,
including CIAS-95 (red), Saladin-97 (red) and Arroyo Loro Negro (black). While these
varieties have been registered and released in the last three years, they have not yet been

incorporated into SEA’s seed multiplication program. CIAS-95, a cross between PC-50 x

107

BAT-1274 which was made by Jim Beaver in Puerto Rico, is more tolerant to mst than
PC-50, is a darker red, matures one week later than P050, and out-yielded PC-50 an
average of 3 percent across six regional locations. Saladin-9 7, a cross between PC-50 x
BAT-1274, has a higher tolerance to rust than PC-50 and yields on average 6 percent
higher than PC-50 across six regional locations.

5.6.2 Distribution Channels of Bean Seeds

As discussed in Chapter 4.6.3, the CRSP and SJV growers play important roles the
SEA seed multiplication program. According to the 1997 CRSP producer survey, 58
percent of SJV farmers used their own stored seed in the 1996/97 season, 34 percent
purchased seed from SEA, 4 percent fi'om another farmer, and 4 percent fiom a
warehouse (CRSP, 1997).

5.6.3 Seed Cost

Prices paid by farmers for bean seed ranged from RDS900/qq for SEA seed to
about RD$1,000 for seed purchased fiom farmer associations". Farmers in the RK survey
who reported using own seedwere assumed in the cost of production analysis to have
seed costs of RD$1,000/qq. .

5.6.4 Seed Density

Data fi'om various surveys indicate that SJV farmers use more seed than is
recommended by the CRSP. The 1998 Producer Survey respondents used an average of

132 kg/ha (18.4 Wu), and the RK farmers used an average 136 kg/ha of seed (18.9 lb/ta),

 

" Farmer associations have access to free cold storage in the San Juan
Department of Seeds storage facility. Thus, the association md price does not
include the physical cost of storage from February to November (8 to 9 months).

108

which is higher than the 94 to 108 kg/ha (13 to 15 lb/ta) recommended by the CRSP.
Similarly, the 1997 CRSP Producer survey (Amaud, 1997) found that more than one-half
of the respondents used more than 108 kg/ha (15 lb/ta). Dr. Amaud attributed this
primarily to the fact that close to one-half of the respondents used discs that left a short
distance between rows planted, resulting in a high seed density. However, since seeds
constitute the primary purchased input cost component (Table 5.10), lowering the seed
density could. efi‘ectively lower input costs per hectare without threatening yield.
5.7 Chemical Inputs
‘ 5.7 .1 Fertilizer

Fertilizer was used by nearly all (94 percent) of the 1998 CRSP Producer Survey
respondents and constituted the second largest purchased input cost component in bean
production". Although the 29 RK farmers applied twelve different basal formulations, the
most popular were 16-20-6 (12 farmers) and Urea (12 farmers). For the total RK sample,
farmers spent an average of $121.41/ha on basal fertilizer and applied an average of 447.6
kg/ha (Table 5.4). The unit cost of 16-20-6 averaged $1.95/kg and ranged fi'om 81.73/kg
to $2.21/kg, while the unit cost of urea averaged $2.26/kg. Seventy-seven percent of
fertilizers were purchased from private input companies such as F ersan or F erquido.

Twenty-one RK farmers used foliar fertilizers in addition to basal. Six difi‘erent

foliar formulations were used by RK farmers, the most popular being a generic foliar mix

 

‘9 References to “valley farmers” are drawn fi'om the 1998 CRSP Producer
Survey, whereas references to RR farmers are drawn from the 1998 RK Cost of
Production Survey. While the Producer Survey collected data on frequency and level
of input use, the RK Survey is the sole source of primary cost of production data.

109

(10 farmers). Farmers applying foliars spent an average of $7.82/ha and applied an
average of 79 grams/ha.

. Total equivalent nitrogen, phosphate, and potassium fi'om aggregate basal and
foliar fertilizer application averaged 81.2 kg/ha of nitrogen, 65.2 kg/ha of phospate, and
20.3 W of potassium for the RK farmers. For small farmers (<2.5 ha), these equivalent
averages were 74.2 kg/ha of nitrogen, 58.5 kg/ha of phosphate, and 23.7 kg/ha of
potassium. In contrast, for large farmers (>63 ha), who applied about 3.0 percent more
nutrients than small farmers, these equivalent averages were 100.51 kg/ha of nitrogen,
70.51 kg/ha of phosphate, and 27.0 kg/ha of potassium.

5.7.2 Insecticide

Nearly all farmers (98 percent) used insecticide, averaging 3 .0 applications per
farmer (1998 CRSP Survey). Although RK farmers used 23 different insecticides, the
most pOpular were Nuvacron (9 farmers), Decis (8), and Sistemin (6). RK farmers spent
an average of $30.06/ha on insecticides. As products were applied in various forms
(kilograms, liters, etc.), application rates varied by product. The average use rate was
0.63 liter/ha (l/ha) for Nuvacron, 0.24 1/ha for Decis, and 1.08 l/ha for Sistemin. The price
for Nuvacron averaged $15.45/1iter, Decis averaged $33.17/1iter, and Sistemin averaged
$11.14/liter. RK farmers purchased 64 percent of these insecticides fi'om private

companies such as Bayer, Ferquido, Monsanto, Fersan, Shell, and Brugal.

110

Table 5.10 Average Quantity and Cost of Purchased Input Use Per Hectare by

 

T."E.9fF.“."‘.‘.°_" WW???” .8. SJV

 

  
    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Seed (kg/ha) 136.0 $215.09 144.1 $224.76 128.4 $220.66
Basal 447.6 $121.41 418.5 $118.84 533.4 $144.35
Fertilizer

(ks/ha)

Foliar 79.0 $5.66 52.0 $3.38 100.0 $11.56
Fertilizer

(is/ha) _

Insecticide na $30.07 na $32.60 na $23.27
Fungicide na $17.99 na $14.46 na $21.97
Herbicide na $4.82 na $3. 14 na $9.78
Other na $0.72 na $0.45 na $0.65
Total Cost $395.75 $396.84 $432.25
(USS/ha)

Note: Total Sample, Small, and Large Farm averages are sample/subsample averages, not the average
of those farmers within the sample/subsample who used the item. Exchange rate used is 14.5 RDS/
USS, the average rate during the season.

Source: 1998 RK Survey

 

 

Small RK farmers spent more per hectare on insecticides than did large farmers,
primarily because the former average 3.3 applications compared to 2.9 for the latter. In
addition, the RK data indicates. that small-farmers also tended to apply higher
concentrations of identical products compared to large farmers. By contrast, small
farmers (N=3 7) in the Producer Survey averaged 2.9 applications, compared to 3.4 for
large farmers (N=30). While the prices of identical products in the RK data varied across

farmers, they did not seem to vary by farm size. Small farmers may make more

111

applications due to less technical knowledge, or they may simply have more exposure to
insects since they generally plant later (due to poorer access to tractors and seeds). Thus,
when larger farmers’ plants are mature, smaller farmers’ beans are still in the early growth
stage and attract more insects (notably the whitefly).

5.7.3 Fungicide

- Many farmers (79 percent) also applied firngicide, averaging 1.6 applications per
farmer (Producer Survey, 1998). Although RK farmers used 13 difi‘erent products, the
most popular were Alto 100 (14 farmers), Dithane (12), and Anvil (4). These farmers
spent an average of $19.33/ha on fungicides. As these products were applied in various
forms (kilograms, liters, etc), application rates varied by product. The average
application rate‘was 0.007 liter/ha (l/ha) for Alto 100, 0.112 llha for Dithane, and 0.036
We for Anvil. The price of Alto averaged $96.00/liter, Dithane averaged 87.30/1iter, and
Anvil averaged $20.41/liter. RK farmers purchased 67 percent of these filngicides fi'om
the local afiliates of private companies such as Bayer, F erquido, Monsanto, Fersan, Shell,
and Brugal.

Small RK farmers spent less on filngicides than did large farmers, although the
former average 2.0 applications, compared to 1.8 for the latter. By contrast, small farmers
who used fungicide (N=27) in the Producer Survey averaged 1.5 applications compared to
1.7 applications for large farmers (N=25). As with insecticides, while prices varied across

farmers, they did not vary by farm size.

112

5.7.4 Herbicides

Herbicideseare not used by many farmers (11 percent; Producer Survey), although
CRSP researchers feel that increased use could help to decrease production costs,
especially given increasing wage rates. Among the RK' farmers, there is considerable
variation in use rates, herbicide prices, and costs per hectare, indicating that there is not a
standard procedure used by farmers for integrating herbicide use into crop management.
Nine RK farmers applied herbicides, with six using one application and three using two
applications. Among the Producer Survey farmers who used herbicide, 9 made one
application and two made two applications. Although the RK farmers used 7 different
products, the most popular was Gamoxone (6 farmers). Other products included Dual,
Afalon, Gamozil, and Paradox. RK farmers using herbicides spent an average of
$15.53/ha on herbicide and $5.50 on the labor for application. The average application
ratewas 0.075 liter/ha (l/ha) for Gramoxone, while its unit cost averaged $10.84/1iter.
Herbicide costs per hectare (without labor included) ranged from $2.03 to $63 .33/ha
(RD$1.85 to $57.75/ta). The economics of substituting herbicides for manual weeding is
discussed later in this chapter.

1 5.7.5 Total Purchased Input Costs
5.7.5.1 W

The primary cost components of the average RK farmer’s purchased input costs
are seed (54.3 percent of input costs) and basal and foliar fertilizers (31 percent). In
contrast, insecticides (7.6 percent) and fungicides (4.5 percent) are small purchased input

cost items, compared to seed and fertilizer (Figure 5.1). While RK farmers used the

113

 

Figure 5.1 Purchased Input Cost Shares
1998 RK Average, San Juan Valley

   

Fertilizer (30.7%

 

 

 

CRSP’s recommended fertilizer levels, they planted 28 kg more of seed per hectare than is
recommended. At an average unit cost of $1 .73/kg, this excess seed density amounts to
$48.46/ha in additional input cost. Thus, planting at the recommended 108 kg/ha swd
density would reduce the average farmer’s input costs by 10 percent, and his/her total
production costs by 5 percent.
5.8 Machine and Labor use

5.8.1 Introduction

Labor accounts for 63 percent of total machine and labor expenses, while machine
accounts for 37 percent (RK survey, 1998). The primary machine and labor operations
are harvesting (31 percent of operation costs), land preparation (29 percent), irrigation (13
percent), and manual weeding (13 percent). The only machine-intensive activity is land

preparation, 94 percent of the cost of which is mechanized (tractors). By contrast, only

114

20 percent of the cost of harvesting is machine cost (threshers).

Table 5.11 presents the average cost of each operation for the RK sample (N=29),
large farmers (N=9), and small farmers (N=9), separated into equipment and labor
components: The table presents operation averages for those farmers who used each
operation (ssTotal columns), as well as averages calculated for the total sample (Total
columns). The “ssTotal” column indicates the average expense for that operation for the
subsample of farmers who carried out the respective operations. To calculate the total
sample average expense for that operation, the number of farmers who used that operation
was multiplied by their average operation expense, and then the subsample total was
divided by 29, the number of RK farmers in the sample.

' A farmer’s cost for each operation category depends upon the number of
operations performed during the season, the hourly wage or per hectare fee paid to the
labor or machine performing the operation, and the hours/hectare of labor or machine
employed by the farmer. Table 5.12 reports sample and subsample (large and small
farmers) averages for hours/ha for each operation and the average sample wage paid for
that operation. Average hours/ha multiplied by the average wage does not equal average
operation cost per farmer as reported in Table 5.11 (just as multiplying average farmers’
yield by average farmers’ price does not necessarily equal average gross returns). The
average wage rate across operations is weighted by the hours/ha. As family labor makes
up a very small percentage (of even small farmer’s labor costs), family labor was included

with hired labor in this table. Family labor was valued at RD$10/hour, close to

115

TABLE 5.11: and Labor Use SJV 1998
armer TOTAL

55 otal and Labor

1.
00

and Labor

TOT

and Labor

and Labor

and Labor

farmers who
those

 

116

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

able 8,12 Hours per hectare and wage rates for Bean Production
Operations by Farmsize, San Juan Valley, OR, 1998
~ ' Average Small Large
' - LABOR Hours/ha Wage Hours/ha Wage Hours/ha Wage
[Operation $USIhr $US/hr $USIhr
and Preparation ‘ 9.2 $1.02 14.0 $0.82 13.4 $1.07
lanting - _ 41.5 $0.83 45.8 $0.84 41.2 $0.85
ertilizer Application 8.9 $0.98 7.2 $1.48 11.8 $0.87
nsecticlde Application 12.2 $1.25 13.5 $1.80 10.8 $1.08
ungicide Application 5.2 $1.32 4.3 $1.92 5.2 $1.18
anual Weeding 81.8 $0.93 74.9 $0.89 75.2 $0.88
erbicide Application 5.8 $0.95 8.4 $1.47 4.9 $0.82
rrigation ' 73.0 $0.92 78.5 $0.82 73.8 $1 .07
awest ’ 148.5 $0.77 150.4 $0.78 147.4 ‘ $0.75
385.7 $0.93 397.0 $0.92 383.7 $0.87
Average Small Large
Hours/ha Rent fee Hours/ha Rent fee Hours/ha Rent fee
$US/hr $USIhr $US/hr ‘
15.9 $13.82 20.4 $8.63 20.2 $14.18
18.3 $0.80 18.9 $0.83 19.8 $0.83
1.9 $18.93 2.1 $18.21 2.2 $11.98
38.1 $8.38 39.3 $5.14 42.0 $8.57
urce: 1998 RK Survey (N=29)

 

 

 

but slightly below the average wage paid (RD$13 .49/hour) by the average RK fanner for
all operations.

On average, small farmers (<2.5 has) pay a higher wage for labor used in bean
operations (5 percent more than the average wage paid by large farmers), yet employ
fewer hours of labor per hectare. In contrast, large farmers (>6.3 has) pay a rental rate for
machine and utilize more hours per hectare of that machine. There are two potential
explanations for why small farmers pay higher wages. First, larger farmers can negotiate a
lower wage because they have a larger number of hours to offer each laborer (although
not necessarily more hours per hectare). Secondly, higher labor wages paid by small

117

farmers may be the result of a different service being provided. For example, for
insecticide, fungicide, and herbicide application, smaller farmers may pay for the laborer’s
time plus the laborer’s application machine use, while large farmers may own several
backpack sprayers and just pay for the laborer’s time.

Higher machine rental fees paid by large farmers is very likely explained by the fact
that large farmers infi'equently hire PROSEMA for land preparation (subsidized state
tractors) due to the poor quality and availability of these tractors/operators. Another .
possible explanation is that larger farmers simply are hiring better quality services and/or
are willing to pay more to secure the services as early as possible in the planting date
period. For example, the 1998 Producer Survey indicates that 30 percent of small
farmer’s hired land preparation from PROSEMA, compared to only 16 percent of large
farmers. In addition, the average small farmer planted on 13 November, while the average
large farmer planted on 8 November. This suggests that large farmers pay higher land
preparation fees because they do not utilize the services of PROSEMA. They are willing
to pay higher rental rates to insure higher quality land preparation and fewer delays. The
following sections provide details on the calculations of operation costs.

5.8.2 Land Preparation

The seven RK farmers who owned tractors reported only gas expenses for their
land preparation, and data were unavailable to calculate depreciation. Given this situation,

a rental rate was calculated” and used to estimate farr'ners’ land preparation expenses in

 

, ‘ 2° The rental rate was calculated as the median of the rented land preparation
rates for large farmers, so as to avoid the influence of low and high rates.

118

the place of the lower reported figures.

Land preparation is the source of the largest difi‘erential in operation costs between
large and small farmers. Besides the fact that larger farmers pay higher tractor rental fees
(for reasons discussed above), larger farmers average 6.2 land preparation operations
before planting whereas small farmers average 4.2 (the average for all RK farmers is 5.2).
Thus, the total land preparation cost difi‘erential is not surprising. '

5.8.3 Planting

Farmers did not report any costs for machine use (although hours of machine were
recorded) because each RK farmer owned a horse. However, a “ren ” rate was charged
to. each farmer (for a horse and boy at RD$12lhour together), based on local rates
reported for the previous season (Amaud, 1997).

5.8.4 Application of Insecticide

Small farmers spend more on insecticides, both because they apply them at higher
rates and because they average 3 .3 applications, compared with 2.9 for larger farmers.
Larger farmers probably make fewer insecticide applications because they plant earlier
(thus their plants flower before the whitefly vector multiplies), prepare their fields better
(thus have fewer weeds), and follow the spraying recommendations (both application rates
and number of applications).

5.8.5 ' Manual Weeding, Irrigation, and Harvest

These three operations account for 78 percent of the average farmer’s labor
expenses. Although the average large farmer manually weed morefi'equently (2.2) than

do small farmers (1.2), large farmers spend $10/ha less on manual weeding due to lower

119

wages paid for this operation (25 percent less). Irrigation includes both cleaning the
irrigation canals and managing the irrigation water when it arrives at the farm. Larger
farmers performed more irrigation operations (7) than did small farmers (S), and paid
higher wages. However, at harvest, large and small farmers hire approximately the same
amount of labor (hours/ha) and pay very similar wages. This finding implies that large and
small farmers attain similar yields.

