CAN A CREDIT-INSURANCE PACKAGE INCREASE THE ADOPTION OF A
MODERN VARIETY PACKAGE?: AN APPLICATION TO HONDURAN DRY BEAN
FARMERS
By
Wolfgang Baudino Pejuan Ucles

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Agricultural, Food and Resource Economics-Doctor of Philosophy
2015

ABSTRACT
CAN A CREDIT-INSURANCE PACKAGE INCREASE THE ADOPTION OF A MODERN
VARIETY PACKAGE?: AN APPLICATION TO HONDURAN DRY BEAN FARMERS
By
Wolfgang Baudino Pejuan Ucles
As farmers face an expected income-risk tradeoff, they can only obtain a higher expected
income by increasing their exposure to risk. Many farmers choose low-yielding technologies
(e.g., traditional varieties) with lower yield variability over high-yielding technologies (e.g.,
modern varieties) with higher yield variability. Farmers’ lack of liquidity at planting time
exacerbates the income-risk tradeoff by raising the cost of investing in modern varieties--hence,
making yield uncertainty even more daunting.
The objective of this study is to assess the potential of a credit-insurance package to
increase farmer adoption of modern bean varieties and complementary inputs (e.g., fertilizer) and
thereby increase bean farmers’ gross income in Honduras. The specific objectives are to: 1)
assess if the change from a traditional variety to a modern variety is profitable, and if the change
from a traditional variety package to a modern variety package is profitable, 2) assess farmers’
exposure to price risk, perceived yield risk, and consumption risk, 3) identify how farmers
currently cope with shocks, 4) evaluate the existing insurance policies for bean farms adopting
modern varieties to observe the extent of risk transfer and load in excess of the actuarial fair rate,
and propose alternative policies that improve risk transfer, and 5) assess the benefits of a creditinsurance package for bean farmers in Honduras.
The change from a traditional variety package to a modern variety package is profitable
as it is shown with partial budgets together with the marginal income from fertilizer compared to
its marginal cost. However, farmers face price risk at a level of approximately 20% coefficient

of variation (CV) for both beans and corn, and a yield risk at a range of approximately 32%-34%
CV for beans depending on the bean technology, and 32% CV for corn.
Currently farmers in Honduras are offered catastrophic crop insurance policies (i.e.,
coverages of 45%) with high loads to the premium (i.e., 3-5) that do not provide high levels of
risk transfer, although it provides the means to grow another crop in case of a peril striking.
Thus, farmers decide to mainly cope with shocks by working in another job besides their crop
enterprise (50%), selling their assets (24%), and asking for a loan from friends, family or other
person (14%).
Simulations, using a multiperiod stochastic model where households maximize expected
consumption utility, were used to evaluate farmers’ decisions on obtaining credit and purchasing
crop insurance. Farmers’ elicited probability distributions together with secondary data and
expert opinion were used for specifying the stochastic element in the simulation.
Farmers under general conditions would be benefited by a subsidized credit-insurance
package (i.e., load of 1.0) that enables them to change from a traditional variety package (i.e.,
low fertilizer and traditional seed) to a modern variety package (i.e., high fertilizer level and
modern seed). The specific conditions in which insurance would benefit are when farmers are
highly risk averse and do not need credit for crop production and living in villages with relatively
high prices, or, moderately risk averse farmers that need credit in villages with relatively high
prices. It is advisable to revisit the approach of estimating the yield probability distributions due
to the problems encountered.

Copyright by
WOLFGANG BAUDINO PEJUAN UCLES
2015

I would like to dedicate this dissertation to my mother, Ilma, who has always encouraged
me to pursue great things in life.

v

ACKNOWLEDGEMENTS

I would like to express my special gratitude to my major professor and member of my
dissertation committee evaluation, Dr. Richard Bernsten, for his advice throughout all my stay at
MSU. Thanks for revising my many drafts.
I would also like to express my gratitude to my dissertation supervisor, Dr. Roy Black,
for mentoring me throughout my Ph.D. degree. Thanks for devoting many, many hours guiding
me through the process.
I would also like to thank my committee members, Dr. Eric Crawford and Dr. Irvin
Widders, for the helpful contributions to my dissertation.
I would like to thank the BEAN/COWPEA Collaborative Research Support Program for
funding my studies at MSU and research in Honduras. Also, the Bean/Cowpea representatives in
Honduras that contributed to do my research: Juan Carlos Rosas.
I would like to thank MSU Graduate School for funding part of my research.
To Rommel Reconco, Byron Reyes and David Mather who helped me analyze sections of my
dissertation.
To Calixto Trujillo for helping me carry out my survey.
To my family for supporting me and hearing me out when I went through difficult times.

vi

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................................ xii
LIST OF FIGURES .......................................................................................................................xx
KEY TO SYMBOLS OR ABBREVIATIONS .......................................................................... xxii
1.CHAPTER I: Introduction ............................................................................................................1
1.1.Problem Statement .....................................................................................................................1
1.2.Objectives ..................................................................................................................................2
1.3.Contribution ...............................................................................................................................3
1.4.Dissertation Outline ...................................................................................................................3
2.CHAPTER II: Literature Review..................................................................................................6
2.1.Honduran socioeconomic characteristics ...................................................................................7
2.1.1.Poverty in Honduras and the use of crop insurance in developing countries .........................7
2.1.2.Characteristics of dry bean farmers in Honduras ....................................................................8
2.1.2.1.Dry beans and corn yields ....................................................................................................8
2.1.3.Insurance schemes in Honduras ............................................................................................11
2.2.Consumption Smoothing .........................................................................................................12
2.2.1.Income risk tradeoff ..............................................................................................................12
2.2.2.Modern bean varieties yield higher than traditional varieties ...............................................13
2.2.2.1.Higher income from modern varieties ...............................................................................14
2.2.2.2.Perceived higher risk from modern varieties .....................................................................16
2.3.Risk Management ....................................................................................................................17
2.3.1.Risk management strategies ..................................................................................................17
2.3.2.Risk management strategies used by Honduran farmers ......................................................18
2.3.3.Risk attitudes and portfolio ...................................................................................................19
2.3.4.Diversifiable and covariate risk, and systemic and idiosyncratic risk ..................................19
2.4.Agricultural Insurance .............................................................................................................20
2.4.1.Types of insurance ................................................................................................................20
2.4.2.Adverse selection, moral hazard, and basis risk ...................................................................20
2.4.2.1.Group insurance, moral hazard, and adverse selection ......................................................21
2.4.3.Weather index-based insurance ............................................................................................21
2.4.4.Insurance and probability distributions .................................................................................22
2.5.Credit........................................................................................................................................22
2.5.1.Microcredit history in Honduras ...........................................................................................22
2.5.1.1.Rural cashiers .....................................................................................................................23
2.5.1.2.Grammeen Bank ................................................................................................................23
2.5.2.Use of credit and insurance ...................................................................................................23
2.6.Evaluating Risk ........................................................................................................................24
2.6.1.Income consumption model ..................................................................................................25
2.6.2.Income-consumption model studies .....................................................................................26