5.9 Other Costs

5.9.1 .Credit

Financing costs play a significantrole in determining profitability of bean
production in the San Juan Valley for two (reasons. First, because the vast majority of
production costs are monetary, farmers must obtain loans in order to purchase inputs, rent
labor and machine, etc. (either directly fi'om banks or indirectly from input dealers).
Second, because many farmers sell their harvested beans to the government, which often
delays final payment for months, their outstanding debts continue to‘ increase until they
receive payment. Thus, a farmer’s interest rate and loan size, relative to production cost
outlay, are important determinants of his/her net returns to bean production.

Eighty-six percent of the Producer Survey respondents received some form of
credit for bean production activities during the 1997-98 winter bean season. The
remaining 14 percent did not receive credit, either because they did not need financial
assistance (9 percent) or could not obtain a loan (5 percent). Of those who received I
credit, 69 percent received one loan, 28 received two loans, and 3 percent received three

loans. Seventy-three percent of those receiving at least one loan indicated that they

120

received enough credit for their bean production activities. Loan sources and related
information are summarized in Table 5.13 below. For the 27 percent of respondents who
reported not receiving enough credit, they indicated that they could have used additional
loans for harvest (50 percent), inputs (21 percent), land preparation (14 percent), machine

purchase/repair (7 percent), weeding, and irrigation (each 3.5 percent).

Banco Agricola 80

Commercial _ 88
Banks

Moneylenders 95
FDD 42

Farmer 97
Associations

Input Dealers 3 36 36 100

' % of total loans (value) by each source b % of farmers receiving loans from each source
Source: 1998 CRSP l

 

Loan form varied among sources, although the primary lenders gave credit to
farmers in the form of cash, input vouchers, or inputs themselves. Banco Agricola gave
67- perCent cash and 33 percent input vouchers, commercial banks gave 100 percent cash,
and moneylenders, gave 78 percent cash and 22 percent inputs.

Commercial banks lend exclusively to landowners and mostly (75 percent of

commercial bank loans) to larger farmers, as the risks of bean production are simply too

121

high for commercial banks to lend to small or medium-size farmers (Libre, 1998). Banco
Agricola, a state-run development bank that offers a subsidized annual interest rate of 18
percent, had a more balanced spread across the sample with large farmers receiving 30
percent of Banco Agricola loans, medium farmers receiving 41 percent, and small farmers
29 percent.) In contrast, Moneylenders lent primarily to small farmers (58 percent of
moneylender loans), although they also lent to medium (31 percent) and large (11 percent)
cm.

Dueto the interest rate difi‘erentials across loan sources (Table 5.13), it is not
surprising that average interest rates by land tenure situation are significantly difl'erent.
Land owners paid an average annual interest rate of 38 percent, sharecroppers an average
interest rate of 47 percent, and land reform farmers an average interest rate of 13 5
percent. Because land reform farmers have use rights but no title to their land, they could
not secure loans fi'om commercial banks, as well as few from Banco Agricola. Land
reform farmers obtained 61 percent of their loans fi'om moneylenders, 19 percent fiom
Banco Agricola, 10 percent from their association, and 5 percent from input dealers and
FDD. Thus, these farmers pay more than twice the interest rate of large farmers due to
their heavy reliance on moneylenders for loans.

The proportion of loans to total cost of production (the loan burden) was
estimated using average costs of production by farmer fi'om the RK data, combined with
that RK farmer’s loan information reported in the 1998 producer survey. The results
show that the loan to cost of production ratio averages 69 percent, but was higher for

small farmers (77 percent) compared to large farmers (68 percent). When the loan burden

122

statistics are combined with the interest rate differential across farm sizes, it is clear that
the resource limitations of smaller farmers (less collateral and higher production risks)
imply greater debt and lower profitability. Although more than 98 percent of Banco
Agricola loans went to non-landowners, only 14 percent of Banco Agricola loans went to
land reform farmers. Thus, while the Banco Agricola is reaching some small and medium-
size farmers, it is not reaching land reform farmers with access to irrigated land; the
majority of these farmers rely on moneylenders.

Although these conclusions are not surprising, one policy implication is that
subsidized credit is not reaching the poorest farmers in the SJV. Key informants reported
that theLand Reform Institutes (IAD). continued delay in granting title to these farmers is
a factor that contributes to reducing these farmers’ access to Banco Agricola credit.
WhileBanco Agricola did in fact lend to a greater percentage of bean farmers in 1997/98
than in the previous year (PRODAS, 1997), this is likely explained bythe simultaneous
efi’ect of an increase in the average farm size (as discussed above) and the decrease in the
number of farmers planting beans.

5.9.2 Irrigation Charge

In thewinter season, bean farmers pay $13.16/ha (RDSlZ/ta) for irrigation water
which is provided by INDRHI, the government irrigation agency.

5.9.3 Internal Transport

Each RK farmer also incurred expenditures related to transportation to and fi'om
their bean fields, as well as the use of transport during harvest to move beans from the

field tolareas for drying/sorting/bagging. 'Based upon RK survey data and government

123

cost of bean production budgets, $10.00/ha is assumed to cover the average farmer’s
transport costs.
5.10 Total Costs of Production

5.10.1 Cost Shares

Machine and Labor operations (Land Preparation, Planting, Labor for Input
Application, and Harvest) from Table 5.6 account for 48 percent of total costs, purchased
inputs account for 42 percent, and transport and financing make up the remaining 10
percent. Table 5.15 below uses the average total input costs fi'om Table 5.3 and -
operations costs from Table 5.4 to develop a representative farm budget for the RK
average farmer (small and large farm budgets are included in the appendices). Figure 5.2
shows the shares of the cost components. The following sections discuss these tables and
their implications for the present and future profitability of bean production in the San
Juan Valley.

5.10.2 Decreasing the Costs of Production

While technical change is perhaps the most general method to reduce production
costs in agriculture, the RK data show that there are various measures which farmers
could implement to decrease their costs of production, as well as options for government

policymakers.

124

Table 5.14: Financial Coats of Red Beans

Winter Season 1997 San Juan DR
anner

 

125

5.10.2.1 Inppt Application gates

As noted above in the seed and insecticide sections, input application rates are
higher than those recommended by the CRSP. Farmers’ seed density is especially
significant because seed itself constitutes 22 percent of total production costs for the
average farmer. The average farmer could reduce his production costs by 6 percent, if he
lowered hisseed density fiom the sample average of 18.9 lb/ta to the recommended 14
lb/ta. As noted above, the high seed rate appears to be a function of not using wider discs.

. 5.10.2.2 Mapugl Weging vs. Hflicides

CRSP researchers believe that SJV farmers can reduce their production costs by
replacing manual weeding with herbicide use. Twenty-one RK farmers used only manual
weeding, while 8 RK farmers used manual weeding, combined with one herbicide
application. According to this data, the cost implication of this tradeofi‘ appears to be
negligible because the 21 “manual-only” farmers spent an average of $67.8 l/ha (61 .84/ta)
on manual weeding, while the 8 “manual-herbicide” farmers spent an average of $49.43/ha
(45.06/ta) on manual weeding, $15.53/ha (14.16/ta) on herbicide and $5.51/ha (5.02/ta)
on herbicide application — a total of $70.47/ha (64.23/ta). Analysis of RK data from the
1998-99 season may provide additional evidence by which to test the hypothesis that

switching to herbicides may enable farmers to cut their production costs”.

 

2‘ However, this data do not control for other production management factors.
To test this hypothesis formally, CRSP researchers would need to set up experimental
trials or ask farmers to perform superimposed trials.

126

5.10.3 Government Policy and the Costs of Production
5.10.3.1 Input Prices

As the primary purchased input cost components, the total cost of seed (52 percent
of input ecst) and fertilizers (30 percent) are important components of total production
costs, whether valued in financial or economic terms. It is clear that seed sold by SEA are
subsidized at about $0.35/lb (SOO/qq), which is about 36 percent of the total economic
cost of seeds purchased by a farmer directly from SEA This calculation, however,
assumes that seed is purchased fi'om SEA, although only about 34 percent of the valley’s
farmers buy seed directly from SEA in a given year (Amaud, 1997). Since, many farmers
store their own saved seed in SEA’s cold storage facility (free of charge), the only subsidy
for these producers is free storage. Using a private sector cold storage rate of 80.01/lb
per month of storage (reported by one respondent), the SEA free cold storage is
essentially worth $0.08/lb. In the absence of SEA seed subsidies, farmer financial prices
for seed would thus increase from between 9 percent (for those storing their own seeds) to
36 percent (for those purchasing direct from SEA”). The implication of the loss of either
the storage or seed sale subsidy is that total costs of bean production could increase
between 3.6 to 5 percent for the average SJV farmer.

Several farmers and farmer associations complained that fertilizer prices in the DR
are excessive, thus contributing to higher production costs than necessary. Monthly data

collected at the wholesale level in-country across Central America during 1996 and 1997

 

22 It is difficult to establish the private sector rate for seeds because the private
sector share of the bean seed market is less than 5 percent, according to the 1998
CRSP Producer Survey (N=100).

127

(and including the DR) by CORECA (Consejo Regional de Cooperacion Agricola), an
amliate of IICA, indicates that while Urea prices in the DR are somewhat higher than the
Central America average, Ammonium Sulfate is cheaper and 15-15-15 is essentially the
same. The DR distributor-level price for Urea is approximately 25 percent higher than the
average in Central America during November-December, and the DR price of Ammonium
Sulfate is about 3i) percent less during the same time period. A study by IICA of the
fertilizer subsector found that fertilizer prices are roughly 10 percent higher than they
would likely be with a more competitive subsector (Nunez, 1999).
51032 W

Labor accounts for 30 percent of the average RK farmer’s total costs, compared to
31 percent for small farmers and 28 percent for large farmers. According to key
informants, hourly wages in the San Juan Valley (and in agriculture in general) are
considered to be high and have been increasing in recent years. Several explanations are
given for increasing wages, including high labor demand during critical periods (planting,
harvest),.a general shortage of labor due to emigration to higher-paying jobs in urban
areas (Santo Domingo and Santiago’s Free Trade Zone jobs), and emigration to the US.
or other countries.

5.10.3.3 WW

As the price of seed is not likely to be lower in the firture, 'wage rates will likely
continue. to rise, and input use levels (aside fi'om seed density) are not excessive,
decreasing the costs of bean production (without technical change) will require either

finding a way to reduce financing costs, reducing the cost of fertilizer, or increased

128

mechanization of harvesting and threshing. As the sections on profitability below
demonstrate, if bean prices begin to fall due to increased pinto imports, changed consumer
preferences towards pintos, or lower SEA and/or INESPRE support, higher-cost
producers will not remain profitable unless production costs are lowered or associations
find a way to command higher prices for their beans in the private market (through
forward contracting, storage, etc). However, the associations and farmers most capable
of adapting new production or marketing techniques are those farmers whose costs are
low enough to avoid having to make these changes in order to remain competitive.
5.11 Production Problems

Respondents were asked to identify their principal production problem during each
of the past five winter bean seasons. The responses in Table 5.15 indicate that “Bean
Golden Mosaic Virus” (BGMV) remains the principal production problem for most
farmers. However, in 1997/98, “early defoliation” was the primary production problem.
In addition, “water availability” (and excess rain on occasion) is a recurring problem for
some farmers. While BGMV was the principal production problem perceived by farmers
in the past five years, its incidence and/or severity has diminished since the major epidemic

ofl994.

129

BGMV

Early Defoliation
Rust
Drought/water

Vaneo de la
vaina (insect)

Excess rain

gusano (worm)

Weeds

Germination

Problem

Respondents by N=67 N=83
year

-Sour¢e: 1998 CRSP Producer

 

5.12 Marketing

5.12.1 Marketing Channels

There are three primary marketing options for bean farmers in the SJV: the SEA
seed multiplication program, the INESPRE bean marketing program, and private traders
(the market). Twenty-six percent of the Producer Survey respondents sold beans to SEA,
64 percent sold to INESPRE, and 29 percent sold to the market (many farmers sold to
more than one buyer). However, in terms of total volume of sales aggregated across
respondents, SEA accounted for 35 percent of bean sales, INESPRE 37 percent, and the

130

private market 28 percent.

5.12.2 SEA Seed Program -

The highest bean prices in the valley are paid by the SEA bean seed program,
which oficially only purchases bean seed from pre-approved farmers participating in its
seed multiplication program. SEA pays these growers a premium price (above that of
INESPRE and the market price) to ensure that farmers under contract do in fact sell their
seed to SEA While SEA paid US$1.79/kg (RDSl,l75/qq) for seed in 1998 in the SJV,
they did not pay growers until an average of 7.2 weeks alter the time of sale. Thus,
adjusting the SEA seed price (the same for all seed program farmers) by the opportunity
cost of the farmer’s delayed payment (the value of each farmer’s sale, times the
commercial interest rate of 24 percent, times the weeks waiting for payment) and the
farmer’s estimated loan payments during the waiting period (the farmer’s total loan burden
times the farmer’s average interest rate, weighted by loan size, times the weeks waiting for
payment), SEA’s adjusted average price was actually $1.70/kg (RD$1,117/qq). For those
farmers who sold to SEA, this buyer accounted for an average of 83 percent of the
farmers’ total sales.

5.12.3 INESPRE

While INESPRE is authorized to purchase beans for their national consumer food
program fi'om any farmer in the valley, in recent years, they have purchased almost
exclusively through farmer associations. All farmer associations are able to sell to
INESPRE, although INESPRE does have standards for the grain. However, while a

classification system that linked price with quality (3 classes) used to ofi‘er a premium

131

(discount) for higher (lower) quality beans, in 1998, only one classification existed.

. "Eight-seven percent of the survey respondents who sold to INESPRE said they did
so because it ofi‘ered the best price (for those farmers who did not have the option of
selling seed to SEA). For these famiers, INESPRE accounted for an average of 87
percent of their total sales. Although INESPRE purchased beans at $1.57/kg
(RDSI,03 5/qq) in 1998, they did not pay growers in firll until ”an average 13.8 weeks after
the sale (as discussed in Chapter 4.6.2.3). Thus, INESPRE’s adjusted price for the
Producer Survey farmers was actually $1.47/kg (RDS964/qq), a decline of about 7 percent
from the nominal “purchase” price. This point is not lost on SJV farmers, who organized
a protest against the government in July of 1998; by that time, many of these farmers had
not received payment for beans delivered in February. INESPRE has delayed payment in
this fashion several times in recent years, although each year they promise quicker
payments to farmers, and farmers continue to sell to INESPRE.

5.12.4 Private Traders
F arrngate prices paid by private traders averaged $1.29/kg (RDS849/qq), although

the prices ranged fi'om $0.9l/kg to $1.67/kg (RDS600/qq to RD$1,lOO/qq)”. Eighty

 

' 23 It is not clear to the author why this price variance is so large. Transport
cost difi‘erentials are not likely the explanation given the small size of the valley.
Quality differentials are likely part of the explanation, but not all of it, given that retail
quality differentials are not as large proportionately as these farmgate difi’erentials.
This price variance may well be seasonal (wholesale and retail prices fall dramatically
in January, February and March), although this analysis has not been done. Another
potential explanation is that the farmers who received the lowest prices were typically
those who borrowed fi'om moneylenders, who in some cases are also bean traders who
may have simply purchased the farmer’s harvest and taken the farmer’s loans out of
the price paid to the farmer.

132

percent of farmers selling to the market said they chose this buyer because they needed
cash immediately to repay debts. For farmers who sold to traders, this buyer accounted
for an average of 75 percent of their sales.
_ 5.12.5 Average Farmer Prices by Farm Size

Average prices for all beans sold per farmer were calculated across the sample.
These prices. were then adjusted for the payment delay of SEA and INESPRE, as
described in Chapter 4.6.2.3. Adjusted average prices varied considerably among farm
sizes,- with small farmers receiving an average price of $1.43/kg (93 8/qq), medium farmers
$1.47/kg (969qu), and large farmers $1.54/kg (1,012/qq). Farmers selling primarily to
SBA and INESPRE - even considering adjustments - obtained higher prices for their
beans. The average adjusted price for the total sample was $1.48/kg (977/qq), compared
to the unadjusted average price of $1 .55 (1,019/qq).

5.12.6 implications of SBA and INESPRE Bean Purchases

Just as the distribution of land and capital resources has a large effect on the
interest rate paid by a given farmer, a farmer’s financial and production situation determine
the price he receives for his harvested beans. It should be noted that SEA pays a premium
for disease-flee, uniform seed, and INESPRE has legitimate needs for quality, uniform
grain for human consumption. Thus, these prices reflect higher quality domestic grain
delivered by certain farmers. However, a farmer’s financial and production capabilities
determine to whom he can sell beans. Only a limited number of approved growers can sell
seed to SEA each year. Although INESPRE pays a premium for quality beans produced

by anyone, due to delayed payments, farmers with high debt and poor cash flow cannot

133

afford to sell to INESPRE due to the risk of delayed payment.

However, these two government procurement programs do not escape the typical
problems of corruption and quality management that are associated with such high
information cost activities that are managed by political agents. There were reports of
SEA buying seed fi'om non-contracted (i.e. non-approved) farmers or of contracted
farmers delivering to SEA grain fi'om another source, rather than seed grown on pre-
approved parcels. On the other hand, there were also several farmers with contracts to
deliver seed that had to sell to INESPRE instead because their seed did not meet SEA
standards Whileit-is diflicult to determine through rapid appraisal how impartial SEA is
in choosing program growers and in monitoring the quality of seed delivered, it is clear
that this process is not flee from the political partiality that seems to characterize many
government programs in the DR.

Quality management issues can perhaps be best assessed in terms of farmer
decisions regarding seed source. When asked why they stored their own seed (rather than
buying SEA seed at subsidized prices), many respondents said that SEA seed was
unreliable in regards to both quality and timing. Thus, it is not surprising that many SJV
farmers are storing their own seed or buying seed from their association’s seed bank, even
when SEA seed typically sells for RDS 900qu. According to the 1996 producer survey
irnplemented by the CRSP DR stafi’, seed sources for the 1996-97 season included own
(58 percent), SEA (34 percent), other producer (4 percent), and warehouse (4 percent).

In the case of INESPRE, their price supports primarily benefit larger farmers, and

sales volume to INESPRE and SEA may result in thinner markets for small producers.

134

There were also reports of non-bean-growing members of farmer associations selling
beans to INESPRE. These association members simply purchased beans from small valley
or mountain farmers at the lower market price, and then sold these same beans to
INESPRE at the higher negotiated price.
5.13 Profitability of Bean Production

5.13.1 Financial Profitability

The following financial and economic profitability analysis (Table 5.16) uses
average yields and adjusted prices fi'om the Producer Survey, and average cost of
production data from the RK Survey“. It is clear that given 1997/98 average yields, costs
of production, and average adjusted farmgate prices, bean production in the SJV is
financially profitable (Table 5.16). Enterprise Profit in this analysis includes returns to

management”.

 

.” Yields and prices received on average and by farm size are taken fi'om the
Producer Survey because this sample is considerably larger (N=100) than that of the
RK survey (N=29), and because the RK farmers were among the better farmers in the
valley (they have higher yields and receive higher prices, regardless of farmsize).
However, the cost of production data (Table 5.15) is from the RK survey, because the
Producer Survey did not collect this type of detailed extensively. In fact, the genesis
of the RK survey lay in the dimculty of collecting accurate cost of production data
from recall surveys.

2’ The opportunity costs of land, labor, and capital are deducted from gross
revenue in Table 5.15. The opportunity cost of land is assumed to be the
sharecropping payment, equivalent to one-fifth of expected gross returns/ha. The
opportunity cost of labor is not an issue in the SJV because all labor is hired. The
opportunity cost of capital is accounted for in the financing line item in variable costs
(for the 69 percent of variable costs financed with external loans) and in equity capital
in Table 5.15, which is the additional cost of production financed by the producer
himself. Therefore, gross margin less land and equity capital opportunity costs gives
enterprise profit - the returns to management.

135

Although average financial costs of production per hectare are similar for small
and large farmers, large farmers have higher average yields ( 14 percent higher), lower
costs/kg (13 percent lower), and higher adjusted price received/kg (8 percent higher).
Therefore, both higher yields and higher prices explain why large farmers’ average net
revenue/ha is 64 percent higher than that of small farmers.

. 5. 13 .2 Economic Profitability

' By contrast, under economic valuation of inputs and outputs, SJV bean production
is not profitable (Table 5.16). Removing seed, fertilizer, and irrigation subsidies/taxes
(Chapter 6.7) fiom the financial RK budget increases the average RK farmer’s 1998
variable costs per hectare by 7 percent. However, this increase in cost (decrease in
profitability) is quite small compared to the decrease in the farmgate output price
ernbodied in valuing farmer output at the economic import parity price (IPP) (Table 6.2).
Even assuming a 20' percent consumer preference premium (Chapter 6.6.2) on top of the
economic IPP (calculated in Table 6.2) of $0.78/kg in 1998, the inflated economic IPP
price of $0.94/kg is 37 percent lower than the financial farmgate price of 1998 — and

returns per hectare are negative.

136

 

_

_

m

.32:

 

.380 o «:9 :.

 

 

soc; no; o: 3::
8.5 o;

.3: a o >
E953 noxooxm m

can. no.8:
.«oocxo o

 

 

 

onE «69:. 35

 

 

 

”E

 

$02: .00: a ”0050
a 0 «:05. so am. «can 356 a
2 2583 m. ca 0 3:55am

  

 

 

 

 

 

 

 

 

 

137

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

 

 

.__c8..§x 252.com

 

 

 

8.8.3 8.8% 8383 8.88 8.88 3.88 2.88 Eea out new
8.8 8.8 8.8 8.8 8.8 8.8 . $589 .238 25.8 -
8.88 8.88 8.88 3.38 8.88 8.88 _ 86:8 23 -
manoO aEatmmO
$88 8129 8.2» 8.88 8.88 8.88

8.898 8.18.8 8.~8.8 8.398 8.88 2:8 858V 980 28.5; .58 -
:88 8.88 3.28 2.8 8 8.88 $.28 85% £80 o.cn.en> .86 -
8.8 8.8 8.8 288 8.88 9.9» a «26:3 cameo 888.. -
8.88 8.8; 8.3 8 we: 3 we. a 8 3.88 A2838 ecu___tox -
2.88 8.88 8.88 2.88 8.88 8.28 048833 88 -
88:8 883.; 888.8 888.8 888.8 «38.8 268 3:23. 320
3.8 8.8 3.8 8.8 8.8 8.8 @533 a 8E Eon. coup:
[83 . «.8; P2; slum; 8.13 R: mead: 2o; cocci

85 xx =85 xx oasnm xx pea. xx =sz xx oases. xx .22.

 

hawthorn 30:05“.

 

_

 

 

_
E0 .>

 

 

o5 :_ seventeen swam h.e 33:05 o_Eo:oow use 305$". mfm afic—

 

 

5.1.3.3 Alternative Crops in the SJV

There are various cropping alternatives“ to winter bean production in the SJV,
although these alternatives are not as financially profitable (Table 5.17). Apart fi'om this
advantage, the bean market at present is much larger and more stable, as discussed at
length in Chapter 4. However, when economic import parity prices are used to compute
economic net returns, pigeon pea and sweet potato become the most profitable crops for

the winter season.

 

Table 17: Financial and Economic of Selected Annual 1998

 

arm

5.14 Sensitivity Analysis
5.14.1 Introduction
In the following tables, sensitivity analyses are carried out by changing bean yields

and bean prices - using the average net revenue per hectare in the base run - in’ order to

 

2‘ These alternative crops were those most fiequently mentioned by 1998
CRSP Survey respondents as the crops that they would switch to if bean prices fell
substantially. .

138

identify the variables that most affect the level of net revenue per hectare. As this
sensitivity analysis does not take into account the probability of any of the changes
actually occurring, risk is not considered.

5.14.2 Changes in Bean Yields and Prices

' The results of sensitivity analyses are reported for average, small, and large SJV
fanners using average yield and price fi'om the Producer Survey, and average cost of
production from the RK Survey. For each farrnsize group, average bean yield is varied
from -50 to +40 percent of the average yield, while the adjusted bean price is varied
betvlreen -50 and +30 percent, thus showing the sensitivity of net revenue per hectare (by
farmsize) to‘ changes in yield and/or adjusted prices, ceteris paribus.

The average farmer with no yield reduction will remain profitable with up to a 20
percent decline in adjusted price received, while the same farmer with no price reduction
will remain profitable with even a 20 percent decline in yield (Table 5.16). However, once
yield and price are simultaneously reduced by more than 10 percent each, (or a 10 percent
reduction in price (yield) combined with a 20 percent reduction in yield (price)), the
averege RK farmer becomes unprofitable (enterprise profit less than zero).

The average small farmer’s bean operation becomes unprofitable: if price declines
more than 10 percent or if yields decline more than 10 percent (Table 5.17). Ifboth yield
and price decline by more than 10 percent each, the average small farmer becomes
unprofitable. By contrast, the average large farmer is more resistant to declines in yield
and/or price as they ean remain profitable even with a 25 percent reduction in either yield

and price, or combinations of yield and price reductions of 10 and 20 percent respectively

139

(Table 5. 18).

These results suggest that if the DR administered tariff-rate quotas instead of the
current import licensing regime, domestic farmgate prices could fall as much as 20 to 30
percent, to the 1998 Financial Import Parity Price of approximately $1.03/kg. Assuming a
20 percent consumer preference premium for red beans, the farmgate price would be
$1.23/kg - a price slightly above the “break-even” price of $1 . l9/kg with which the
average RK farmer can maintain positive enterprise profits. However, if this consumer
price premium were to fall or disappear over time, then the average RK farmer would not
remain profitable.

Ifthe DR both abolished import licensing and removed the 25% tariff and 8%
value-added tax fiom bean imports, domestic farmgate prices could fall to the 1998
Economic Import Parity Price of approximately $0.7 8/kg ($0.94/kg with the 20 percent
consumer preference premium). In this case, the average SJV bean farmer would no
longer remain be profitable, even with the twenty percent price premium. It should be
noted that these scenarios are for average producers (average prices, yields, costs), thus
lower-cost producers and/or producers who find ways to decrease production costs would

possibly still be able to compete with pintos.

140

 

141

 

5.14.3 Sensitivity to Changes in Input Costs

Table 5.21 presents a similar sensitivity analysis for the average farmer in which
input costs (seed, fertilizers, other chemicals, and irrigation charges) and/or Labor Costs
are allowed to vary between +40 percent to -40 percent. With even a 30 percent increases
in both input and labor costs, the average farmer earns $142/ha in enterprise profit, which
constitutes a 59 percent decline in returns from the 1998 average. Thus, it is apparent that
the average farmer will remain profitable — even under reasonably potential increases in
labor or input costs. One such potential scenario is that if farmers faced the economic
values of both seed and irrigation charges, this would result in an increase in input costs by
18.5 percent. In this case, the average farmer’s enterprise profit would fall to $271/ha,
which amounts to a decline in returns of about 23 percent. Thus, the abolishment of seed
and irrigation subsidies would have a noticeable effect on the profitability of the average
SJV farmer.

142

Perhaps the principal implication of the results of these sensitivity analyses is that
the significant support that bean farmers receive in the DR has little to do with overt
subsidies of inputs (seed, irrigation, etc) and primarily to do with import protection. This

point will be reiterated and firrther supported in the following chapter which presents

benefit-cost analysis of bean research in the DR.

 

Table 5.21: Sensitivity Analysis of Enterprise Profit/ha, Red Beans, SJV, 1997/98
I l Gross Revenue/ha $1,681

Average Farmer Equip 8 Other/ha] $279

Input Cost ($USIha)‘ Op Cost Land/Equity Cap. $359

 

 

 

 

Labor
Cost‘ % $554 $514 $475 $435 $396 $435 $475 $514 $554
($US/ha) Change 40% 30% 20% 1 0% 0% 19% 20% 30% 40%
$178 -40% 31 1 350 390 429 469 429 390 350 31 1
$208 -30% 281 320 360 400 439 400 360 320 281
$237 -20% 251 291 330 370 410 370 330 291 251
$267 -1 0% 222 261 301 340 380 340 301 261 222
$297 0% 192 231 271 31 1 350 31 1 271 231 192
$326 10% 162 202 241 281 321 281 241 202 162
$356 ‘ 20% ‘ 133 172 212 251 291 251 212 172 133
$386 30% ' 103 142 182 222 261 222 182 142 103
$415 40% 73 113 152 192 232 192 152 113 73
$445 50% 44 83 123 162 202 162 123 83 44

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

" The column (row) with a 0% change represents the 1998 average cost
Source: 1998 Producer Survey (yields, pricg); 1998 RK Surveyl

 

 

 

 

 

 

 

These results also indicate that efforts to reduce costs of production, while helping
to maintain or improve profitability, will not be able to compensate for similar percentage
reductions in yields or prices. Of course, every effort to reduce costs of production by
improving allocative efficiency will help maintain profitability. For example, the average

farmer could lower his total production costs by 6 percent simply by reducing his seed

143

density to the CRSP-recommended level.

5.14.4 Farmers’ Sensitivity to Price Declines

In order to measure farmers’ sensitivity to bean price declines, respondents were
presented with a hypothetical situation regarding expected bean prices for the upcoming
seasbn. Respondents were asked if their primary buyer reduced their price RD$100/qq '
below the price they paid this year (1998), would the respondent continue to plant beans?
Ifthe respondent said they would continue to plant beans, they were then asked how they
would respond to an expected price decline of RD$200 pesos/qq, and so on until a decline
of RD$500 pesos/qq was reached (by which point all had indicated they would. switch to
an‘alternative crop with the exception of one farmer who said he’d plant beans regardless
of the price). The results of this hypothetical market situation are shown in Table 5.22.

These results imply that if the farmgate bean price fell by 10 percent”, ceten's
pm'r'bus, 38 percent of SJV farmers would not plant beans. Ifthe farmgate price fell 20
percent (approximately), 67 percent of SJV farmers would not plant beans. As the
responses of small farmers indicate, they are more sensitive to an initial 10 percent drop in
price, compared to large farmers. However, some small farmers claim that they would
continue to grow beans, even if prices dropped as much as 30 to 40 percent. By contrast,
large farmers are initially less sensitive to a 10 percent price drop, although they more
quickly switch out of beans as the magnitude of price declines increase. While this type of

hypothetical market situation is certainly not a substitute for actual price variation in

 

2’ The price elasticity of supply estimates of these farmers are derived fiom the
price levels of each individual farmer in the survey. Thus, these results are not
generalizable across other price levels.

144

assessing farmer price responsiveness, all of these farmers are commercial producers.
Thus, this hypothetical market provides a usefirl insight into their likely response to falling
bean prices.

Table 5.22 Farmer Reactions to Declines in Hypothetical Farmgat‘e Bean Prices,

919 38 912 46 967 26
819 67 812 63 867 72
719 84 712 78 767 87
619 90 612 82 667 94
519 99 512 98 567 100

‘ The average price received by each group was: sample $RD1019/qq, small 1012, and large 1067.

Source: 1998 Producer

 

Respondents were also asked what alternative crops they would plant, if they
stopped growing beans. For small farmers, the primary alternative activities included
pigeon peas (34 percent), maize (26 percent), nothing (12 percent), livestock grazing (8
percent), sweet potato (7 percent), rice and peanuts (4 percent each). For large farmers,
the primary alternative activity included rice (29 percent), nothing (19 percent), livestock
grazing (12 percent), maize (11 percent), sweet potato and pigeon peas (10 percent each),

papaya, onion and aguacate (3 percent each).

145

5.16 Alternative Cost of Production Figures

_ 5.16.1 San Juan Valley

There are several alternative bean cost of production figures available for the SW.
First, SEA reports bean cost of production to be $1,304/ha for “high input level” farmers
in the southwest region (i.e the San Juan Valley) (SEA, 1998). This figure is
approximately 29 percent higher than that obtained for large farmers in this study
($1,015/ha). Another figure comes from a report published by the farmer association
Productores Agricolas, Inc., and estimates the SJV “high input” bean cost of production
to. be $1,053/ha". This figure is approximately 4 percent higher, than the result obtained in
this study. The CRSP production cost budget considered the same cost items as both the
SBA and farmer. association budgets. '

5.16.2 Rainfed Bean Production

The distinction between the yield and profitability performance of San Juan Valley
farmers, eompared with rainfed farmers in other regions, can only be demonstrated
through use of government estimates of rainfed bean production costs. SE ’3 “natio ”
rainfed bean budget estimates cost of production for these farmers to be $763/ha, which is
27% less than the average SJV cost of production figure ($971/ha). However, this SEA
figure is likely to be inflated, given the difl‘erence between the CRSP and SEA bean
production budgets for the SJV. Using the SEA figure for an average rainfed farmer

(arbitrarily adjusted downward 25 percent to $5 72/ha, and then adding the opportunity

 

2"Land rental cost is removed from the Productores Agricolas’ original total in
order to make a comparison with the other two cost of production budgets.

146

cost of land yielding $686/ha), and assuming an average rainfed yield of 600 kg/ha, the
cost per unit of the average rainfed farmer is $1 . 14/kg, which is actually lower than that of
the average SJV farmer ($1.17/kg). At the adjusted rainfed cost of production level and a
farmgate price of $1 .32/kg (SEA’s reported annual farmgate price for 1998), the average
rainfed farmer with a yield of 520 kg/ha will just break even.
5.17 Chapter Summary

While the financial rate of return at the farm level is high, the financial profitability
of smaller farmers is sensitive to variation in yield and output price. Although this chapter
highlights various opportunities to decrease farm-level production costs, the prospect of
lower farmgate bean prices due to increased imports (because of GATT commitments as
well as falling domestic production levels) does not bode well for producers with below
average yields and above average costs of production. The long-term competitiveness of
beans at financial prices in a more open economy (i.e. economic profitability) is
questionable, given the nature and extent of government involvement and support in
supporting output prices as well as significant subsidies of bean inputs such as seed. While
economic analysis in Chapter 6 shows that societal welfare would increase with lower
bean prices, the adjustment costs of such a policy adjustment will be borne by those
producers least apt to find suitable alternatives without public investments in alternative

crop production and marketing assistance.

147

CHAPTER SIX: ECONOMIC IMPACT OF BEAN RESEARCH IN THE D.R.
6.1 Introduction
’CRSP bean research in the DR has focused primarily upon improving productivity

through the development of both improved varieties and management practices. Varietal
development has balanced demands for increased yields with greater diseases
tolerance/resistance. The first technology analyzed in this study is the red bean variety
PC-50, whose primary benefits are improved yields, tolerance to rust and common
bacterial blight, and better grain uniformity.l The principal benefit of P050 is higher
yields for irrigated producers. The second technology, analyzed in tandem with PC-50
adoption, is the fallow period in the San Juan Valley (SJV), which is credited with
controlling BGMV in the SJV and thus averting enormous bean yield losses.