vii

2.6.3.Utility of wealth models using simulation ............................................................................27
2.6.4.Stochastic dominance (SD) ...................................................................................................29
2.6.4.1.Cumulative Density Functions and intuition to stochastic dominance ..............................30
2.6.5.Improved bean package ........................................................................................................33
2.6.6.Insurance and credit ..............................................................................................................34
2.6.6.1.Yield and revenue contracts ...............................................................................................36
2.6.7.Subjective yield distributions ................................................................................................38
2.7.Factors Affecting Adoption .....................................................................................................39
2.8.Simulation ................................................................................................................................39
2.8.1.Reasons for using simulation ................................................................................................39
2.8.2.Objective ...............................................................................................................................40
2.9.Chapter Summary ....................................................................................................................40
3.CHAPTER III: Conceptual Model and Methods ........................................................................42
3.1.Why Simulation? .....................................................................................................................42
3.2.Simulation Model.....................................................................................................................43
3.3.Specifics of the SPVWUM ......................................................................................................43
3.4.@Risk General Setup ...............................................................................................................50
3.4.1.Utility function ......................................................................................................................52
3.4.2.Quasi-fixed choices...............................................................................................................52
3.4.3.Monte Carlo simulation ........................................................................................................53
3.4.4.Random variables, correlation matrix and parameters .........................................................53
3.4.4.1.Elicitation of subjective probabilities ................................................................................54
3.4.4.1.1.Visual Impact Method.....................................................................................................54
3.4.4.2.Insurance and credit ...........................................................................................................56
3.4.4.3.Insurance contract parameters............................................................................................56
3.4.4.3.1.Premium evaluation ........................................................................................................56
3.5.Data Collection ........................................................................................................................57
3.5.1.Key informant interviews .....................................................................................................57
3.5.1.1.Insurance companies ..........................................................................................................57
3.5.1.2.Banks and other financial institutions ................................................................................58
3.5.1.3.Scientists ............................................................................................................................58
3.5.2.Farmer survey .......................................................................................................................58
3.5.2.1.Farmer sample selection ....................................................................................................58
3.5.2.2.Farmer data collection........................................................................................................59
3.5.2.3.Farmer data ........................................................................................................................61
3.5.2.3.1.Income.............................................................................................................................61
3.5.2.3.2.Capital .............................................................................................................................61
3.5.2.3.3.Costs................................................................................................................................61
3.5.3.Secondary data ......................................................................................................................63
3.6.Factors to Compare ..................................................................................................................63
3.7.Certainty Equivalents Comparison of Technology Choices ....................................................64
3.7.1.Certainty equivalent comparison for risk averse and not capital constrained ......................65
3.7.2.Certainty equivalent comparison for risk averse and capital constrained.............................66
3.7.3.Gains from additional inputs.................................................................................................67
3.8.Chapter Summary ....................................................................................................................69

viii

4.CHAPTER IV: Data Analysis and Information Collection ........................................................71
4.1.Key Informant Interviews ........................................................................................................71
4.1.1.Crop insurance history in Honduras .....................................................................................71
4.1.2.Crop insurance rules and regulations ....................................................................................72
4.1.3.Insurance companies .............................................................................................................73
4.1.3.1.ProAgro ..............................................................................................................................73
4.1.3.2.HSBC .................................................................................................................................75
4.1.3.3.Seguros Interamericana ......................................................................................................76
4.1.3.4.Seguros Equidad ................................................................................................................76
4.1.3.5.Seguros Atlantida ...............................................................................................................77
4.1.4.Credit and microcredit ..........................................................................................................78
4.1.4.1.Credit..................................................................................................................................78
4.1.4.1.1.BANADESA ...................................................................................................................79
4.1.4.1.2.BAMER ..........................................................................................................................80
4.1.4.1.3.Banco Ficohsa .................................................................................................................81
4.1.4.1.4.Banco HSBC ...................................................................................................................81
4.1.4.1.5.Banco de Occidente ........................................................................................................82
4.1.4.1.6.Banco Atlantida ..............................................................................................................83
4.1.4.2.Microcredit.........................................................................................................................84
4.1.4.2.1.ODEF ..............................................................................................................................84
4.2.Farmer Survey..........................................................................................................................85
4.2.1.Farmers’ elicited yield distributions analysis .......................................................................86
4.2.1.1.Enumerator bias .................................................................................................................86
4.2.1.2.Incomplete distributions.....................................................................................................88
4.2.1.3.Villages and farmer categories...........................................................................................91
4.2.1.4.Completing the distributions ..............................................................................................93
4.2.1.5.Fitted distribution and distribution parameters ..................................................................93
4.2.1.6.Rescaling ............................................................................................................................94
4.2.2.Farmer crop revenues and costs ............................................................................................95
4.2.2.1.Corn revenues and costs.....................................................................................................96
4.2.2.1.1.Corn yields ......................................................................................................................97
4.2.2.1.2.Corn prices ......................................................................................................................99
4.2.2.1.3.Corn production costs ...................................................................................................100
4.2.2.2.Bean revenues and costs ..................................................................................................101
4.2.2.2.1.Bean yields ....................................................................................................................101
4.2.2.2.2.Bean prices ....................................................................................................................105
4.2.2.2.3.Bean costs .....................................................................................................................106
4.2.2.2.4.Input use and costs ........................................................................................................106
4.2.2.2.5.Varieties and seeding rate .............................................................................................107
4.2.2.2.6.Fertilizer use..................................................................................................................107
4.2.2.2.7.Herbicide use ................................................................................................................108
4.2.2.2.8.Fungicide use ................................................................................................................108
4.2.2.2.9.Insecticide use ...............................................................................................................109
4.2.2.2.10.Input expenditures.......................................................................................................110
4.2.2.2.11.Partial budgets for variety change in each village ......................................................111
4.2.2.2.12.Activity expenditures ..................................................................................................111

ix

4.2.2.2.13.Area planted to beans and corn ...................................................................................113
4.2.3.Gross income from other sources .......................................................................................113
4.2.3.1.Remittances ......................................................................................................................114
4.2.4.Consumption .......................................................................................................................114
4.2.5.Farmer assets .......................................................................................................................114
4.2.6.Farmer use of credit ............................................................................................................115
4.2.7.Farmer use of crop insurance ..............................................................................................115
4.2.8.Coverages in Honduras .......................................................................................................117
4.2.9.Farmer coping with disaster ................................................................................................118
4.3.Parameters used for simulation ..............................................................................................118
4.4.Correlation Matrix .................................................................................................................125
4.5.Analysis of Crop Insurance in a Portfolio Setting .................................................................126
4.6.Base Village Simulation and Cumulative Probability Distribution Functions ......................127
4.6.1.La Cienega farmers distribution results for with and without insurance ............................127
4.6.1.1.Revenue distributions.......................................................................................................128
4.6.1.2.Gross income distributions ..............................................................................................134
4.6.1.3.Portfolio distributions ......................................................................................................137
4.6.1.4.Chance of not meeting consumption expenditures ..........................................................140
4.7.Chapter Summary ..................................................................................................................141
5.CHAPTER V: Results and Analysis .........................................................................................143
5.1.Certainty Equivalent Ratios ...................................................................................................143
5.2.Effects of Variety Packages, Credit Insurance Packages, and Farmer Characteristics. ........145
5.2.1.Modern variety package change .........................................................................................145
5.2.2.Use of insurance..................................................................................................................151
5.2.2.1.Insurance premiums .........................................................................................................153
5.2.2.2.Coverages .........................................................................................................................154
5.2.3.Input level ...........................................................................................................................154
5.2.4.Type of variety ....................................................................................................................159
5.2.5.Farmers’ risk aversion and rate of time preferences ...........................................................163
5.2.6.Use of credit and insurance .................................................................................................163
5.2.7.Price of beans ......................................................................................................................165
5.3.Percentage Chance of Going Insolvent ..................................................................................165
5.4.Results for Cases of Interest ..................................................................................................173
5.4.1.Results for a risk averse and not capital constrained farmer ..............................................173
5.4.2.Results for a risk averse and capital constrained farmer.....................................................174
5.5.Chapter Summary ..................................................................................................................175
6.CHAPTER VI: Conclusions and Recommendations ................................................................177
6.1.Conclusions ............................................................................................................................177
6.2.Limitations and Recommendations for Future Research .......................................................181
APPENDICES .............................................................................................................................184
Appendix 1. Determining yield probability distributions ............................................................185
Appendix 2. Completing the distributions due to problems encountered ....................................194
Appendix 3. Village Information .................................................................................................198

x

BIBLIOGRAPHY ........................................................................................................................243

xi

LIST OF TABLES

Table 1. Average area and average of farmers planting corn and bean in Postrera (2001) and
Primera (2002), by department, Honduras. ...........................................................................9
Table 2. Mean yield (kg/ha) of traditional and modern bean varieties, 1999-2000, Honduras. ....10
Table 3. Partial budget analysis: change in income for substituting one hectare of a modern bean
variety for a traditional bean variety. ..................................................................................15
Table 4. Choices (i.e., c=1, …,11) resulting from a combination of technology categories and use
of credit and/or insurance that will be compared on the metrics certainty equivalent and
insolvency percentage, within each village. ........................................................................46
Table 5. Acronyms of choices (i.e., c=1, …,11) resulting from a combination of technology
categories and use of credit and/or insurance that will be compared on the metrics
certainty equivalent and insolvency percentage, within each village. ................................47
Table 6. Proper certainty equivalent comparisons of choices for each of the cases of interest
where a credit-insurance package could potentially be beneficial. .....................................48
Table 7. Example of a bean yield distribution, using the visual impact method. ..........................55
Table 8. Example of a corn yield distribution, using the visual impact method............................55
Table 9. Mean bean yields and coefficient of variation for Postrera season 1999 and 2000,
Honduras. ............................................................................................................................60
Table 10. Acronyms of choices related to technology packages to compare for a risk averse and
not capital constrained farmer. ............................................................................................65
Table 11. Acronyms of choices related to technology packages to compare for a risk averse and
capital constrained farmer. ..................................................................................................66
Table 12. Acronyms of choices related to level of fertilizer to compare for a risk averse and not
capital constrained farmer. ..................................................................................................69
Table 13. Copay (% of loss) schedule for type of peril insured. ...................................................74
Table 14. Number and percentage of uniform probability distributions that were elicited by
enumerators. ........................................................................................................................88
Table 15. Trimmed mean corn yields (kg/ha) in Primera by type of land preparation and type of
variety. .................................................................................................................................97
Table 16. Corn yield differences (kg/ha) between La Cienega and every other pooled village. ...98
xii

Table 17. Corn prices in Primera by pooled villages. ....................................................................99
Table 18. Trimmed mean corn costs (L./ha) in Primera by type of land preparation and type of
variety. ...............................................................................................................................100
Table 19. Trimmed mean bean yields (kg/ha) by type of land preparation, type of variety and
technology applied. ...........................................................................................................101
Table 20. Bean yield differences (kg/ha) between La Cienega and other villages. .....................103
Table 21. Average national bean yields (kg/ha) for period 1990-2007. ......................................103
Table 22. Expected minimum bean yields by category. ..............................................................104
Table 23. Bean prices (L. / kg) in Postrera by pooled villages and type of variety. ....................105
Table 24. Bean varieties* used by farmers, Honduras 2007. .......................................................107
Table 25. Fertilizer used by farmers, Honduras 2007. .................................................................108
Table 26. Herbicides used by farmers, Honduras 2007. ..............................................................109
Table 27. Fungicides used by farmers, Honduras 2007. ..............................................................109
Table 28. Insecticides used by farmers, Honduras 2007. ............................................................109
Table 29. Trimmed mean input expenditures (L./ha) by type of land preparation, type of variety
and technology applied. ....................................................................................................110
Table 30. Trimmed mean residual activity expenditures (L./ha) by type of land preparation, type
of variety and technology applied. ....................................................................................112
Table 31. Trimmed mean value (L.) of farmers’ capital assets by type of land preparation, type of
variety and technology applied. ........................................................................................115
Table 32. Farmer awareness and use of crop insurance...............................................................116
Table 33. Percent of farmers quoting different reasons for buying bean insurance. ...................116
Table 34. Percent of farmers quoting different reasons for not buying bean insurance. .............116
Table 35. Farmers experiencing shortage of income different percent of time. ..........................118
Table 36. Parameters* used in @Risk for the stochastic present value of wealth utility model for
Alauca farmers, Honduras. ................................................................................................120
Table 37. Parameters* used in @Risk for the stochastic present value of wealth utility model for
Arauli and Chirinos farmers, Honduras. ...........................................................................121