' The standard hypothesis posed by ex post benefit-cost analysis is whether or not
the incremental’ economic benefits of the adoption of a technology - such as PC-50 -

outweigh the costs of developing and extending that technology (as well as any

 

’ More uniform grains typically receive higher prices. While it is possible that
producers and/or consumers have benefitted fiom the value of more uniform grain,
measuring the farmgate value of grain quality improvement would require a sample
including both PC-50 and Pompadour Checa growers, which is not possible for the SJV.

2 The economic concept of opportunity cost is central to benefit-cost analysis. The
benefit of a new technology is not the total revenue it generates, but the incremenml
revenue relative to the farmer’s next most renumerative use of his/her resources, which is
assumed to be the “without technology” scenario. For example, take a farmer who
received $1 .39/kg for Pompadour Checa and who had an average yield of 1,080 kg/ha.
Alter adopting PC-50, assuming he received the same price for his beans, and assuming
his new average yield was 1,242 kg/ha (15 percent higher than with Pompadour Checa),
then the benefit generated by adoption was not his new total revenue ($1.39/kg * 1242
W) but rather his incremental revenue ($1.39/kg * (1,242 - l,080)kg/ha)).

148

incremental costs to producers or consumers resulting from adoption). However, to be
considered profitable, the investment must not only cover its operating costs, but also
generate more return than the next-best alternative investment, as reflected in the
opportunity cost of capital. In addition, donors are often interested in more than solely
aggregate-level indicators of efficiency. Thus, the distribution of benefits and costs among
various producer and consumer groups is also of interest. The distribution of benefits and
costs depends ilpon the nature of the technology, the resource allocations of producers
and consumers, and the institutional and policy environment that shapes the opportunity
sets of producers and consumers.

This chapter presents the data and results from benefit-cost analysis of the two
principal technologies developed by CRSP bean research in the DR: the new bean variety
PC-50, and the fallow period in the SJV. In this chapter, the returns to PC-SO alone
(assuming the fallow period as a given) are presented first in their entirety, before the
returns to PC-50 combined with the fallow period are discussed. Each of these analyses
start with discussion of underlying assumptions made, clarifying the difl‘erences between
the with- and without-technology scenarios that are central to benefit-cost analysis.
Second, the source and estimation of the variables used in the analysis and the estimation
of the financial and economic rates of return are presented. Finally, sensitivity analysis and

the distribution of benefits are discussed.

149

6.2 Assumptions of the PC-50 Analysis

6.2.1 Farm-level Benefits fi'om the Adoption of P060

Agricultural technology is only welfare-improving if it is adopted’. Since benefits
accrue only fiom adoption, the adoption rate and the incremental net benefit per hectare
must be estimated. However, it is widely known that the farm-level productivity
improvement derived from a given technology often varies across farms with difi’erent
resource endowments. This is especially true if the technology is developed specifically
for a given resource environment. For example, many first-generation Green Revolution
rice and wheat varieties were developed specifically for use with fertilizer and irrigation as
complements.

Based on the recommendation of CRSP scientists, this analysis assumes that PC-
50 only improves yields on irrigated bean area (Amaud, 1998). This assumption is
significant because this excludes the more than half of the DR’s annual bean area planted
that is rainfed (at least some of which is planted to PC-SO, according to the 1993 CRSP
nationwide survey). However, the CRSP/DR program has followed a common
agricultural research model which assumes that the biggest aggregate return will come by
focusing research on higher-resource farmers. The distributional results of following this

strategy given the DR situation are discussed below.

 

3 There can certainly be spillover benefits of technology developed in one country
or area which is either directly or indirectly adopted and used in another to improve
productivity, which is an implicit goal of the CRSP’s regional initiatives. For example, the
new black bean variety Arroyo Loro Negro developed in the DR was not widely adopted
in the DR (due to overwhelming consumer preference for red beans), but this variety has
recently been extended to Haiti, where it is in high demand by farmers and consumers (see
the forthcoming CRSP impact study by Shields/Bemsten).

150

Estimates of the percentage of bean area in each region under “high”, “medium”,
and “low” bean production technology (Sterling, 1998; UEPA, 1990; SEA 1999) were
used to divide actual aggregated (irrigated and rainfed) data on red bean area in each
region into three production technology categories (Appendix B), as defined by SEA
“High” technology implies high input levels and irrigation, “medium” technology implies
medium input levels and-irrigation, and low technology implies low input levels and
rainfed production. Therefore, only the PC-50 area under high and medium technology is
considered for this analysis.

6.2.2 Incremental Farm-level Production Costs

The benefit-cost analysis is constructed by starting with representative farm-level
budgets and aggregating them to the project level. In this analysis, the incremental farm-
level production costs of adopting PC-50 are assumed to be negligible because this variety
is typically used with the same input package as was used with the previous variety
(Pompadour Checa or Jose Beta). The only incremental cost of PC-SO adoption is
assumed .to be increased harvesting costs per hectare (given higher yields). Because the
following analysis assumes a 15 percent incremental yield, the assumed increase in
harvesting costs per hectare is 15 percent.

6.2.3 Incremental Farm-level Input Use

Key informant sources suggested that attributing yield increases since the early
1990s to PC-50 alone (ceteris paribus) ignores the potential role of changes in other

factors of production on improved bean productivity, such as irrigation improvements and

151

increased fertilizer use. In fact, annual data from 1989 to 1998 on estimated‘ bean costs
of production (fi'om SEA) do not indicate increased fertilizer rates in bean production in
the Southwest region during this period. At least one rapid appraisal source in the SJV
concurs that fertilizer usage has remained more or less constant before and after the
introduction of PC-50 (Amaud, 1998). Therefore, this analysis assumes that fertilizer use
(and thus costs) did not increase due to PC-50 adoption. With regard to the potential
effect on productivity of irrigation improvements, less than 10 percent of the SJV
irrigation system was improved during the time period under review (Paniagua, 1998).

6.2.4 Economic Surplus Approach

The following benefit-cost calculations follow the economic surplus approach to
benefit-cost analysis (Masters, 1996). The supply shift associated with adoption of new
agricultural technology is usually assumed to be the result of an increase in total
production, resulting in a horizontal shift in the supply curve. The supply curve can also
shift vertically; up if costs of production per hectare increase and down if they decrease.
In the case of ex post analysis, observed prices and quantities over the period under review
already include the efl‘ects of research. Thus, the supply shift is based on estimating the
magnitude of net cost reductions given the observed level of output, and then making an
adjustment for the change in quantity associated with a change in price. However, the

estimate of the net cost reductions is generally the most important determinant of the

 

‘ SEA’s bean costs of production data were estimated via annual consultations
with regional extension agents. While the data were not collected with the same attention
to detail as the CRSP’s record keeping data, SEA’s cost of production figures are fairly
representative.

152

results (Masters, 1996). The equation that follows (Table 6.1) does not include net
production cost increases at the farrnlevel (as in Masters’ version); this analysis aggregates
these farrnlevel costs to the natiOnal level (although first aggregated by region and type of
production system), and then subtracts them fiom the total benefit stream (as with ’
research costs).

Table 6.1 Economic Surplus E nation Definitions and Formulas

 

 

 

 

Incremental production j = (AY " t "' A) / Q AY: incremental yleld (kg/ha)
due to the new - . t: adoption rate

technology, as a A: total bean area (ha)
proportion of total Q: total bean production (kg)
production

Net change in k = (j / e8) e8: price elasticity of supply
production Costs, as a

proportion of the

product price

Change in the AQ = (Q * eS * eD * k) eD: price elasticity of demand
equilibrium quantity /

produced due to the (eS * eD)

new technology

Economic benefits fi'om SG = (k "' P * Q) - the second term is subtracted for
the adoption of the new V: (k * P " AQ) ex post analysis

technology -

 

 

 

 

 

Source: Masters, 1996.

Because the economic theory underlying the economic surplus method of benefit-
cost analysis proposes that a shifi in the supply curve is responsible for changes in
producer and consumer surplus, the analyst must make various assumptions regarding the
supply and demand curves of the given commodity. Assumptions include the choice of

firnctional form of the supply curve, the nature of the shift of the supply curve, and supply

153

and demand elasticities. The choice of functional form is generally considered to be much
less significant than the choice of the nature and magnitude of the supply shift (Alston,
1995). .F9‘ this reason, the supply and demand curves are assumed here to be linear for
expediency. By contrast, the nature of this supply shift is a key determinant of the
distribution of benefits from research between producers and consumers, and on the
measured level of total research benefits (Linder and Jarrett, 1978). A parallel shift of the
supply curve implies the same absolute reduction in average costs for both low cost

and high cost producers. In other words, a parallel shift describes the case in which the
innovation lowers the per unit production costs of low cost producers (those representing
the lefi portion of the supply curve) by a. substantially greater percentage than for those
producers with higher costs (ibid).

Given that only irrigated producers are assumed to benefit fi'om PC-SO, and that
high technology bean farmers obtain greater benefits than do medium technology farmers,
a parallel shift seems appropriate for this analysis. Therefore, in this analysis, the supply
curve is assumed to be linear, and the supply shifi is assumed to be parallel. Supply is
expected to be price inelastic, given that bean is a staple crop for rainfed farmers, and
given. that valley farmers (for whom bean is a cash crop) enjoyed high and stable prices
prior to 1904. Econometric estimation of supply response confirms this expectation; the
price elasticity of supply is estimated to be 0.50 (Mather, 1999b)’. Demand is also

assumed to be inelastic (0.7) because of the strong tradition of beans as part of the mid-

 

’ However, a structural change appears to have occurred in 1994, the year that the
Dominican government signed the GATT agreement and a pro-reform party came to the
presidency. Alter 1994, the price elasticity of bean supply becomes elastic (1.15).

154

day meal and because bean remains the staple protein of the poor.
6.3 Financial Benefit Stream
6.3.1 Adoption Rates and Time-line
6.3.1.1 Inggdugion

A PC-50 was publicly released in 1989 to growers in the San Juan Valley (SJV) and
around the country. Although SEA funded a “PC-50 adoption” study in June 1990, the
sample included only farmers “known” to have sown PC-SO. Therefore, farmer adoption
rates were not reported (Diaz, 1990). In June-July 1992, a CRSP-funded nationwide
adoption. study of PC-50 found the following farmer adoption rates at that point in time:
San Juan Valley (SJV) 59.5%; Cibao Valley 30.6%; Higuey Valley 12.3%; and other
regions 3.9% (Heikes, 1995). The 1998 CRSP Producer survey in the SJV provides
additional information about adoption of P050 in the SJV, the main bean producing
region in the DR

Adoption studies traditionally fit a logistic curve to known data (adoption levels at

given points in time determined through surveys) to estimate the difiiision of technology
adoption among farmers over time. The logistic curve contains three parameters: origin,
slope, and ceiling, as defined by the curve Pt = K / 1 + e" " ”0 (Griliches, 1957). P, is the
percentage of either area or farmers employing the new technology at a given time t, K is
the ceiling, b is the rate of grth coefi'rcient, and a is the constant of integration which
positions the curve on the time scale. Use of the logistic curve to approximate technology
adoption over time makes intuitive sense in that the parameters capture several aspects of

empirical diffusion dynamics. For example, lags in either local seed demand (knowledge

155'

of and/or desire for) or supply (availability) in given areas help explain differences across
farms and across regions in the time and scale of adoption (differences in origins). The
curve also captures differences in how quickly adoption proceeds (slopes or rates) toward
a long-run equilibrium use of a given technology (ceiling). The use and method of
estimating the parameters of the logistic function for use in this cost-benefit analysis are
discussed in the following sections.
6.3.1.2 W

PC-50 adoption rates from the 1998 CRSP producer survey in the SN are similar6
to those of the 1992 study. The actual adoption of P050 in the SJV (based on the 1998
survey results) proceeds more quickly to the ceiling value than would be expected from
the traditionally-assumed smooth logistic curve7 (Figure 6.1). Because many farmers in
the SN were familiar with P050 before it was ofiicially released', adoption therefore
proceeded rapidly; the percentage of SJV farmers who adopted P050 in the first year of

release (1989) was 23 percent (1998 CRSP survey).

 

‘ The 1998 survey asked respondents when they first used PC-50, thus an adoption
curve for the SJV was created from the cumulative adoption information provided by the
respondents. However, because the 1998 sample only included producers who grew
beans in 1998, the adoption curve calculated fiom this sample may not include individuals
who did not adopt PC-50 and have since gotten out of been production. Thus, the
percentage of adopters in the first few years may be biased upwards. Nevertheless,
because the 1992 level of 60 percent in the 1998 survey is almost identical to the 1992
level found in Heikes’ 1993 survey, the 1998 survey results are thus considered unbiased.

7 Logistic curve calculated assuming a ceiling of 95% in 1995.

' Many SJV farmers were familiar with PC-50 prior to its public release because
they were already growing PC-50 for seed multiplication, they had participated in semi-
commercial testing of PC-SO, or they simply knew about it from contact with CRSP
personnel or association colleagues. .

156

The logistic curve has an initial period of increasing adoption at an increasing rate,
followed by a period of increasing adoption at a decreasing rate, followed by a peak
adoption rate or ceiling. In addition, it is sometimes assumed that the ceiling rate can be
followed by a decline in adoption as the variety either degrades or is replaced by farmers
with newly improved varieties. In fact, the PC-50 adoption rate experienced such a
decline following the release of IB-17 8, a new red variety developed by the CRSP, which
was first released to several producer associations in the SJV in 1996. Adoption rates of
JB-178 were 15% in 1996 and 36%; in 1997, with 1997 being the first year that JB-l78
seed was available directly from SEA”. .While most farmers planting JB-17 8 continue to
devote some of their bean area to PC-SO, the per-farmer use rate of PC-50 nevertheless
has fallen (Table 6.2). Because INESPRE purchases JB-178 at the same price as P050,
and SEA contracts some growers in its multiplication program to deliver JB-l78 seed, it is
therefore not surprising that per farmer use of PC-50 area has declined since 1996, the

year JB-l78 was introduced to the SJV.

 

9 JB-178 adoption rates in the SJV are from the 1998 Producer Survey and Amaud
(199s). ‘

157

 

Figure 6.1: PC-50 Adoption
San Juan Valley

 

 

 

 

 

 

 

 

W ,3 2
0%4 r“ i

888990919293949596979899000102
‘ Year

l+ Actual (SJV) -x- Logistic Curve

Cumu‘late % farmers adopting PC-5Q,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

However, while JB-l78 has taken area away from PC-SO in the SJV, it is not clear
at this point whether or not it will replace PC-SO entirely due to consumer demand for
specific bean attributes. The Jose Beta market class (JB-178 is similar in shape, coloring
and cooking time to Jose Beta) has never been nearly as preferred by consumers as
Pompadour Checa (PC-50). On the other hand, at least one-third of consumers of various
income levels in Santo Domingo and the SJV indicate that freshness and cooking time are
their principal desired attributes (Mather, 1999). These consumers often preferred
imported pintos to local red varieties because of better relative fi'eshness, even though they
preferred the color of PC-50. Because JB-‘178 is known to have a shorter cooking time

than PC-SO, it is possible that consumers will increasingly demand JB-178 as they become

158

more aware of this variety.

T bl 62 P d ' ' ' ' SJV 1997-98

 

PC-SO ' P050 79 64
JB- 1 78 $2178 36 24
Jose Beta Jose Beta 7 2

Other White: Blanca, Anacaona 9 10

Black: Negra, Venezuela 44
Large Red: ConstanZa I
Other Red: Pompadour Checa, 131 % 100%
Checa, Indiana
Pinto: Pinta, Gira
Cranberry: Jacomela

 

 

 

 

' N= 100. Percentages add to more than 100 because some producers plant more than one bean variety.
1 Area represents all land dedicated to bean production for the winter 1997-98 season. This data are
aggregated from parcel level data.

S—o-urce: 1998 CRSP Producer Survey

 

 

 

The San Juan Valley PC-50 adoption curve, based on data collected during the
1998 producer survey, serves in this cost-benefit analysis as the PC-SO adoption rate for
irrigated bean area in the Southwest Region (i.e. a logistic curve was not fit to data points
given that the survey provided all the data points). The SJV adoption curve is likely to be
different (i. e. rise faster and have a higher ceiling) from those of the other two main bean
producing valleys (Cibao and Higuey) because of the presence of many SEA seed
multiplication program growers in the SJV, and because the CRSP has been active and/or

based in San Juan since the project’s" inception in 1984.