xiii

Table 38. Parameters* used in @Risk for the stochastic present value of wealth utility model for
Talanga and Guaimaca farmers, Honduras. ......................................................................122
Table 39. Parameters* used in @Risk for the stochastic present value of wealth utility model for
La Cienega farmers, Honduras. .........................................................................................123
Table 40. Parameters* used in @Risk for the stochastic present value of wealth utility model for
Villa de San Francisco and Sabaneta de Valle de Angeles farmers, Honduras. ...............124
Table 41. Correlation matrix of simulation random variables. ....................................................125
Table 42. CE ratios of modern variety package over traditional variety package for each land
preparation system and credit-insurance package in La Cienega. ....................................146
Table 43. CE ratios of modern variety package over traditional variety package for each land
preparation system and credit-insurance package in Talanga and Guaimaca. ..................147
Table 44. CE ratios of modern variety package over traditional variety package for each land
preparation system and credit-insurance package in Alauca. ...........................................148
Table 45. CE ratios of modern variety package over traditional variety package for each land
preparation system and credit-insurance package in Arauli and Chirinos. .......................149
Table 46. CE ratios of modern variety package over traditional variety package for each land
preparation system and credit-insurance package in Villa de San Francisco and Sabaneta
de Valle de Angeles. .........................................................................................................150
Table 47. CE ratios of high input use over low input use for each land preparation system and
type of farmer (i.e., risk aversion and discount rate) in La Cienega, and Talanga and
Guaimaca...........................................................................................................................156
Table 48. CE ratios of high input use over low input use for each land preparation system and
type of farmer (i.e., risk aversion and discount rate) in Alauca and Arauli and Chirinos.157
Table 49. CE ratios of high input use over low input use for each type of farmer (i.e., risk
aversion and discount rate) in Villa de San Francisco and Sabaneta de Valle de Angeles.158
Table 50. CE ratios of modern variety use over traditional variety use for each type of farmer
(i.e., risk aversion and discount rate) in La Cienega, and Talanga and Guaimaca. ..........160
Table 51. CE ratios of modern variety use over traditional variety use for each type of farmer
(i.e., risk aversion and discount rate) in Villa de San Francisco and Sabaneta de Valle de
Angeles, and Alauca..........................................................................................................161
Table 52. CE ratios of modern variety use over traditional variety use for each type of farmer
(i.e., risk aversion and discount rate) in Arauli and Chirinos. ..........................................162

xiv

Table 53. Percentage chance of going insolvent for each land preparation system and creditinsurance package in La Cienega. .....................................................................................168
Table 54. Percentage chance of going insolvent for each land preparation system and creditinsurance package in Talanga and Guaimaca. ..................................................................169
Table 55. Percentage chance of going insolvent for each land preparation system and creditinsurance package in Villa de San Francisco y Sabaneta de Valle de Angeles. ...............170
Table 56. Percentage chance of going insolvent for each land preparation system and creditinsurance package in Alauca. ............................................................................................171
Table 57. Percentage chance of going insolvent for each land preparation system and creditinsurance package in Arauli and Chirinos. .......................................................................172
Table 58. Probability Distribution Analysis of Variance for Traditional Varieties, Honduras 2007.
...........................................................................................................................................185
Table 59. Probability Distribution Analysis of Variance for Modern Varieties, Honduras 2007.186
Table 60. Probability Distribution Analysis of Variance for Traditional Varieties, Honduras 2007.
...........................................................................................................................................187
Table 61. Probability Distribution Analysis of Variance for Modern Varieties, Honduras 2007.188
Table 62. Probability Distribution Analysis of Variance for Traditional Varieties (enumerator
five removed), Honduras 2007. .........................................................................................189
Table 63. Probability Distribution Analysis of Variance for Modern Varieties (enumerator five
removed), Honduras 2007. ................................................................................................190
Table 64. T-tests for differences between reported mean and different points of the elicited
distribution. .......................................................................................................................191
Table 65. Traditional Variety Probability Distribution by Village and Interval (Half Distribution).
...........................................................................................................................................192
Table 66. Traditional Variety Probability Distribution by Village and Interval (More than Half
Distribution). .....................................................................................................................193
Table 67. Bean and corn yield (kg/ha) response to location, land preparation system, type of
variety, and input levels. ...................................................................................................198
Table 68. Regression of expert opinion on bean and corn yield (kg/ha) responses to rainfall,
elevation, and fertilizer......................................................................................................199
Table 69. Partial budget analysis: change in income for substituting one hectare of a modern
bean variety for a traditional bean variety in La Cienega. ................................................199

xv

Table 70. Partial budget analysis: change in income for substituting one hectare of a modern
bean variety for a traditional bean variety in Talanga and Guaimaca...............................200
Table 71. Partial budget analysis: change in income for substituting one hectare of a modern
bean variety for a traditional bean variety in Arauli and Chirinos. ...................................200
Table 72. Partial budget analysis: change in income for substituting one hectare of a modern
bean variety for a traditional bean variety in Villa de San Francisco and Sabaneta de Valle
de Angeles. ........................................................................................................................201
Table 73. Partial budget analysis: change in income for substituting one hectare of a modern
bean variety for a traditional bean variety in Alauca. .......................................................201
Table 74. Percentage of farmers quoting their main way of coping with perils. .........................202
Table 75. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in La Cienega; with a RTP*** of 7%, insurance coverage of
75%, load=1.0, and two CRRAC****. .............................................................................203
Table 76. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in La Cienega; with a RTP*** of 7%, insurance coverage of
75%, load=1.4, and two CRRAC****. .............................................................................204
Table 77. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in La Cienega; with a RTP*** of 7%, insurance coverage of
45%, load=1.0, and two CRRAC****. .............................................................................205
Table 78. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in La Cienega; with a RTP*** of 7%, insurance coverage of
45%, load=1.4, and two CRRAC****. .............................................................................206
Table 79. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in La Cienega; with a RTP*** of 14%, insurance coverage of
75%, load=1.0, and two CRRAC****. .............................................................................207
Table 80. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in La Cienega; with a RTP*** of 14%, insurance coverage of
75%, load=1.4, and two CRRAC****. .............................................................................208
Table 81. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in La Cienega; with a RTP*** of 14%, insurance coverage of
45%, load=1.0, and two CRRAC****. .............................................................................209
Table 82. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in La Cienega; with a RTP*** of 14%, insurance coverage of
45%, load=1.4, and two CRRAC****. .............................................................................210

xvi

Table 83. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Talanga and Guaimaca; with a RTP*** of 7%, insurance
coverage of 75%, load=1.0, and two CRRAC****. .........................................................211
Table 84. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Talanga and Guaimaca; with a RTP*** of 7%, insurance
coverage of 75%, load=1.4, and two CRRAC****. .........................................................212
Table 85. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Talanga and Guaimaca; with a RTP*** of 7%, insurance
coverage of 45%, load=1.0, and two CRRAC****. .........................................................213
Table 86. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Talanga and Guaimaca; with a RTP*** of 7%, insurance
coverage of 45%, load=1.4, and two CRRAC****. .........................................................214
Table 87. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Talanga and Guaimaca; with a RTP*** of 14%, insurance
coverage of 75%, load=1.0, and two CRRAC****. .........................................................215
Table 88. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Talanga and Guaimaca; with a RTP*** of 14%, insurance
coverage of 75%, load=1.4, and two CRRAC****. .........................................................216
Table 89. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Talanga and Guaimaca; with a RTP*** of 14%, insurance
coverage of 45%, load=1.0, and two CRRAC****. .........................................................217
Table 90. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Talanga and Guaimaca; with a RTP*** of 14%, insurance
coverage of 45%, load=1.4, and two CRRAC****. .........................................................218
Table 91. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Villa de San Francisco and Sabaneta de Valle de Angeles;
with a RTP*** of 7%, insurance coverage of 75%, load=1.0, and two CRRAC****. ....219
Table 92. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Villa de San Francisco and Sabaneta de Valle de Angeles;
with a RTP*** of 7%, insurance coverage of 75%, load=1.4, and two CRRAC****. ....220
Table 93. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Villa de San Francisco and Sabaneta de Valle de Angeles;
with a RTP*** of 7%, insurance coverage of 45%, load=1.0, and two CRRAC****. ....221
Table 94. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Villa de San Francisco and Sabaneta de Valle de Angeles;
with a RTP*** of 7%, insurance coverage of 45%, load=1.4, and two CRRAC****. ....222

xvii

Table 95. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Villa de San Francisco and Sabaneta de Valle de Angeles;
with a RTP*** of 14%, insurance coverage of 75%, load=1.0, and two CRRAC****. ..223
Table 96. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Villa de San Francisco and Sabaneta de Valle de Angeles;
with a RTP*** of 14%, insurance coverage of 75%, load=1.4, and two CRRAC****. ..224
Table 97. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Villa de San Francisco and Sabaneta de Valle de Angeles;
with a RTP*** of 14%, insurance coverage of 45%, load=1.0, and two CRRAC****. ..225
Table 98. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Villa de San Francisco and Sabaneta de Valle de Angeles;
with a RTP*** of 14%, insurance coverage of 45%, load=1.4, and two CRRAC****. ..226
Table 99. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Alauca; with a RTP*** of 7%, insurance coverage of 75%,
load=1.0, and two CRRAC****. ......................................................................................227
Table 100. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Alauca; with a RTP*** of 7%, insurance coverage of 75%,
load=1.4, and two CRRAC****. ......................................................................................228
Table 101. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Alauca; with a RTP*** of 7%, insurance coverage of 45%,
load=1.0, and two CRRAC****. ......................................................................................229
Table 102. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Alauca; with a RTP*** of 7%, insurance coverage of 45%,
load=1.4, and two CRRAC****. ......................................................................................230
Table 103. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Alauca; with a RTP*** of 14%, insurance coverage of 75%,
load=1.0, and two CRRAC****. ......................................................................................231
Table 104. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Alauca; with a RTP*** of 14%, insurance coverage of 75%,
load=1.4, and two CRRAC****. ......................................................................................232
Table 105. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Alauca; with a RTP*** of 14%, insurance coverage of 45%,
load=1.0, and two CRRAC****. ......................................................................................233
Table 106. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Alauca; with a RTP*** of 14%, insurance coverage of 45%,
load=1.4, and two CRRAC****. ......................................................................................234

xviii

Table 107. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Arauli and Chirinos; with a RTP*** of 7%, insurance
coverage of 75%, load=1.0, and two CRRAC****. .........................................................235
Table 108. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Arauli and Chirinos; with a RTP*** of 7%, insurance
coverage of 75%, load=1.4, and two CRRAC****. .........................................................236
Table 109. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Arauli and Chirinos; with a RTP*** of 7%, insurance
coverage of 45%, load=1.0, and two CRRAC****. .........................................................237
Table 110. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Arauli and Chirinos; with a RTP*** of 7%, insurance
coverage of 45%, load=1.4, and two CRRAC****. .........................................................238
Table 111. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Arauli and Chirinos; with a RTP*** of 14%, insurance
coverage of 75%, load=1.0, and two CRRAC****. .........................................................239
Table 112. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Arauli and Chirinos; with a RTP*** of 14%, insurance
coverage of 75%, load=1.4, and two CRRAC****. .........................................................240
Table 113. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Arauli and Chirinos; with a RTP*** of 14%, insurance
coverage of 45%, load=1.0, and two CRRAC****. .........................................................241
Table 114. CE* values (L.)* and IP** for different technology categories, and credit and crop
insurance use for farmers in Arauli and Chirinos; with a RTP*** of 14%, insurance
coverage of 45%, load=1.4, and two CRRAC****. .........................................................242

xix

LIST OF FIGURES

Figure 1. Cumulative density function of yield (kg/ha) from a modern dry bean variety. For
interpretation of the references to color in this and all other figures, the reader is referred
to the electronic version of this dissertation. .......................................................................32
Figure 2. Cumulative density functions of yield (kg/ha) from a modern and a traditional dry bean
variety. .................................................................................................................................32
Figure 3. Cumulative density functions of gross income (L.) from a modern* and a traditional**
dry bean variety without insurance and a modern variety with insurance***. ...................33
Figure 4. Half distribution (pdf) classification found in 2007 farmer survey, Honduras 2007. ....89
Figure 5. More than half distribution (pdf) classification found in 2007 farmer survey, Honduras
2007. ....................................................................................................................................89
Figure 6. Bean revenue (L./ha) cdfs for revenue insurance and no insurance; for Nonmechanized+traditional variety+low input use category, no copay, load=1.00, 75%
coverage, high indemnity price. ........................................................................................130
Figure 7. Bean revenue (L./ha) cdfs for revenue insurance with and without copay; for Nonmechanized+traditional variety+low input use category, load=1.00, 75% coverage, high
indemnity price..................................................................................................................130
Figure 8. Bean revenue (L./ha) cdfs for revenue insurance and yield insurance; for Nonmechanized+traditional variety+low input use category, no copay, load=1.00, 75%
coverage, high indemnity price. ........................................................................................131
Figure 9. Bean revenue (L./ha) cdfs for yield insurance and no insurance; for Nonmechanized+traditional variety+low input use category, no copay, load=1.00, 75%
coverage, high indemnity price. ........................................................................................132
Figure 10. Bean revenue (L./ha) cdfs for yield insurance and no insurance; for Nonmechanized+traditional variety+low input use category, with copay, load=1.00, 75%
coverage, high indemnity price. ........................................................................................133
Figure 11. Bean revenue (L./ha) cdfs for yield insurance with and without copay; for Nonmechanized+traditional variety+low input use category, load=1.00, 75% coverage, high
indemnity price..................................................................................................................133
Figure 12. Bean revenue (L./ha) cdfs for yield insurance with high and low indemnity price; for
Non-mechanized+traditional variety+low input use category, with copay, load=1.00, 75%
coverage. ...........................................................................................................................134

xx

Figure 13. Bean revenue (L./ha) and gross income (L./ha) cdfs with no insurance; for Nonmechanized+traditional variety+low input use category. .................................................135
Figure 14. Bean gross income (L./ha) cdfs for yield insurance and no insurance; for Nonmechanized+traditional variety+low input use category, with copay, load=1.00, 75%
coverage, high indemnity price. ........................................................................................136
Figure 15. Bean gross income (L./ha) cdfs for yield insurance and no insurance; for Nonmechanized+traditional variety+low input use category, with copay, load=1.00, 75%
coverage, low indemnity price. .........................................................................................136
Figure 16. Gross income cdfs (L./ha) for beans and total production (beans and corn); for Nonmechanized+traditional variety+low input use category and no insurance. .....................137
Figure 17. Total production gross income (beans and corn) (L./ha) cdfs for bean yield insurance
and no insurance; for Non-mechanized+traditional variety+low input use category, with
copay, load=1.00, 75% coverage, low indemnity price. ...................................................138
Figure 18. End of first year wealth (L.) cdfs for yield insurance and no insurance; for Nonmechanized+traditional variety+low input use category, with copay, load=1.00, 75%
coverage, low indemnity price. .........................................................................................139
Figure 19. End of twentieth year wealth (L.) cdfs for yield insurance and no insurance; for Nonmechanized+traditional variety+low input use category, with copay, load=1.00, 75%
coverage, low indemnity price. .........................................................................................140