159

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 6.2: PC-50 Adoption
Cibao/Higuey & Other Regions
810W.
9 v
E 80%
‘6.
0 ch-
'8 60%
8
g ..
g 40%
a! ..
0
3 20%
g
3 In-
E
”0‘70- .iiiiiiiiii
0 . as so 92 94 96 98 oo 02
- Years 7
I--- Cibao Valley -x- Higuey Valley -— Other Regions]

 

 

6.3.1.3 MW

Logistic adoption curves for the Cibao (N orthcentral Region) and Higuey (East
Region), the two other major areas of production, are estimated by using point estimates
fiom the 1992 Heikes/CRSP nationwide survey (31 percent adoption in the Cibao and
12.3 percent adoption in Higuey), and assuming that in 1988 the adoption rates in these
two valleys were 1 percent each (i.e. essentially zero). Logistic curves are then fit using
estimates of b derived fiom these two point estimates, and assuming an upper limit of 80
percent. The upper limit of 80 percent is implicitly less than that of 91 for the San Juan
Valley given that the SJV has the majority of SEA seed program growers and the presence
of the CRSP. Given that the SEA seed program switched from multiplication of

Pompador Checa to that of P050 in 1990, and given available SEA Seed Department

160

data on red bean seed distribution by region (see Appendix B), these adoption curve
assumptions seem reasonable. These curves are presented in Figure 6.2, along with the
curve for Other Regions.
6.3.1.4 W

The adoption curve for Other Regionslo are estimated by using point estimates
fiom the 1992 nationwide survey (3 .7 percent adoption), and assuming that in 1988 the
adoption rates in these regions were 1 percent each (i.e. essentially zero). The adoption
ceiling of 80 percent is again assumed. This is a reasonable assumption given SEA seed
distribution data and the fact that only irrigated area is included in this analysis.

63.2 Red Bean Area by Region

Data fiom SEA on red bean area by region fi'om 1985-99 is used along with the
estimated. adoption curves described above to estimate the P050 area by region fiom
1989-2002. Although the proportion of irrigated area, as a percentage of total bean area,
has varied over the period in question (1989-1999), it has averaged about 45 percent".
Therefore, this analysis assumes that annual irrigated area is 45 percent of total red bean

area, as reported by SEA.

 

‘° Other Regions refers to irrigated bean area outside of the San Juan, Cibao, and
Higuey valleys.

“ Regional and/or national data on irrigated/rainfed proportions of bean area are
only available for 1985-1990, 1993, and 1999; 45 percent irrigated is the average of 1990,
1993 and 1999.

161

6.3.3 Red Bean Yield by Region
6.3.3.1 Yield Difl‘erentige
Annual red bean yields by region fiom 1985-99 (SEA) are used to account for
annual yield fluctuations across the DR Because this data aggregates regional yields, it
therefore includes production from both rainfed and irrigated area in a given region.
Using the actual regional (aggregate) yields each year, the high technology PC-SO area in
each region is assumed to have yields 30 percent higher than the regional aVerage, while
the yield of the medium technology PC-SO area is assumed to have yields 15 percent
higher than the regional average. '2 Due to the widespread adoption of PC-50 in the SJV,
the 1998 SJV producer survey could not estimate the counterfactual yield situation (i. e.
yields during the 19903 of irrigated farmers who never adopted PC-SO). Therefore,
incremental yield estimates are based “on research station trials and SJV farmers’
perceptions of the incremental yield of 'PC-SO compared with Pompadour Checa (fi'om the
1998 Producer Survey).
6.3.3.2 Incrementej Yielge
Semi-commercial yield trials by the CRSP project from 1985-89 showed that PC-
50. yielded an average 30 percent more. than Pompadour Checa (CRSP, 1990) across five
trial sites (in various regions). In the 1998 Producer Survey, 76 percent of farmers said

that their yields increased with PC-50 compared to Pompador Checa, 14 percent said their

 

‘2 Thirty percent was chosen as the yield differential between high technology
yields and the regional average because this is the average differential between irrigated
and aggregate yields across all regions. Fifteen percent was arbitrarily assumed to be the
yield difi‘erential between medium technology yield and aggregate regional yield.

162

yields remained the same, and 10 percent said their yields were lower with P050. Of
those who claimed to have higher yields, 72 percent said that their yields increased
between 30-60 percent.13 Given that experimental yields are rarely achieved by farmers in
practice, and given the potential rounding-error problems associated with the Producer
Survey question regarding the percentage increment, an incremental yield of 23 percent is
assumed, which is 75 percent of the experimental incremental yield. When this
incremental yield of 23 percent is weighted by respondents’ indications of yield increase,
stability, or decline due to PC-50 adoption, the average incremental yield is approximately
15 percent“. Thus, the incremental yield attributable to PC-50' alone is conservatively
assumed to be 15 percent for high and medium technology bean areas.
6.3.3.3 Ineremflal Preeuetien

Regional incremental production attributable to PC-50 can then be calculated

given the total bean area under high and medium technology, the regional adoption rate,

the incremental yield, and the actual regional yield for that technology area. IncMental

 

‘3 It should be noted that there were problems with the question asked of farmers
in the 1998 producer survey, “How many more/less quintales per tarea do you normally
get fiom PC-50 compared with your previous variety?” Bean yields in the DR are
reported in “quintals per tarea” (quintal = 100 or 110 lbs, depending on the buyer", 15.9
tarea = hectare), and the typical yield range is between 0.5 - 2.5 quintals per tarea.
Because coded response categories went up and down by half-quintals, it is very likely
that there is an upward rounding bias in the responses to this question. The implication is
that while 76 percent of producers said their yields increased, “incremen ” yields
reported by some farmers to be as high as 60-80 percent are not plausible and are likely
the result of the upward bias of rounding. Ifyield gains were as high as the average
farmer reported, these gains would surely be so noted in aggregate regional yield
improvements.

’4 That is, (0.77.123) + (0.14*1) + (0.101077) = 1.15

163

production fOr a given high technology area in a region is therefore calculated as:

Incremental production (kg) = area adopted (ha) * regional yield (kg/ha) * yield

difi‘erential for high tech area (%) "' incremental yield (%)

Thus, incremental production is calculated on a regional basis using actual regional
aggregate yield data and the assumptions presented. This incremental production is then
aggregated across regions and plugged into the Economic Surplus model as the numerator
of the j parameter (see below).

In sum, this analysis uses actual regional data, assumes that the annual yield for
high technology area is 30 percent higher than the annual regional mean, and then assumes
that 15 percent of the resulting high technology PC-SO production is attributable to variety
alone. This incremental yield (15% of the high technology yield for a given region in a
given year) is-multiplied by the adopted high technology area in that region for the given
year to estimate the incremental regional production under high technology area. A
similar calculation is then performed for the medium technology area in each region,
although in this case, the medium technology yield is assumed to be 15 percent higher than
the annual regional mean, and that 15% of the medium technology PC-50 production is
attributed to the variety alone. Once the incremental production due to P050 is
estimated, the remaining steps involve calculating “a real, discounted value of the national
incremental production for each year.

6.3.4 Red Bean Financial Prices

This analysis uses an annual national average farmgate price series for red beans in

the DR (SEA, 1998) as the nominal financial price series for the time period of adoption

164'

1989-1999. The price series for 2000-02 is assumed to increase by 5 percent a year to
approximate inflation.

6.3.5 CPI, Exchange Rate, and Discount Rate

The consumer price index (CPI)" used to deflate benefits and costs from 19.84-
1999 is from the Inter-American Development Bank Economics and Statistics Database
(base year -1999=100). The annual average exchange rate used is also obtained from the
Inter-American Development Bank. The discount rate is assumed to be 10 percent, as
inflation in the 19903 ranged fi'om 8 to 20 percent (averaging about 10 percent), and the
nominal interest rate has been 24 to 30 percent during this period.

Key informants as well as a World Bank study (V aldes, 1995) suggest that the
Dominican currency has not been overvalued since 1990. Because the figures prior to
1990 that'this analysis uses are research costs, which are either in SUS or are non-
tradeables", the use of a shadow exchange rate is not necessary.

6.4 Financial Costs Stream

6.4.1 CRSP Research Costs

The project-level costs involved in the development and difiirsion of PC-50 include
CRSP expenses for research and training, as well as financial contribution fiom SEA in the
form of technician salaries and the use of experiment station land. CRSP DR Project cost

data was obtained fi'om the CRSP Management omce at MSU, and includes primarily

 

" The CPI 1s used because the Dominican government does not calculate a
producer price index.

“ CRSP research expenses are reported in SUS, and SEA research expenses are
non-tradeables such as technician salaries and land values.

165

salaries, training, and operational costs. Both host-country and US-based expenditures

(charged to the DR project) from 1984-1989 are assumed to be relevant to the testing,

TOTAL Nebraska Nebraska P.Rico TOTAL Technician Land
YEAR CRSP HC US for HO HO SEA Salaries

1 83 480
1 1
1 1 1 959

0 $33 085 $25 12 473

 

screening and development of PC-50 and basic material for the SEA seed program.
However, CRSP DR technical reports of research activities from 1984-1989 indicate that
approximately one-half of CRSP researcher activities during this time period were related
to PC-50". Therefore, it is assumed that 50 percent of the total CRSP costs on the DR
program (those shown in Table 6.3) are attributable to the development and release ‘of PC-
50.

6.4.2 SEA Costs

As the CRSP is a collaborative project, the Dominican government (SEA)
contributed land and technician salaries to the project. The amount contributed by SEA is
estimated to be US$19,000 per year (nominal) for technician and worker salaries (this

figure is based on the amounts contributed by SEA in 1992 and 1993), and $5,544 for the

 

‘7 One-half of the total experimental area during this period was used for
screening, trials, multiplication, or other activities that involved comparisons of P060
with other material.

166

opportunity cost of 5.6 hectares of land (the total area of experiment station land used for
bean research). This assumes that a hectare of irrigated land could have brought financial
net return of $1,000/ha with two bean crops in a year in the late 19803. Extension costs
are not included as the extension service existed prior to the project, which implies that
extension costs are not incremental costs. The “extension cost” of PC-50 is assumed to be
embodied by the substantial SEA seed subsidy.
6.5 Financial Rate of Return to PC-50

Given the base assumptions discussed below, the financial rate of return to the
development of PC-50 during the period 1984-2002 is 96.2 percent. As financial ROR’s
are consistently higher than economic ROR’s in the case where production is
subsidized/protected, sensitivity analysis will only be reported for the economic ROR’s.
The Net Present Value of the investment is USS 63.9 million (valued at financial prices).
6.6 Economic Benefit Stream

6.6.1 Economic Import Parity Prices

The economic benefits stream uses the same area, yield, and adoption rates as does
the financial benefits stream. However, the economic bean price series are economic
import parity prices calculated using US. (Colorado) pinto bean season-average prices
(Table 6.4) during the benefit period (1989-2002)". Pintos typically constitute more than

90 percent of the DR’s bean imports, and in recent years, almost all pintos have been

 

" The cut-ofl‘ year of 2002 is arbitrary but represents a typical time-line of
evaluation of a technology (13 years fi'om release) as well as the reality that discounting
reduces firture benefits quickly beyond five years from the present. Adding more years of
benefits to the following analyses analysis does not change the conclusions and has very
limited .efi‘ect on the magnitude of the benefit-cost indicators.

167

imported fi'om the US. During the period 1939-99, the financial import parity price (IPP)
of Colorado pintos averaged 22 percent less than the annual DR average farmgate prices
for domestic red beans (PC-50, Pompadour Checa), while the economic IPP averaged 62

percent less than the farmgate price during this same period”.

season

% US$lcwt RD$lcwt

Rate 15.90
season

 

6.6.2 MarketPrice Premium for Domestic Red Bean
The use of the pinto as a parity good for the Pompadour market class assumes that

these goods are perfect substitutes and that the price difference is due solely to a

 

‘9 IPP calculations are made by the author using import costs from Table 6.5 and
pinto price series fiom USDA/AMS.

168

differential in the costs of production. However, this assumption would belie the fact that
for many consumers, pintos and domestic red beans are not perfect substitutes. In fact,
using retail price (data. that distinguishes between domestic reds and pintos”0 (fi'om August
1998 to June 1999), DominiCan consumers were willing (through 1999) to pay
approximately 20 percent more for domestic reds compared to pintos. However, these
market prices are not “ ee” market prices because pinto imports are still heavily
regulated, although import levels have increased dramatically the past few years. This
analysis will therefore assume that no consumer preference premium for red beans exists in
the base scenario, but will add a premium in sensitivity analysis to incorporate the disparity
between censumption value of the pinto compared to the domestic red.

6.7 Economic CostStream

6.7.1 Introduction .

There are several inputs in Dominican bean production that are either directly or
indirectly subsidized or taxed. Thus, economic analysis requires that these inputs be
valued withlsubsidies/taxes removed (Tables 6.5). Additional economic costs include
farm-level input subsidies (aggregated to the project level) fiom SEA (seeds), INDRHI
(irrigation), and PROSEMA (land services). The calculations of these costs/ha are
discussed below (Table 6.6). Additional economic benefits include lower fertilizer prices
that would occur under more competitive pricing. For computational convenience,

fertilizer savings per hectare are included in the economic cost section of the analysis. The

 

2° Prior to August 1998, SEA did not report retail pinto prices as separate fi'om the
“red bean” price category.

. 169

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

82.82 .888" ca .289 68522 E» 82 88885 an» noose»

38.83 8.23 88.8» 8.8» »2.8» 8.2» 08.8» 8.8» 88
68.83 8.23 28.8» 8.8» 838» 8.2» 838» 8.8» 88
$8.83 8.23 28.8» 8.8» 80.8» 8.2» e828 8.8» 88
88.33 8.23 38.8» 8.8» 88.8» 8.2» 2:8» 8.8» 82
£8.83 8.23 ~88» 8.8» 82.8» 8.2» 28.8» 8.8» 82
82.83 8.23 «8.8» 8.8» 88.28 8.2» 238 8.8» 82
838% 8.23 80.8» 8.8» 2888 8.2» 8.8» 8.8» 82
82.83 8.23 8.8» 8.8» 838? 8.2» 8.8» 8.8» 82
98.83 8.23 88.8» 8.8» 8.8» 8.2» 8.8» 8.8» 82
$8.23 8.23 . 88.2» 8.8» 88.8» 8.2» 2.8» 8.8» 82
€8.23 8.23 838 8.8» «8.3» _ 8.2» 85;» 8.8» 82
2:83 8.23 28‘» 8.8» 2%» . 8.2» 82.8 8.8» 82
#283 8.23 85» 8.8» 88.8 8.2» 28.8 8.8» 82
88.23 8.23 «8» 8.8» 8}» 8.2» 25» 8.8» 82
o» o» o» o» 82
o» o» o» o» 39
8 o» o» o» 82
o» o» o» o» 82
o» o» o» o» 82

8: 883 85 883 18: a; a 82 8: a; $w3m<m>

. . £38 85 8.3 53% 8:88 888 883 8285

can .to o> emEo ll. an FED—z. . humming
53>» :33 .3225". . _
3:253:50 «moo £50.38 .31 m6 2::

 

 

 

170

same CRSP and SEA project costs used in financial analysis are utilized for the economic
analysis... '

6.7.2 Indirect Fertilizer Tax

According to an on-going study of the [Dominican fertilizer industry (Nunez,
1999), fertilizer prices in the DR have been approximately 10 percent higher over the last
decade than they would have been under a more competitive pricing environment (two
firms share the entire DR fertilizer market and are considered to be oligopolists).
Therefore, the economic valuation of these inputs would be the economic import parity
fertilizer price (i.e. the international price plus international and domestic transport costs,
insurance, etc), which is approximately 10% lower than existing market prices.

6.7.3 Seed Subsidy

On the other hand, farm-level production costs would increase if farmers were
charged the full (economic) cost of seeds purchased from SEA, the full cost of irrigation
services from INDRHI, and the full cost of land preparation by PROSEMA In this
analysis, the value of these subsidies are aggregated to the project level (i.e. nationwide),
and are calculated using a subsidy value per unit of the input, multiplied by the amount of
that input used per adopted hectare each year.

The value of the SEA seed subsidy aggregated to the national level is calculated as
the subsidy per kilogram of seed multiplied by the total amount of SEA seed planted that
year. The subsidy per quintal (cwt) of seed is calculated as follows. .While SEA pays
RDSI, 17S/qq. for seed, the acquisition cost of seed-quality bean is assumed to be near the

upper end of the distribution of market prices for grain (RD$1,000/qq). The cost of

171

private “cold storage” is approximately RD$7/qq per month (RD$56/qq for 8 months
of storage)", and the costs of cleaning and treatment are approximately RDSSOIqq. The
opportunity cost of not selling the grain at RD$1,000/qq in February (interest over eight
months @ 2 percent/month) is RD$195/qq, and a 10 percent profit margin on operation
costs is RDSI lO/qq. Under these cost assumptions, a private firm would sell seed for no
less than RD$1,216/qq (U 881 .85/kg). Because the SEA seed price is RD$900/qq
(US$1.37/kg), this implies a subsidy of RDSB 16/qq (U SSO.48/kg) of all SEA seed actually
planted by farmers. To calculate the total subsidy per year, the subsidy per kilogram of
seed ($0.48/kg) is multiplied by an assumed average seed rate (123 W or 17 lb/ta)
across the DR Thus, each farmer who purchased seed from SEA in a given year received
a subsidy of (US$58.9l/ha). This result is then multiplied by the total PC-SO hectares
adopted each year, then multiplied by 30 percent (the assumed turnover rate for
purchasing new seed)22 (Table 6.5 and 6.6). .

6.7 .4 Irrigation and Land Preparation Subsidies

To calculate the aggregate value of subsidies on irrigation, the subsidy of a
season’s irrigation per hectare supplied by INDRHI (US$19.75/ha in the SJV) is

multiplied by the adopted hectares each year (all adopted hectares are irrigated).

 

2' This monthly storage fee is the rate quoted by a bean wholesaler in San Juan
(who has a cold storage facility) in August, 1998.