xxi

KEY TO SYMBOLS OR ABBREVIATIONS

BAC

Central American Bank

BANADESA

Honduran National Agrarian Development Bank

CEPAL

Economic Commission for Latin America and the Caribbean

CDF

Cumulative Density Function

CE

Certainty Equivalent

CER

Certainty Equivalent Ratio

CV

Coefficient of variation

FAO

Food and Agriculture Organization of the United Nations

FDSD

First Degree Stochastic Dominance

HSBC

Hong Kong Shanghai Bank Corporation in Honduras

ICCYA

Investment Cost Coverage with Yield Adjustment

INE

Honduran National Statistics Institute

IS

Insurance Superintendence

ODEF

Feminine Entrepreneurial Development Organization

PCGI

Percentage Chance of Going Insolvent

ProAgro

Agrarian Protection Insurance Company in Honduras

SBI

Superintendence of banks and Insurance

SDSD

Second Degree Stochastic Dominance

SIMPAH

Honduran Agricultural Product Information System

SPVWUM

Stochastic Present Value of Wealth Utility Model

YIPCA

Yield Insurance with Production Cost Adjustment

xxii

1. CHAPTER I: Introduction
1.1. Problem Statement
Small-scale farmers in Honduras, with a significant share of land dedicated to dry bean
production, are thought to be constrained from adopting a modern variety package by cash-flow
and risk constraints (Smeltekop, 2005). Smeltekop reported that farmers stated that they do not
want to risk a low yield with modern varieties in a given year due to terminal drought, the factor
farmers expressed to be more constraining to increase yields. Thus, farmers face an income-risk
tradeoff-- they can only obtain a higher expected income by increasing their exposure to risk. As
a result, they choose traditional varieties with lower yield variability over modern varieties with
higher yield variability. This income-risk tradeoff is exacerbated by farmers’ lack of liquidity at
planting time, which raises the cost of investing in modern varieties and hence makes yield
uncertainty even more daunting. In contrast, experts in Honduras suggest this perception is due
to farmers’ experience planting older improved varieties compared to the newly released
varieties (Rosas, 2009; personal communication). Still, the potential liquidity constraint remains.
Consequently, farmers potentially are in the ‘poverty trap’.
Crop yield insurance potentially increases farmers’ choices by transferring at least some
risk to the insurer (Ray, 1981). Access to credit allows capital-constrained farmers to acquire
cash to purchase modern varieties and complementary inputs prior to planting. Crop insurance
reduces risk and increases the farmers’ willingness to purchase modern varieties and
complementary inputs for risk averse farmers. Thus, a successful credit-insurance package is
hypothesized to increase farmers’ incomes through the purchase of modern varieties and
complementary inputs (i.e., modern variety package) in three out of four cases from a full
factorial combination of the two factors, (i.e., risk aversion and capital-constraint) with two

1

levels within each factor (i.e., capital constrained or not capital constrained; risk averse or not
risk averse). The three relevant cases from the full factorial combination of these two factors are
the following: 1) farmers not constrained by capital but who are driven by risk aversion to
purchase traditional varieties instead of modern varieties because of failure of the latter in
drought situations, 2) farmers constrained by capital and driven by risk aversion to not access
credit to purchase a modern variety and complementary inputs in case of a bad event happening
and not being able to repay the loan, and 3) farmers who are constrained by capital and not risk
averse, would be required by banks to purchase crop insurance to be able to acquire the loan to
1

buy the modern variety package . This study will look at the two cases of risk aversion (i.e., first
two cases) within those three relevant cases.
In these scenarios, credit demand is directly assumed for the capital constrained farmers
as opposed to derived from financial statements as is practiced in other studies. Whether or not
credit-insurance packages are successful depends on socioeconomic conditions.
This study is motivated by the need to increase farmers’ expected gross income, the low
participation of crop insurance in Honduras, and the potential ability of crop insurance to
increase bean farmers’ gross income by facilitating the adoption of modern varieties.
1.2. Objectives
The objective of this study is to assess the potential of a credit-insurance package to
increase farmer adoption of modern bean varieties and complementary inputs (e.g., fertilizer) in

1

The fourth scenario consists of a farmer not constrained by capital and not risk averse, which is
not relevant since she would obtain the modern variety package if it outperforms the traditional
variety as it is portrayed in this study.
2

2

order to increase bean farmers’ gross income in Honduras. The specific objectives are to: 1)
assess if the change from a traditional variety to a modern variety is profitable, and if the change
from a traditional variety package to a modern variety package is profitable, 2) assess farmers’
exposure to price risk, perceived yield risk, and consumption risk, 3) identify how farmers
currently cope with shocks, 4) evaluate the existing insurance policies for bean farms adopting
modern varieties to observe the extent of risk transfer and load in excess of the actuarial fair rate,
and propose alternative policies that improve risk transfer, and 5) assess the benefits of a creditinsurance package for bean farmers in Honduras.
1.3. Contribution
The research will contribute to our understanding of the limitations farmers face in
adopting a modern bean variety package and provide insights regarding farmers’ incentives of
adopting or not adopting improved bean technologies. If the hypothesis that modern varieties are
perceived to increase risk is correct, this study will contribute to identifying institutional
adjustments (e.g., changes in insurance coverages) necessary to increase adoption. This research
will also contribute to the empirical work on crop insurance in developing countries. More
specifically, it will inform insurance companies about farmers’ reactions to incentives or
disincentives to purchasing crop insurance.
1.4. Dissertation Outline
Chapter 2 presents a review of the theory, concepts, and empirical work used as the
foundation of the study. The study, which develops a multiperiod stochastic simulation model to
capture the impact of stochastic yields and prices, uses the expected utility framework,

2

Gross income is defined in this study as revenue minus variable costs, except for family labor.
3

insolvency metrics, risk management, and portfolio theory concepts. This chapter presents
concepts of risk management, crop insurance, and credit to provide insight to farmer modern
bean variety adoption.
Chapter 3 develops the framework of the multiperiod stochastic simulation model used,
from its expected utility model to the specific structure and parameters of the simulation model
used. It includes the variables, parameters and correlation matrix that need to be collected and/or
established to estimate the model metrics (e.g., certainty equivalent, expected utility). It presents
the sources of information (i.e., key informant interviews, farmer survey, and secondary
information) and procedures followed for data collection (e.g., sampling procedure for the
survey). It also presents technology comparisons, which are used to answer the question of the
suitability of the use of crop insurance and credit to enhance modern bean variety adoption.
Finally, it discusses the elicitation of farmer subjective yield probability distribution functions
for use as a source of randomness in the simulation model, as well as to obtain the farmers’ view
of their perceived risk.
Chapter 4 presents estimates of the basic statistics and distribution of the variables,
establishes the parameters, and correlation values for the correlation matrix to be used in the
simulation model. It describes the general policies and parameters of credit and crop insurance
contracts used by banks and insurance companies. The chapter describes problems that emerged
from the farmer yield distribution elicitation and the steps followed to overcome those problems
and obtain subjective crop yield probability density distributions. It also indicates parameters
that were estimated from data obtained from the farmer survey and used in the simulation model
regarding bean and corn costs, farmer assets, farmers’ off-farm income, etc. Finally, it presents

4

differences between types of crop insurance contracts and differences between policies within
the same type of insurance contract.
Chapter 5 presents the results and analysis of the application of the simulation model to
provide insight into the objective of assessing the benefit of a credit-insurance package.
Comparisons between technology categories are presented and discussed.
Chapter 6 presents the study’s conclusion, policy recommendations, limitations, and
recommendations for future research. It discusses possible problem mitigation techniques for
future researchers when eliciting farmers’ yield distributions.

5

2. CHAPTER II: Literature Review
This chapter reviews the literature on Honduran socioeconomic characteristics,
consumption smoothing, risk management, crop insurance, credit, and simulation as a tool for
generating and understanding of decision making.
The Honduran socioeconomic characteristics section provides a sense of the possible
demand and impact of introducing a credit-insurance package by documenting the levels of
poverty. It also provides some general background information of farmers’ research on the yield
differentials between traditional and modern varieties. Additionally, this section discusses the
establishment of crop insurance in Honduras and the firms offering insurance.
The consumption smoothing section discusses farmers’ balance of their spending and
saving, but where risk and credit constraints are present, farmers might make production choices
(i.e., plant traditional varieties) that result in suboptimal consumption levels.
The risk management strategies, together with the crop insurance and credit sections,
present the options farmers have to manage risk and how these complement each other; thereby
enabling farmers to make higher yielding production choices (i.e., grow modern varieties) and
move to a higher level of consumption smoothing, higher expected utility, and higher income.
The evaluating risk section presents a way to evaluate farmers’ choices (i.e., type of
variety to grow) within a context of risk exposure. It presents why a stochastic present value of
wealth utility model (SPVWUM)--derived from simulated outcomes in multiperiod scenarios
with quasi-permanent choices—is appropriate for evaluating the choice of varieties. Certainty
equivalent is the metric of comparison between choices. Additionally, the chapter includes a
description of other ways of evaluating decisions under risk. More specifically, it discusses the

6

use of dynamic programming within an income-consumption model, and also the use of
stochastic dominance as a criterion to evaluate choices.
The section discusses factors affecting adoption and briefly describes why a creditinsurance package could increase adoption of modern varieties and increase farmers’ expected
gross income.
Finally, the simulation section describes why simulation is a good approach to emulate
the farmer decision-making process and identifies studies that have used Excel add in @Risk for
simulation.
2.1. Honduran socioeconomic characteristics
2.1.1. Poverty in Honduras and the use of crop insurance in developing countries
According to the Grameen Foundation (2007), more than 66 percent of Honduras’
population lives below the poverty line of $2 a day and in rural areas and 61 percent of families
live below $1 a day. The Statistics National Institute in Honduras (INE) in (2010) reported 34%
of the rural population lives below $1 a day. The 2012 Honduran population was 8,385,072.
The rural population was 4,004,162 (INE, 2012). This provides an idea of how great an impact
an efficient credit-insurance package could have on poor households in rural areas. It also
provides a sense of the potential demand for credit-insurance package.
Various studies have found that the adoption of modern bean varieties increases farmers’
income. However, some experts have noted that due to the modern varieties longer maturation
period, compared to the traditional varieties (i.e., 10 day difference), modern varieties have a
higher yield variance--which increases farmers’ income variance (Smeltekop, 2005) because of
weather exposure. Crop yield insurance potentially reduces the downside risk which in turn
reduces income variance, and at the same time encouraging the use of modern varieties to

7

increase farmers’ income. Many efforts are currently in place to introduce crop insurance into
developing countries. Most of these efforts focus on introducing weather index insurance.
2.1.2. Characteristics of dry bean farmers in Honduras
While this study of a credit-insurance package focuses on bean farmers with a significant
portion of land devoted to dry beans, a significant proportion of the bean farmers in Honduras
also grow corn. Approximately 300,000 families in Honduras grow both bean and corn (IICA,
2009). Thus, general background information of these two crops sets the stage for showing how
the credit-insurance package would need to work under those conditions.
2.1.2.1.

Dry beans and corn yields

In Honduras, staple crops are grown during two seasons—the Primera (May to midAugust) and the Postrera (mid-August to December). National corn yields for the Primera and
Postrera (i.e., from 1986 to 1999) averaged 1,573 kg/ha and 1,395 kg/ha, respectively. Bean
yields for the Primera and Postrera (i.e., from 1986 to 1999) averaged 632 kg/ha and 702 kg/ha,
respectively (INE, 2006).
In 2002, 72 bean farmers were surveyed in El Paraiso and Olancho departments. These
two departments were chosen because they have the largest bean area planted (i.e., El Paraiso 13%
and Olancho 15%, of the total planted area; SECPLAN, 1994). Eighty six percent and fifty
seven percent of these bean farmers grew corn in the Primera (2002) and Postrera (2001),
respectively (Pejuan, 2005). While, 35% of the bean farmers grew corn in one season, 54%
grew it in both seasons, and 11% did not grew corn at all. Farmers in El Paraiso and Olancho
planted a similar area to beans in the Postrera (2.12 ha and 2.02 ha, respectively), but in the
Primera, farmers in El Paraiso planted more area to beans than farmers in Olancho (1.37ha and
0.76 ha, respectively) (Table 1).