. 22 Although data on annual SEA seed sales by region is available and presented in
Appendix B, the method used here to calculate the seed subsidy does not use these figures
due to their suspect accuracy._ However, annual SEA seed sales by region as compared
with annual area planted by region does support the assumption that approximately 30
percent of bean area any given year is planted to seed purchased fiom SEA that year.

172

Likewise, the subsidy per hectare of land prepared by PROSEMA ($40/ha in the SIV) is
multiplied by the adopted hectares each year and then by 25 percent, as this is
approximately the percentage of national bean area serviced by PROSEMA (SEA, 1998)
(Table 6.5 and 6.6).
6.8 Economic Analysis of PG”

681 Economic Rate of Return

Given the base scenario assumptions discussed below, the economic rate of return
to the development of PC-SO during the period 1984-2002 is 63.2 percent. Therefore, this
ex-postanalysis concludes that given the base assumptions made - namely, taking the
fallow period as a given, the development of PC-SO has made the Bean/Cowpea CRSP
project a profitable investment through 2002 (the economic ROR is greater than the
discount rate/opportunity cost of capital of 10 percent). The Net Present Value of this
investment over the same time period is US $28.0 million.

Although bean production in the SJV is uneconomic (Chapter 5), the economic
ROR is positive because of a seemingly paradoxical situation: PC-SO yielded positive
returns because it made a “bad” economic situation “less bad.” That is, the import
protection policy of the DR creates large deadweight losses and results in large transfers
from consumers to producers, which is the result of the production of uneconomic crops
(Chapter 5). Taking the trade policyregime situation as a given, a farmer using PC-SO has
higher (yet still negative) economic returns than a farmer who does not use P050 (and
has lower negative economic returns). Thus, PC-SO reduces the cost to society of import

protection (the deadweight loss). In this case, the “with P050" scenario is an

173

improvement in economic terms over the “without PC-SO" scenario.
8 6.8.2 Sensitivity Analysis

As several key variables in the rate of return function are based on assumptions,
analysis of the sensitivity of the economic ROR to the assumed values of these variables is
presented below. The initial value of these variables include: discount rate (10 percent),
elasticity of demand (0.7), elasticity of supply (0.5), incremental yield (15 percent), price
difi‘erential between imported pintos (IPP) and domestic reds (20 percent), total
CRSP/SEA project costs from 1984-89 (50 percent attributed to the development of PC-
50), and cost of production increase due to increased harvesting costs (10 percent). Table
6.7 demonstrates the sensitivity of the economic ROR to variation in the values of the said
variables, ceteris paribus (assuming all other variables remain at their base levels - in bold
below).

As demonstrated in Table 6.7, the economic ROR is most sensitive to the assumed
values for the price elasticity of supply, the percentage of CRSP/SEA total bean research
costs attributed to PC-SO development, and the incremental yield incease. However, even
when these variables'are changed (ceteris paribus) to maximum/ minimum values, the
investment remains profitable. The economic ROR is relatively insensitive to changes in
the price elasticity of demand, level of government input subsidies (seed, land prep, and

irrigation), and incremental farm-level harvesting and fertilizer costs, ceteris paribus.

174

Table 6.7 Sensitivity Analysis of the Economic ROR to PC-SO Adoption in the DR,
1984-

. an... n..“‘6'._....._-.u'\._u‘.'u ._. . 5...I 6.. 'u'.‘-\'.‘.‘.' .I. g I .I. .' .I- a - . a . -.- .l.'u_o_«'-_-.. n I'd-I.I.l.l.l':-I'l.:.l.-’;'l 5:. ...........

    

 
 

 

 

 

 

 

 

 

% Adopted 0 53.6 % Total 50 53.6
Area CRSP
(decrease) Costs
. - - - _20 , 49.3 (1984M) 100 40.6
-30 ’ 46.8 % Import . 0 49.9
Parity Price ,
0.4 58.2 Adjustment 5 50.8
Price 0.5 . 53.6 (13... - Pm.) 10 51.8
. . - (increase)
Elastmty of 0.6 49.9 15 52.7
Supply
- 0.7 46.8 20 53.6
0.8 44.2 . % s ' 32.5
Incremental ‘ -
0.9 41.9 Yield . 10 45.5
(“”3“) 15 53.6
20 59.6

 

 

 

 

 

6.8.3 (Distribution of Benefits

6.8-3.1 1413;945:3192
The distribution of the benefits and costs of technical change depends upon the

policy situation and the structure of production and marketing systems (which includes
consumer income, preferences, and their distribution): Critics of Green. Revolution
agricultural technologies decried the bias of these technologies towards larger, resource-
rich farmers who had easier access to requisite complementary inputs such as irrigation
and fertilizer. However, Scobie and Posada demonstrated that given the right

circumstances, technology with thisbias can still improve the welfare of a vast majority of

175

urban and rural poor (even many non-adopting farmers). For example, in their study of
the impact of technical change in rice production in Colombia, Scobie and Posada (Scobie,
1990) showed national supply increases large enough to substantially drive down rice
prices, thus benefitting'all consumers and all farmers who were net rice buyers. While

. they conceded that some commercial upland rice farmers were net losers fiom this
teclmical change, the national welfare improvement for the urban and rural poor was far
greater than the isolated losses.

This same economic argument is often employed to silence criticism of working
exclusively with resource-rich farmers, as is the nature of CRSP research in the DR. The
argument, goes that the “bang for the buck” of working with rainfed producers is simply
too low to justify the investment, and, more importantly, when resource-rich farmers’
yields increase enough to drive down domestic prices of that staple, then all net buyers of
the staple (including all of the urban poor and many of the rural poor) are better ofl‘.
Therefore, given that the analysis above demonstrates that CRSP bean research in the DR
had a profitable aggregate impact, it is still worth asking the question of whether or not
consumers received any benefits from this technical change, and which producers have
benefitted.

6.8.3.2 Consumers

Due to the nature of the Dominican government’s supply control policies
implemented via import restrictions, it is assumed that incremental domestic production
did not expand total supply, but rather merely displaced imports. The DR scenario is not

similar to the typical import quota scenario (Alton, 1995) in which incremental production

176

expands total supply (because the size of the import quota remains the same), thus
lowering retail prices. In fact, the DR scenario is more similar to the typical import tarifl’
scenario (Altson, 1995) in which incremental production displaces potential imports and
does not lead to an expansion of total supply - retail prices remain constant, and
consumers receive no benefits from the technical change. Therefore, this analysis indicates
that consumers have received little if no benefits from incrementalbean yields; producers
(farmers, importers, marketing agents) have captured all the benefits fiom technical
change. ‘

Real Dominican retail bean prices in fact only began to fall in the late 1990s as
Dominican production fell dramatically and the government began to substantially increase
imports of pintos, which — even after the inefficiencies of tariffs, rents, etc. (described in
Chapter 4) -, are cheaper than domestic reds and thus put downward pressure on domestic
red retail prices. One scenario does exist, however, in which Dominican consumers could
benefitted fi'om the technical change. Assuming a price premium for domestic red beans
would exist under free market prices, then the incremental production due to PC-SO
multiplied by this premium would constitute consumer surplus because “displaced
imports” are not valued as highly by consumers. However, this assumes that consumers
who favor the domestic reds (and are willing to pay a premium) are not made better ofl‘
when import levels increase- driving down the price of their favored domestic reds. In

short, analysis into consumer surplus due to a price premium is not considered here.

177

6.8.3.3 Producer}

Contrary to the usual result in agricultural technical change, the benefits of
incremental yields in the DR have accrued primarily to producers. Among producers,
benefits have accrued principally to those with irrigation. The financial benefits to SJV
farmers have been the greatest, as they have received the highest farmgate prices due to
INESPRE and SEA purchases at above-market prices. Since small farmers sell at lower
prices, most of the benefits accrue to medium and larger farmers in the valleys. As noted
in Chapter 5, smaller farmers receive lower prices for their beans in part due to generally
lower quality beans, as well as to financial and organizational limitations.

6.9 ’Financial and Economic Analysis of The Fallow Period
I 6.9.1 Introduction

By the accounts of many key informants, the fallow period has enabled the San
Juan Valleyto efl‘ectively control thewhitefly and to thus continue to grow bean in the
winter. season. However, the fallow period in the San Juan Valley has meant that since
1991-92, farmers no longer have the option of planting both an early fall bean crop (or
other host crop for the white fly) prior to the main winter bean season. In short, if the
fallow period is to be considered an output of CRSP research, then the “without-fallow
period” benefit-cost scenario would have to include both more bean area and lower yields
(due to continued BGMV problems that would have occurred in the absence of the fallow

period) than would the “with-fallow” scenario”. Therefore, the fallow period represents

 

’3 It should be noted that not all bean growers planted in both seasons, and the
winter season has always been the largest for bean production in the SJV.

178 *

the opportunity cost — foregone income from an early bean season - of the “with-fallow”
scenario that enables a lower incidence of BGMV and thus higher yields in the winter bean
season.

With regard to benefit-cost analysis of the fallow period, the “with-fallow”
scenario is historical; that is, the fallow period was enacted in 1991, and the approximate
SJV annual bean area since that time is known. While good data on yields specific to the
SJV are. not available, historic regional yields on irrigated area provides a close
approximation, as the SJV makes up almost all the irrigated area in the Southwest region.
By contrast, the “without-fallow” scenario is hypothetical; the SJV area under both an
early fall and a winter bean season must be assumed, as well as BGMV incidence and
subseciuent yield loss. To estimate the area loss of the early fall season embodied in the
with fallow period scenario (i. e. a gain in the without scenario), the average bean area for
the SJV fi'om 1986 to 1990 was compared with the average bean area for 1992 to 1997
(the fallow period began in 1991). The difi‘erence between these averages is
approximately 900 ha. This small change is likely due to the fact that many farmers did
, not plant beans in both seasons. Thus, by moving to one season, area planted did not
decrease substantially. Therefore, the “without fallow” scenario includes 900 ha/year in

addition to the historical bean area for the SJV (Table 6.8).

179.

 

Table 6. 8 The SJV Fallow Period In With- and Without-Fallow Scenarios, 1990-02

 

 

With the fallow period historical 1998 survey average yield'

 

 

 

Without the fallow historical + 900 1998 survey average yield - 33.75
period ha/year" percent yield loss per year due to
increased incidence of BGMV

 

‘ Actual yields embody losses to BGMV; 25 percent of 1998 CRSP Survey respondents
reported BGMV as their principal production problem. The 1998 survey average yield
is used as time series data for alternative crops in the SJV is not available.

" Yield loss calculated as follows: 25% chance of 75% yield loss due to BGMV + 25%
chance of 50% yield loss + 25% chance of 20% loss + 25% chance of 10% loss =
average of 38.75% yield loss per year. However, this analysis assumes that historical
yields embody a 25% chance of 20% yield loss per year + 75% chance of 0% loss, thus
the incremental yield loss is 38.75% - 5% = 33.75%.

 

 

 

With respect to BGMV incidence and yield loss, this analysis assumes that without
a fallow period, BGMV incidence and yield losses in the SJV would have been higher.
Specifically, this analysis assumes that the “without fallow” scenario would have resulted
in an expected average of 33 .75 percent lower production than what actually occurred in
the “with fallow” scenario (Table 6.8). The frequencies and respective yield percentage
losses are based on the assumption that in any given year (without the fallow period),
there would be a 25 percent chance of a devastating BGMV incidence (75 percent losses),
a 25 percent chance of a serious incidence (50 percent losses), a 25 percent chance of a
“normal”.year (20 percent losses), and a 25 percent chance of only 10 percent losses. The
assumed fi'equencies and losses are based very roughly on anecdotal estimates (Amaud,
1995, 1996, 1997) of BGMV incidence and bean crop losses in the SJV since 1988, as

well as survey responses (25 percent of respondents in 1998 claimed that BGMV was

180

their principal production problem) (CRSP Survey, 1998).

However, under these expected average yield losses in bean production, bean
would not have remained more profitable than alternative crops using financial valuation in
the hypothetical “without fallow” scenario: For example, net returns analysis”
using bean and alternative crops (Table 6.9 to 6.12) and assuming 33.75 percent yield loss
on bean in'the “without fallow” scenario”, demonstrates that SJV bean farmers facing
financial (market) prices in 1991 would have switched out of winter production of beans
and into alternative crops, to a combination of maize (66 percent of area previously

planted to bean) and pigeon pea (34 percent)“.

 

2‘ The crop budgets in Tables 6.9 to 6.12 were constructed as follows: SJV
cropping area is based on the historical data on SJV bean area; crop prices and alternative
crop costs of production are national averages from SEA (SEA, 1998); bean yield is the
average for 1997 fiom the SJV Producer Survey; alternative crop yields for the SJV are
from a local development project (JICA, 1998); bean costs of production are from the RK
survey; economic prices for bean are from the import parity price series discussed in
Chapter 6.6.1; the alternative crops pigeon pea and maize are not protected. The
alternative crops considered in this analysis were two of the crops indicated by the 1998
CRSP Producer survey respondents as the crops they would switch into if bean prices fell
dramatically. The other alternative crops indicated by respondents have limited market
demand and were therefore not considered as viable alternatives for the majority of
producers. This analysis assumes that producers are profit-maximizers that make
decisions based upon expected prices, costs of production, and yields. Producer risk
preferences are not considered, and no price effects of increase maize or pigeon pea are
considered. Nevertheless, it seems clear that pigeon pea and maize demand is large
enough to accommodate these growers, and that any price efi‘ects would not make these
crops less profitable than bean in the without fallow scenario.

2’ Tables 6.9 to 6.12 assume constant yield over time because no time series data
on- alternative crop yields is available for the SJV. Assuming constant yields for bean and
alternative crops for this simple analysis would not likely change the result found, given
that the area in question is irrigated.

2‘ This proportion between maize and pigeon pea is based on the 1998 CRSP
farmer survey respondents’ “likely alternative crops” as well as the fact that the pigeon pea

181

To examine this result more closely, consider that the time series of bean net
returns under the “with fallow” scenario (Table 6.9) in comparison with the net returns to
pigeon pea (Table 6.11) and maize (Table 6.12) demonstrates that bean is the most
financially renumerative crop in the srv” (net returns of uss 972/ha for bean in 1991,
compared with US$ 679/ha for pigeon pea and USS 669/ha for maize), assuming no yield
loss from BGMV. It is also clear bean has lower economic returns than the alternative
crops”. However, once yield losses to BGMV (averaging 33 .7 5 percent per year) are
assumed for the “without fallow” scenario (Table 6.10), it is clear that both pigeon pea
and maize would be more financially renumerative (net returns of USS 679/ha for pigeon
pea in 1991, "USS 669/ha for maize, and uss 316/ha for bean). Thus, assuming no
transaction costs, risk-neutral farmers would have switched to these alternative crops
during the 1991 to 2002 period. While the data used in this simple net returns analysis is
not ideal, even these rough estimates of net profitability over time indicate that SJV
farmers do have alternatives to bean, and would have switched to these crops if BGMV

had not been controlled.

 

market is smaller than that for maize.

2" While there are other alternative crops which have much higher returns
(potatoes, sweet potatoes, tomatoes, etc), these are not considered viable alternatives for
this analysis given their small market demand and the transaction costs associated with
developing stable marketing coordination.

2' The net returns to bean and the alternative crops do not include the opportunity
cost of land or equity capital, which explains why bean net economic returns are not
negative as they are in Chapter 5.

182

Table 6.10: Financial and Economic Returns to Bean SJV: Without Fallow

 

assumes

183

 

184

Given this result, the ensuing benefit-cost analysis computes the incremental benefit of the
fallow period for a given year by subtracting the value of the “without fallow” scenario
(the aggregate value of net returns to alternative crops) from the value of the “with '
fallow” scenario (the aggregate value of net returns to bean production) (Table 6.13).

Table 6.13 With- and Without-Fallow Scenarios Assuming Switch to Alternative
Crops, 1990-02

 

 

   
 

 

 

 

With the fallow historical bean area value of bean production: net returns

 

 

 

 

 

period to bean production (SUS/ ha)
i - multiplied by the bean area
Without the fallow historical bean area + value of alternative crop production:
period 900 ha/year net returns to alternative crop
production (W S / ha) multiplied by
the (former bean area + 900 ha) .

 

6.9.2 Intuitive Results

‘ Given from what has already been presented regarding the financial and
economic returns to bean and alternative crops, it is possible to intuit the results of
benefit-cost analysis of the fallow period. In the “with fallow” scenario, bean is financially
more profitable than alternative crops, yet economically less profitable. By contrast, in the
“without fallow“ scenario, bean (with large yield losses) is less financially profitable than
alternative crops. Thus, we conclude that farmers would have switched to alternative
crops in this scenario. This means that farmers are expected to be better ofl‘ in financial
terms with the fallow period (U 88 972/ha for bean, compared with US$676/ha for pigeon
p.ea) However, because the alternative crops under economic valuation are more

profitable than bean, society is much better off In economic terms with out the fallow

185

 

 

period, because farmers would have switched to alternative crops which are more
renumerative in economic temis (USS 676/ha for pigeon pea compared with USS - 346/ha
for bean).
. Benefit-cost analysis subtracts the value of the “without” scenario from the “with”
scenario. Thus, under financial valuation, the fallow period represents a large gain for
SJV bean farmers, while underoeconomic valuation, it represents a large loss to society.
The reason for this dichotomous result is that since bean is much less profitable in
economic valuation than alternatives, and since uncontrolled BGMV would have forced
farmers to switch to alternative crops (and take financial losses), thus the fallow period
enables farmers to maintain their relatively high returns fi'om bean while preventing society
from gaining the benefits of more emcient allocation of resources (namely, if the SJV
stopped producing beans, the country would import more beans which are cheaper, thus
benefitting consumers greatly).

-. 16.9.3 Aggregate Financial and Economic Benefits of the Fallow Period

‘ To aggregate the financial benefits of the fallow period, the annual net returns per

hectare of beans and alternative crops (Tables 6.9 to 6.12) under financial valuation are.
multiplied by the number of hectares of bean in the SJV, as per the format of Table 6.13.
The positive difference between the net returns to bean per hectare and the net returns to
alternatiVe crops per hectare in a given year is the benefit to SJV bean farmers of the
fallow period, which enabled them to continue to plant bean. The results (Table 6.14)
concur with the intuitive result, that the fallow period enabled SJV farmersto avoid

switching to less renumerative alternative crops, thus resulting in very large benefits which

186

Area

7,332

9 393
10

9.

6 164
6,412
7 634
,634
7,634
.634

' "with fallow" scenario,
fallow" scenario
are

'“the

able 6.14:

Net Return
to Beans'

744
1 930
1 476

of the
Financial
Net R
to alt

,457,02
18,132,
15,139,
12, 12,

.058,

,029,
.311.
.311,
.313,
.313,
been

able 6.1

bean

allow

alt

,443,986
,447,768
,443,668
288,943

,484
,367,
,139,071
.961.

, 1,008
,961,008

are

maize

13,038
10,684

.223:
31 ,
,2

172,355
,418
301
301

are assumed to ,1

to

 

accrue to the farmers when comparing the difference between aggregate returns to bean

relative to alternative crops.

. The same analysis is then completed by aggregating net returns under economic

valuation. In this case, the difi‘erence between the net returns to bean per hectare and the

 

net returns to alternative crops per hectare are large and negative. Thus, the actual results
(Table 6.15) concur with the intuitive result, that the fallow period enabled SJV farmers to
continue to grow bean, resulting in large financial gains to bean farmers (as compared with
alternative crops) and large economic losses to society (as compared with alternative

crops).

187

 

6.15:
Economic
Returns Net Returns
ear to Beans bean to all

' 1 131' 1 .443: , 1
1993 1 1,890,111 ,447,768 ,557,657

1995 2 ' :286 :288:943 ' ,643
193 1,992,65 ,738,898 ,746,24

1998 ' -- ,171 367,386 ,810,21
‘ 1999 121 1, 19 ,139,071 ,060,590
2001 1 117,117 ,961,008 1
1 117,11 ,961, , ,1
*, the , are assumed to 1,
“ in the "without fallow" scenario, bean are assumed to be 7
are. able 6.1 and maize able 6.13

 

6.10 Financial and Economic Analysis of P050 and the Fallow Period

6.10.1 Combining the Two Technologies in Benefit-Cost Analysis

. Whereas the previous benefit-cost analysis has isolated the financial and economic

returns to PC-50.and the fallow period as separate technologies, the following analysis
focuses on the returns to these two technologies implemented in tandem. Simply looking
at the net benefit streams of each technology separately indicates that the fallow returns -
both under financial and economic valuation - are larger than those of PCs-50. Thus, the
fallow period greatly magnified both the financial benefits to irrigated farmers and the
economic losses to society.

However, when considering the twotechnologies together, the analysis. must be

divided into two parts; one for the SJV and one for the rest of the DR. The following

188

 

 

analysis assumes that the scenarios for the rest of the DR remain unchanged by the fallow
period”. from the original with- and without-PC-SO scenario. For the STV scenarios, this
analysis assumes that the incremental benefit of P050 adoption is essentially embodied in
the constant yield assumed for the fallow period analysis. In other words, we assume that
this constant yield would be approximately 15 percent lower over the 1991 to 2002 time
period were it not for widespread PC-SO adoption. In the previous PC-SO analysis, we
assumed that for adopters, a given percentage of historical yields across the DR was
attributable to PC-SO. This yield was then multiplied by the adopted area in a given year.
Therefore, the following analysis simply combines the fallow result for SJV bean area with
the PC-SO result for bean area outside the SJV.
. 6.10.2 Research Costs of the Fallow Period

This analysis assumes that there were minimal research costs either by the CRSP
or by SEA in the development of the idea for the fellow period. The idea of prohibiting
the growing of host crops for a specific insect or disease during a period of time is not
new. While the CRSP did incur some costs in optimal planting date experiments in 1991
and againin 1993-94, these costs were very small and thus not quantified in this analysis.

6.10.3 Financial Return to PC-SO and the Fallow Period

Given the base scenario assumptions discussed above, the financial rate of return

to the development of P050 and the srv fallow period during the period 1984-2002 is

 

' ’9 This assumes there would be no price effects from the “without fallow” scenario
in which SJV farmers shifi out of beans (foregone bean production in the SJV would be
replaced with cheaper pintos, which would likely lower the financial price of beans across
the DR).

189

 

 

87.8 percent. The Net Present Value of this investment over the same time period is
$91.4 million. This financial ROR is larger than that for PC-SO alone (73.2 percent),
which follows the intuitive prediction above. .

6.10.4 Economic Return to PC-SO and the Fallow Period

Given the base scenario assumptions discussed above, the economic rate of return
to the development of PC-SO and the SJV fallow period during the period 1984-2002 is
not calculable as it is negative. This is because theidiscounted “net benefits” of the fallow

period under economic valuation are negative and much larger in magnitude than the

 

positive benefits from PC-SO. The Net Present Value of this investment over the same
time period is USS -47 .2 million. Therefore, this ex-posr analysis concludes that these two
technologies analyzed in tandem 'make the Bean/Cowpea CRSP project an unprofitable
investment through 2002 (the economic ROR is negative and uncalculable, and therefore
less than the discount rate/opportunity cost of capital of 10 percent).

6.10.5 Paradoxical Result

This result says that PC-SO alone - with the fallow period as a given - yielded '
positive returns because it made is “poor” situation “less worse.” That is, the import
protection policy of the DR creates large deadweight losses and results in large transfers
from consumers to producers. Taking that situation as a given, a farmer using PC-SO has
higher (negative) economic returns than a farmer who does not use PC-SO (and has lower
negative economic returns). Thus, PC-SO reduces the cost to society of import protection
(the deadWeight loss). In this case, the with scenario is an improvement in economic

terms over the without scenario.

190

 

 

 

However, because the without-fallow period scenario would have been an
economic improvement over the with-fallow scenario, the enactment of the fallow period
meant that society missed an opportunity for net welfare improvement (transfers from
producers to consumers, and reductions in deadweight loss). Therefore, this foregone
opportunity is counted as a loss to society, and since this effect is larger in magnitude than
the gains to society fi'om PC-SO adoption, the overall economic efi'ect is negative.

6. 110.6 Distribution of Benefits

Analyzing PC-SO combined with the fallow period simply exacerbates further the
benefit distribution of PC-SO — irrigated producers are even better off than they were just
from PC-SO adoption, consumers still do not share in the benefits, and society now
actually loses because the opportunity for more efficient allocation of resources is
foregone.

6.11 Chapter Summary

6.11.1 Aggregate Results

The ex post benefit-cost analysis presented in this chapter demonstrates that the
development and adoption of PC-SO has resulted in an economic ROR superior to the
opportunity cost of capital, as well as an .even higher financial ROR The combination of
high adoption rates of PC-SO with a favorable price premium in the market place with
respect to the lower-priced imported pintos, helped to make the economic rate of return to
CRSP research on PC-SO profitable - notwithstanding that import parity prices of pintos
are significantly lower than financial prices of domestic red beans. Thus, the CRSP DR

program from 1984-2002 is considered economically profitable on the basis of P050

191

 

 

alone. Other activities and achievements of the CRSP, such as development and
promotion of improve cultural practices in the SN, the institutional development of the
SlV bean research team, potential spillover effects of DR research to the US, other CRSP
countries, and to Haiti, or the existence value of the genetic base compiled by the DR stafl'
are not included in this analysis. However, if the SJV fallow period is considered an
output of the CRSP, joint analysis of P050 and the fallow period lead to the result that
the CRSP "research in the DR is not profitable in economic terms.

Benefit-cost analyses are typically undertaken to justify future investments in
agricultural research or continued funding of the existing investment under investigation.
Aside fi'om performing the necessary benefit-cost analysis, a goal of this study was to
demonstrate the role played by the policy environment in shaping the past, present, and
firture net benefits of technical change, as well as on the distribution of these net benefits.

Import protection is used by the DR because production costs are higher than
import costs, but the, import protection program has an associated economic cost. As
development and adoption of new bean technology reduces production costs, the
discrepancy between production costs and import costs falls, thus reducing the economic
costs of the import protection program. Therefore, taking the import protection as given,
bean research lowers the economic cost of the import protection program and thus has a
positive economic impact.

6.11.2 Distributional Results

Ifthe goal of benefit-cost analysis is simply to discern whether an investment is

profitable or not,“ then CRSP bean research in the DR has indeed been successful.

192

However, the socioeconomic irnpact of bean research in the DR almost exclusively has
been to increase the per hectare returns to irrigated bean farmers in the San Juan Valley.
This distributional result is due simply to the realities of the economic environment facing
bean farmers, the structure and demographics of the bean production and marketing

system, and the CRSP approach towards working within that environment and system.

193

CHAPTER SEVEN: SUMMARY OF FINDINGS
7.1 Introduction

The general objective of this study is to assess the ex post impact of PC-SO and the
SJV fallow period. Ex post benefit-cost analyses are typically undertaken to justify future
investments in agricultural research or continued funding of the existing investment under
investigation. In addition to performing this necessary ex post analysis, a goal of this
study was to demonstrate the role played by the policy environment in shaping the past,
present, and future net benefits of technical change, as well as on the distribution of these
net benefits. This chapter summarizes the results of the subsector, profitability, and
benefit-cost analyses in this study and provides policy recommendations to the CRSP.
7.2 Summary of Findings

7.2.1 Subscctor Analysis

Subscctor analysis (Chapter 4) demonstrated that the policy environment in the DR
from the mid-19809 to the present has been conducive to adoption of new bean
technologies for several reasons. First, bean production has remained profitable in the DR
due io import protection. Second, SEA’s role in multiplying seed and distributing it
across the country (combined with a subsidized price) played an important role in
technology dissemination. Third, farmer associations in the SJV helped facilitate high
adoption rates of varietal and management technologies (1'. e. the'fallow period), and
lobbied on behalf of the nation’s bean producers to maintain import protection. The first
and third reasons were especially important, given that the CRSP technologies were

developed specifically for irrigated producers, for most of whom bean is a cash crop.

194

Thus, in the absence of favorable output prices, the adopters of CRSP technologies would
have switched to alternative crops years ago (as has happened in the Cibao Valley, which
shifted into tobacco and other export crops).

7.2.2 Profitability Analysis

profitability analysis in Chapter 5 demonstrated that bean production in the San
Juan Valley is profitable under financial valuation but not under economic valuation,
considering the costs of local production relative to international prices, thus highlighting
the sensitivity of Dominican bean production to maintenance of current import restrictions
and input subsidy policies. Although the chapter highlights various opportunities to
decrease farm-level production costs, the prospect of lower farmgate bean prices in the
future does not bode well for those producers with below average yields and above
average costs of production. The long-term competitiveness of domestic bean production
in a more open economy is doubtful.

7.2.3 Benefit-Cost Analysis of PC-SO

Benefit-cost analysis (Chapter 6) demonstrated that the development and adoption
of the new bean variety PC-SO was a profitable investment for the CRSP during the 1984-
2002 period under economic valuation. Although import parity prices of pintos are
significantly lower than financial prices of domestic red beans, high adoption rates
generated a benefit stream large enough to more than cover investment costs. These high
adoption rates are principally the result of relatively high farm-level prices due to import
protection, as well as ample seed distribution via subsidized government-managed seed

multiplication and distribution. Although bean isuneconomic in the SJV (Chapter 5), PC-

195

50 adoption decreased the magnitude of economic loss that would have occurred without
PC-SO, thus generating a positive incremental benefit.

This seemingly paradoxical result can be explained in more general terms. Import
protection is used by the DR because domestic production costs are higher than import
costs, but the import protection program has an associated economic cost. As
development and adoption of new bean technology reduces production costs, the
discrepancy between production costs and import costs falls, thus reducing the economic
costs of the import protection program. Therefore, taking the import protection regime as
given, bean research lowers the economic cost of the'import protection program and thus
has a positive economic impact.

7.2.4 Benefit-Cost Analysis of P050 and the Fallow Period

When the SJV fallow period is considered an output of the CRSP, joint analysis of
120-50 and the fallow period lead to the result that the casp research in the DR is not
profitable in economic terms. In sum, the CRSP has generated technologies in the SW
which have been widely adopted and have greatly enhanced the financial returns to
irrigated producers. However, regardless of this success in scientific toms, because the
CRSP is working with an uneconomic crop, the economic return to the successfill
generation and adoption of CRSP technologies in the SJV is negative. Continued import
protection means that SJV farmers are able to maintain high financial returns at the

expense of welfare gains for consumers in the form of lower priced imported beans.

196

 

 

7.2.5 Distribution of Benefits from Technological Change

If the goal of a benefit-cost analysis is simply to discern whether an investment is
profitable or not, then CRSP bean research in the DR has indeed been successful.
However, deciding that any given profitable project is the best use of scarce resources for
investment in agricultural research is akin to claiming that since the Dominican economy
has been growing at 7 percent a year since the mid-1990's that most poor Dominicans
must be better ofi‘. That is to say that the socioeconomic performance of the investment
should be compared with the stated socioeconomic goals prescribed by the CRSP to see
whether or not the project impact meets those stated ends. The socioeconomic impact of
bean research in the DR has primarily been to increase the per hectare returns to irrigated
bean farmers. Consumers have received little benefits from the investment in PC-SO, and
the SJV fallow period lefi consumers and society worse ofi‘. This is due simply to the
realities of the economic environment facing bean farmers, the structure and demographics
of the bean production and marketing system, and the CRSP approach towards working
within that environment and system.
7.3 Prospects for Future Economic Impact of the CRSP in the DR

7.3.1 General Outlook

casp research has successfillly developed improved varieties and disease
prevention practices since 1984. However, past and existing government policies towards
beans have created a situation in which production and marketing risk has increased,
resulting in declining area planted and hence decreased regional and national production.

Future domestic bean production (and filture CRSP impact) hinges primarily on the nature

197

of future government activities in bean purchasing and marketing, the government’s
policies regarding import levels and import administration, and consumer preferences for
domestic. beans .with respect to imported pintos. For example, if SEA continues to scale
back INESPRE’s bean purchases in the San Juan Valley, the decline SJV bean area since
the mid-19903 will continue. Second, if imports are allowed to remain at 1997 to 1999
levels, there will not be a resurgence of bean area in the near future because retail,
wholesale, and farmgate bean prices will continue to fall. Third, if import licences are
administered by true public auction, wholesale pinto prices in Santo Domingo will fall 10
to 30 percent, and'domestic red retail prices will subsequently fall. Finally, if consumer
preferences change such that pintos are seen as near perfect substitutes (or preferred) to
domestic red beans, this will put further downward pressure on domestic bean retail
prices.

These likely events all point toward continued decline in irrigated bean area both
within the SJV - the primary focus of CRSP research - and across the country. In fact,
national irrigated bean area has decreased 50 percent since 1989, and 35 percent since the
mid-1990s. In 1998, national irrigated bean area totaled 29,000 ha. Declining irrigated
bean area means lower prospects for future economic impact of the CRSP in the DR

7.3 .2 Research for Rainfed Producers

The CRSP could significantly increase its potential future impact in the DR by
focusing at least some of its resources on technologies that could benefit rainfed
producers. The project argues that they are prohibited by Dominican law fiom promoting

bean production in zonas Iaderas (steep-sloping zones). In June of 1990, the Dominican

198

legislature passed a regulation prohibiting SEA fi'om distributing seeds and other inputs to
producers in zones with higher than 30% slope. The rationale was based on the fact that
bean production on lands with greater than 30% slope produced low yields and was
environmentally destructive. The resulting soil erosion was filling up the country’s dams
at an alarming rate, threatening bean and other production in the irrigated valleys (Heikes,
1995; SEA Breve, 1990).

However, there are two reasons why it is not clear why this should prevent the
CRSP fiom promoting technologies for rainfed farmers‘. First, according to the
nationwide 1993 CRSP Producer survey, approximately 75 percent of rainfed producers
(42 percent of the total sample) had sloped land, and 25 percent had steep-sloped land.
Thus it is far from clear that all rainfed producers farm on steep-slopes. Second, the SEA
law says, in fact, that the government may not work with steep-sloped bean producers
unless it involves development and extension of soil conservation technologies (SEA
Breve, 1990). In Honduras, both CIAT and the CRSP are carrying out research to
develop and extend new varieties and management practices (including the use of
vegetative strips to alleviate soil erosion) appropriate to the needs of limited-resource
hillside farmers.

7 .33 Extension of Technologies

It seems clear fiom both the successes and limitations of the DR project that the

CRSP can no longer afi‘ord to ignore the critical issue of technology extension. The

 

1

It should be noted that the author did not specifically research this possibility in the field.
This is an area that would require further investigation.

199

success of the DR project in meeting many of its stated goals is to some extent due to
proactive measures by project personnel to use farmer associations as vehicles for
technology dissemination. However, a serious limitation of the DR project is that first, its
impact to date has been diluted by its sole focus on the SJV, and second, that its future
impact will increasingly depend upon its ability to develop and extend technologies to
farmers outside of the San Juan Valley, given declining irrigated bean area in general
(most of which is in the SJV).

Unfortunately, the typical USAID response to this situation is to ask the CRSP to
simply do more with the same (or less) funding. Given that USAID will not likely increase
firnding to the CRSP in the near future, a more realistic approach would be for the DR
CRSP to reallocate resources toward activities that will facilitate the development and
dissemination of CRSPrdeveloped technologies to rainfed bean producers. For example,
given situations common across the CRSP, in which the national public extension system
is non-existent or behaves as such, CRSP fitnds designated to train NGO stafl‘ -
organizations that work directly with small rainfed farmers - could perhaps firrther extend
at least some of the CRSP technologies (those appropriate for the conditions of the small
rainfed producers) far, beyond the time, logistical, and knowledge constraints of the CRSP
personnel themselves.

7.4 Limitations of This Study
I The principal limitation of this study is its insuficient focus on bean production
outside of the San Juan valley. In addition, the study lacks sumcient ex ante focus

intended to guide the CRSP in a direction that would increase fixture impact. The irony of

200

this study-is that the ex post result confirms the worth of the investment to date in the DR
project, yet present economic realities show that the project will not likely continue the
same success given its current strategy. Another limitation is that other notable activities
and achievements of the CRSP, such as development and promotion of improved cultural
practices in the SJV, the institutional development of the SJV bean research team,
potential spillover effects of DR research to the US, other CRSP countries, and to Haiti,
or the existence value of the genetic base compiled by the DR stafi‘ are not included in this
analysis. However, while these achievements could potentially improve the project’s
economic return, without refocusing the strategy of working exclusively with and for
irrigated producers in the SJV, only large spillover efi‘ects in other countries (Haiti) would
enable the CRSP to claim that the DR project was improving welfare of poor producers or
consumers:
7.5 Recommended Socioeconomic Research

First, ex ante benefit-cost analysis using various area and yield projections would
help demonstrate more clearly why the project needs to extend its focus beyond the San
Juan Valley. Second, future research needs to focus on institutional and other constraints
to increasing the project’s ties with NGOs and government omcials outside of the SJV
that are necessary to expand the project’s potential for future impact. Specifically, rapid
appraisal outside of the SJV is necessary to determine the potential for the development of

CRSP technologies in areas beyond the SJV.

201

APPENDIX A: Validity of the Sample Frame

The validity of the sample fiame, assumed to consist entirely of 1,001 farmers fi'om
the lists of the 17 farmer associations in the SJV, is an important concern due to the bias
that association of a farmer could represent for resulting “population” statistics. There are
several reasons to believe that 95 percent or more of the SJV bean farmer population is
included within the farmer association membership lists (all of which were obtained before
sampling was initiated). First, several key informants, who work in various valley-wide
projects, attest to this figure. Second, the mean farm size of the weighted sample (130.78
ta) multiplied by the 1001 farmers listed yields 130,910 ta, which is quite close the
139,000 ta of irrigated beans planted in the southwest region during the winter 1997-98
bean season (Sterling, 1998). Third, the economic power that association members enjoy
is clear: unequaled access to government subsidized seed, credit, seed storage, and, most
importantly, government purchases of beans by INESPRE at a negotiated (high) price.
Lastly, the group Productores Independientes (Independent Producers) formed several
years ago so that formerly non-associated bean farmers could band together to receive
what amounted to flee seed from the government (at that time), which was only
distributing this valuable input via associations. In addition, at this time, INESPRE was
beginning to purchase beans primarily through associations, meaning that individuals vying
to sell their beans at the high INESPRE prices would have to store their beans'for months
while association members went first. Productores Independientes is the largest and most

geographically diverse farmer association in the SJV, which is representative of the fact

202

that it incorporates members from across the valley who were in most cases not associated

previously.

203

 

 

APPENDIX B: Percentage of Regional Bean Area by Technology Level

% of bean area tech level and

0/o of bean area tech level and re

 

204

APPENDIX C: SEA Seed Sales

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SEA Seed: Distribution of Sales (100 cwt)

Year Central East N. Central N. West N. East North South 8. West Total

1990 3,483 2,370 4,299 700 1,917 1,536 9,728 21,369 45,402
1991 4,376 2,977 5,400 880 2,408 1,929 12,220 26,844 57,034
1 992 5,444 3,704 6,719 1,095 2,996 2,400 15,203 33,397 70,957
1993 1,725 2,600 5,562 500 2,300 2,720 3,500 22,917 41,824
1994 12,743 4,100 2,200 2,547 5,056 2,049 10,700 25,693 65,088
1995 450 800 2, 750 1 00 140 600 6, 300 1,686 12,826
1996 1,430 1,200 508 150 645 450 4,310 29,265 37,958
1997 1,322 750 1,412 256 905 250 3,912 9,100 17,907

 

1990-92 totals from SEA Diagostico; 1993-97 data by region is from SEA Seed Dept Annual Reports

 

1990-92 regional distribution estimated from 1993 distribution percent

 

agiby regio

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l l l l l

Potential PC-50 area by flion @239 lblha (15 Iblta) plus 10% SEA and farmer stor e
Year Central East N. Central N. West l N. East North South S. West Total
1990 1,312 893 1,619 264 722 578 3,663 8,047 17,097
1991 1,648 1,121 2,034 331 907 726 4,602 10,108 21,477
1 992 2,050 1,395 2,530 412 1,128 904 5,725 12,576 26,720
1993 650 979 2,094 188 866 1,024 1,318 8,630 15,750
1994 4,799 1,544 828 959 1,904 772 4,029 9,675 24,510
1995 169 301 1,036 38 53 226 2,372 635 4,830
1996 538 452 191 56 243 169 1,623 11,020 14,294
1997 498 282 532 96 341 94 1,473 3,427 6,743

DR Red Bean Area Planted b r Region (ha)
Year Central East N. Central N. West N. East North South S. West Total—
1990 3,844 2,818 4,231 332 3,764 3,992 7,363 13,436 39,780
1991 2,516 2,908 2,839 539 1,953 4,164 6,980 14,780 36,677
1992 2,554 2,756 4,759 153 1,755 2,441 6,855 14,897 36,171
1993 1,661 3,166 5,074 350 2,385 3,311 4,736 25,929 46,614
1994 1,625 3,264 5,238 377 1,391 2,590 3,1 89 19,700 37,374
1995 1,534 4,856 4,676 376 2,033 1,844 7,875 20,795 43,991
1996 1,620 2,230 4,400 160 1,866 1,341 5,047 20,442 37,105
1997 2,584 4,212 3,736 247 1,190 1,471 4,021 14,946 32,406

Percentage (%) of planted area attributed to SEA direct seed sales
Year Central East N. Central N. West N. East North South 8. West Total
1990 34 32 38 79 19 14 50 60 43
1991 66 39 72 62 46 17 66 68 59
1992 80 51 53 269 64 37 84 84 74
1993 39 31 41 54 36 31 28 33 34
1994 295 47 16 254 137 30 126 49 66
1995 11 6 22 10 3% 12 30 3 11
1996 33 20 4 35 13 13 32 54 39
1997 19 7 14 39 29 6% 37 23 21

 

Source: SEA Departamento de Semilla, Memorias

205

 

Although-there are several cases in the final section of the table above which
indicate inconsistent data (regional percentages of area planted to PC-SO greater than 100
percent), it is clear that ifthis data has any reliability, then PC-SO was in fact distributed

widely across the country and across considerable bean area.

206

REFERENCES

Alston, Julian, George Norton, and Philip Pardy, Science Under Scarcity: Principles and
Practices for Agricultural Research and Priority Setting, Cornell University Press,
Ithaca, NY, 1995.

Amaud Santana, Eladio, Juan Cedano, and Rodys Colon, “Pulso Productivo del Cultivo
de Habichuelas en el Valle de San Juan, 1993 -94,” Proyecto Titulo XII
Habichuela/Cowpea, May 1994.

Amaud Santana, Eladio, Juan Cedano, Rodys Colon, and Segundo Nova, “Pulso
Productivo del Cultivo de Habichuelas en el Valle de San Juan, 1995-96,”
Proyecto Titulo XII Habichuela/Cowpea, May 1996.

Amaud Santana, Eladio, Segundo Nova, and Rodys Colon, “Pulso Productivo del Cultivo
de Habichuelas en el Valle de San Juan, 1996-97,” Proyecto Titulo XII
Habichuela/Cowpea, July 1997.

Amaud Santana, Eladio, CRSP/DR Annual Report 1997, San Juan de la Maguana,
Dominican Republic, 1997.

Banco Central de la Republics Dominicana, Irlforme de la Economia Dominicana, 1997.

Byerelee, Derek, and Steven Franzel, “Institutionalizing the Role of Economists in
National Agricultural Research Institutes,” CIMMYT Economics Working Paper
93-01, Mexico, D.F.: CIMMYT, 1993.

Capalbo, SM. and 1M. Antle, Incorporating social costs in the Returns to Agricultural
Research, American Journal of Agricultural Economics 70 (2):281-288.

CRSP/SEA, PC-50 release brochure (reprint), 1997.

CRSP/SEA, JB-178 release brochure, 1997.

CRSP Record-Keeping (RK) Survey in the San Juan Valley, November 1997 - February
1998, directed by the Bean/Cowpea CRSP Dominican project Principal

Investigator, Dr. Eladio Amaud-Santana, in consultation with Dr. Rick Bernsten,
Bean/Cowpea CRSP Economist.

207

CRSP Producer Survey in the San Juan Valley, July-August 1998, directed by David
Mather (graduate research assistant and MSc/PhD candidate in the department of
Agricultural Economics, Michigan State University), under the direction of
Bean/Cowpea CRSP Economist Dr. Rick Bernsten; in consultation with the CRSP
Principal Investigator for the DR, Dr. Eladio Amaud Santana.

Daniels, L., James F. Oehmke, J. Howard, M. Maredia, and R. Bernsten, The Impact of
Agricultural Research: A Review of the Err-Post Assessment Literature with
Implications for Africa, Department of Agricultural Economics Stafi' Paper No.
92-3 8, Michigan State University, East Lansing, MI, 1992.

Diaz, Vianela, and Ramon Almanzar, “Resultado de la Encue'sta Aplicada a los
Productores de Habichuela de la Variedad PC-SO,” Spectra Consultores, S.A.,
1990.

Echeverria, Ruben, ed., Methods for Diagnosing Research System Constraints and
Assessing the Impact of Agricultural Research: Volume II Assessing the Impact of
Agricultural Research, Proceedings of the ISNAR/Rutgers Agricultural
Technology Management Workshop, 6-8 July 1988, Rutgers University, New
Jersey, USA International Service for National Agricultural Research, The

Hague, 1990, pp. 1-35.

The Economist, Country Profile: lhe Dominican Republic. The Economist Intelligence
Unit, London, UK, 1997.

FAOSTAT (website), Food and Agriculture Organization of the United Nations (F A0),
1999.

Feder, G., RE. Just, and D.Zilberman, “Adoption of Agricultural Innovations in
Developing Countries: A Survey,” Economic Development and Cultural Change,
33:2, 1985, pp. 255-98.

Foreign Agricultural Service (F AS), Agricultural Situation - Dominican Republic,
Agricultural Attache Annual Report, Santo Domingo, Dominican Republic, 1997.

Gittinger, J. Price, Economic Analysis of Agricultural Projects, The World Bank, Johns
Hopkins University Press, Washington, DC. 1982.

Griliches, Zvi,‘ “Hybrid Corn: An Exploration in the Economics of Technological Change,”
Econometrica, 25:4, October, 1957, pp. 501-522.

208

Guerrero, Ing. Ramon A, “PC-50: Nueva Variedad de Habichuela Roja,” (Release
Brochure), CRSP Proyecto Titulo XII/Secretaria de Estado de Agricultura,
Republics Dominicana, 1987.

Heikes, Gregory A, The Adoption of Agricultural Technology: The Case of Beans in the
Dominican Republic, (unpublished/unfinished M.Sc. thesis, MSU Ag Economics)
Michigan State University, E. Lansing, MI, 1995.

Hirschman, Albert 0., Exit, Voice, and Loyalty, Harvard, 1970.
Holtzrnan, J .S. “Rapid Reconnaissance Guidelines for Agricultural Marketing and Food
System Research in Developing Countries,” Michigan State University

International Development Papers, Working Paper No.30, Department of
Agricultural Economics, Michigan State University, East Lansing, MI, 1986.

InterAmerican Development Bank (IDB website), Economic Statistics Database, 1998.

JICA, Study on the Integrated Rural Development Project of the Yaque del Sur River
Basin in the Dominican Republic, June 1998.

Lancaster, Kelvin, Consumer Demand: A New Approach, New York: Columbia
University Press, 1971.

Libre, Frank, Vice-President, Citibank, Santo Domingo, Discussion in August 1998.

Lindner, lick, and ac. Jarrett, “Supply Shifts and the Size ofResearch Benefits,”
American Journal of Agricultural Economics, February 1978.

Masters, William, with Anne Williams and Marnadou Sidibe, “The Economic Impact of
Agricultural Research: A Practical Guide,” Department of Agricultural Economics,
Purdue University, West Lafayette, IN, February 1996.

Mather, David, preliminary results of rapid appraisal and a pilot consumer survey of bean
purchases and desired attributes, Santo Domingo and the San Juan Valley, July-
August 1999.

Mather, David, “Bean Supply Response in the DR”, term paper for EC 823: Applied
Econometrics, Dr. Wooldridge, Fall 1999.

Mellor, John, The Economics of Agricultural Development (Ithaca:Cornell University
Press, 1966).

209

Norton, G. and J. Davis, Evaluating Returns to Agricultural Research: A Review,
American Journal of Agricultural Economics 63 (4): 685-699.

Nui‘lez, Ruben, IICA Economist, informal conversation in Santo Domingo, July, 1999.

Nunez, Ruben, “Estudio Sobre la Comercializacion del Arroz en la Republica
Dominicana,” IICA/USAID/SEA, August, 1998.

Nuitez, Ruben, “Estudio Sobre el Mercado de Leguminosas en La Republica
Dominicana,” IICA/USAID/SEA, July, 1999.

Oehmke, James F., “The Calculation of Returns to Research in Distorted Markets,”
Agricultural Economics, vol. 2, 1988, pp. 291-302.

Rogers, Beatrice. 1987. “Resultados Preliminares del Estudio de los Efectos de Ingreso y
Precios Sobre el Consumo de Alimentos en la R.D.”

Schmid, AA, Property, Power and Public Choice. An Inquirty into Law and Economics,
Praeger Publishers, New York, NY, 1987.

Schmitz, A and G. Seckler, “Mechanical Agriculture and Social Welfare: The Case of the
Tomato Harvester,” American Journal of Agricultural Economics, 52:4,
November 1970, pp. 569-578.

Schuh, GE. and Helio Tollini, “Costs and Benefits of Agricultural Research: The State of
the Arts,” World Bank Stafl’ Working Paper No. 360, October 1979. IBRD.

Scobie, G. M. and TR. Posada, “The impact of technical change on income distribution:
The case of Rice in Colombia,” American Journal of Agricultural Economics 60:
85-92.

SEA Breve, 1990.

SEA, Secretaria de Estado .de Agricultura, Departamento de Economia Agropecuaria,
Diagnostico del Sector Agropecuario 1997, May 1998.

SEA, Secretaria do Estado de Agricultura, Departamento de Economia Agropecuaria,
Estadisticas del Sector Agropecuario 1998, May 1999.

SEA, Secretaria de Estado de Agricultura, Departamento de Economia Agropecuaria,

CoStos Estimados de Produccion de Cultivos Temporeros y Permanentes, 1997,
January 1998.

210

SEA, Departamento de Semillas, San Cristobal, Memorias Anuales, 1993-98, and an
interview with the director.

INESPRE, Instituto de Estabilizacion de Precios, Santo Domingo, Memorias Anuales,
1983-1998.

Scobie, Grant and Rafael Posada, “The Impact of Technical Change on Income
Distribution,” Agricultural Development in the Third World, ed. Carl Eicher and
John Staatz, Johns Hopkins University Press, pp. 400-415, 1990.

Shafi‘er, James D., .“On the Concept of Sub-Sector Studies,” American Journal of
Agricultural Economics 55: 333-335, 1973.

Sterling‘Diaz, Noe, Economist, Secretaria de Estado de Agricultura, Departamento de
Economia Agropecuaria, discussions in August, 1998.

UEPA, Unidad de Estudios de Precios Agropecuarios, Proyecto Analysis de Politics
Agropecuaria, “Analysis de Precios de la Habichuela Roja: Estacionalidad y
Tendencia,” Santo Domingo, 1991.

Valdes, Alberto and Barry Schaefi‘er, in collaboration with Jesus de los Santos,
“Surveillance of Agricultural Prices and Trade: A Handbook for the Dominican
Republic,” World Bank Technical Paper No. 267, August 1995, IBRD.

Williamson, Oliver, The Economic Institutions of Capitalism, Free Press, 1985.

World Bank, Country Assistance Strategy for the Dominican Republic, April 1995.

World Bank, World Development Indicators 199 7, 1997.

211