8

Table 1. Average area and average of farmers planting corn and bean in Postrera (2001)
and Primera (2002), by department, Honduras.
Average of farmers planting crop area
Average area (ha)
(ha)
Department
Department
El Paraiso
Olancho
El Paraiso
Olancho
Season
Bean
Corn
Bean
Corn
Bean
Corn
Bean
Corn
Postrera
2.12
1.16
2.02
1.81
2.12
2.45
2.02
2.72
36
36
36
36
36
17
36
24
N1
Primera
1.37
2.37
0.76
2.93
1.70
2.59
1.04
3.63
36
36
36
36
29
33
26
29
N2
N = 72
Source: Pejuan farmer survey, 2002.
Selection bias is a problem in comparing varieties yield differences in Honduras (Mather
2003). Mather (2003) reported that average yields for modern bean varieties tended to be higher
than average yields from traditional varieties, although there was no significant difference
between the yield of two types of varieties (Table 2). However, this absence of a yield
difference is due to selection bias—farmers who planted traditional bean varieties lived primarily
in locations where disease (BGMV) pressure was low. However, farmers who planted modern
varieties lived primarily in areas where disease pressure was high. Thus, farmers could have used
modern varieties either because of lower expected net revenue from planting BGMV-susceptible
traditional varieties or because of a low yield from traditional varieties in years of high disease
incidence, or both. To correct for selection bias, Mather applied Lee’s (1978) two-step
3

procedure to estimate yield differentials between modern and traditional varieties . Mather
found the yield gain percentage difference from growing modern varieties if a farmer is located
in a disease risk area.

3

The counterfactual to modern variety yields is defined as what the yields of modern variety
growers would be if they had continued growing traditional varieties.
9

Table 2. Mean yield (kg/ha) of traditional and modern bean varieties, 1999-2000, Honduras.
Type of Variety
Primera
Postrera
1999

2000

1999

2000

Traditional

678

632

615

459

Modern

857

769

667

446

N = 188

N = 203

N = 242

N = 268

Source: Mather (2002). CRSP/PROFRIJOL Farmer Survey, 2001 (N = 210).
Smeltekop (2005) asserted that farmers have the misperception that improved varieties
always do poorly in drought conditions. This is because based on their limited experience of
observing modern varieties perform poorly under drought, farmers generalize to similar but not
identical circumstances. Nevertheless, Smeltekop recommended that scientists give priority to
developing bean varieties that perform well in drought situations (especially terminal drought) by
selecting varieties for early maturity or some other source of drought tolerance.
Pejuan (2005) reported that a larger percentage of the bean area in both El Paraiso (65%)
and Olancho (53%) was planted to traditional varieties—which indicates that a significant
portion of farmers still plant traditional varieties.
There have been government programs that distribute seed. The World Bank did an exante Poverty and Social Impact Analysis (PSIA) on Technical Production Voucher Program in
Honduras and concluded that it had the potential to positively impact the livelihoods of the
beneficiaries on the short-term (World Bank, 2010).
With Mather, Smeltekop and Pejuan’s findings lead to four questions: 1) what are
farmers’ perceptions of the probability distributions of yield of modern and traditional varieties?
2) what are the varieties’ characteristics that lead to those differences in the distributions, and 3)
how important is the early-maturing characteristic for bean varieties? 4) under what conditions
does crop insurance motivate farmers to change from traditional varieties to improved varieties?

10

2.1.3. Insurance schemes in Honduras
Crop insurance was introduced in Honduras at the end of 2001, but by October 2003
policy makers were dissatisfied with the low participation of farmers (Moneda, Oct. 2003). In
May 2003, the government proposed a bill requiring farmers who sought credit from government
banks to buy crop insurance. This bill was not welcomed by farmer groups, due to the limited
area across the country that the insurance companies covered (Moneda, May 2003), and it did
not pass as a law. In 2008, compulsive insurance was introduced and later eliminated for
individuals seeking credit from the National Development Bank (BANADESA, 2010). Currently,
it is strongly suggested.
In November 2003, two private insurance companies began to offer crop insurance–El
Ahorro Hondureño and Seguros Atlantida. In 2004, two additional companies--Cooperativos
Equidad and Seguros Continental--also began to offer crop insurance (Moneda, Oct. 2003).
However, Cooperativos Equidad only provided insurance on a trial basis. Policy makers
expressed hope that the entry of these new companies into the insurance market would result in
lower premiums and stimulate greater farmer participation (Moneda, Oct 2003).
The two insurance companies with higher market share are Seguros Atlantida and
4

Seguros Ficohsa (which has gained much share from what used to be El Ahorro Hondureño ).
Both offer similar insurance schemes, but there were a few differences in terms of premiums.
At present, these insurance companies insure all perils as a bundle and the premium oscillates
5

around 9-10% of the insured amount (Castillo, 2006). Coverage levels are around 40% .

4

El Ahorro Hondureño became part of BGA and later purchased by HSBC. HSBC stopped
selling crop insurance contracts.
11

6

Insurance contracts in these companies have the participation loss (i.e., 30% for drought
and 15% for excess rain). The policy’s premium ranges from 7.5% to 9.5% of the liability base
(liability before adjusting for coverage), depending on the region and the perils that is being
insured for.
Sometimes insurance companies offer a collective insurance policy to a group of farmers,
each with 1-5 mz (0.7-3.5 ha), but indemnization is paid on an individual basis (i.e., only the
farmers who suffered the losses are compensated).
2.2. Consumption Smoothing
Consumption smoothing refers to the idea that farmers will balance their spending and
saving to achieve the highest expected utility (Morduch, 1995). When positive shocks to the
farmers’ income occur, farmers would save. When negative shocks occur, farmers would
borrow or use their savings. Crop insurance helps smooth consumption by providing
indemnization in the low yielding years.
2.2.1. Income risk tradeoff
The income-risk tradeoff is discussed in the portfolio literature. Sometimes individuals
can only achieve a higher expected income by incurring higher risk (e.g., cash crops). It is up to
individuals, given their risk preferences, to position themselves on the income-risk frontier and
maximize their utility. With complete markets (i.e., ideal market where a complete set of statecontingent claims can be made with existing contracts, and which help smooth consumption) and

5

With an indemnity price of beans at L. 8.87/kg; a L. 3,571/ha liability and 1,056 kg/ha yield for
a low technology farmer (i.e., as considered by the insurance company) and L. 5,000/ha liability
and 1,319 kg/ha yield for a medium technology farmer, highest coverages are around 45% (i.e.,
43%).
6
Participation loss is a term used in the Honduran insurance contracts to refer to the copay.
12

no credit rationing, the Fisher separation theorem, which asserts that investment decisions to
maximize the present value of an enterprise are independent of the owner’s preferences, would
apply to farmers. The problem arises when farmers are credit constrained, and therefore their
preferences play a more important role in where they position themselves on the income-risk
tradeoff line. When risk and credit constraints are present, farmers make production choices
like planting traditional varieties that result in suboptimal production and consumption levels.
This is because farmers cannot or will not buy modern varieties which increases income, and
therefore be located in a lower level of the intertemporal budget line and consumption levels
would be suboptimal.
2.2.2. Modern bean varieties yield higher than traditional varieties
This section is introduced to support the previous point, regarding differences in income
between traditional and modern varieties. First, yields are compared and subsequently, income.
On-farm trials in Honduras indicate that modern bean varieties yield higher than
traditional varieties, especially when the optimal amounts of inputs are applied, including
fertilizer (Smeltekop, 2005). Analysis of survey data, in contrast to on-farm trials, indicates that
the yield differential is mainly due to disease resistance of modern varieties and susceptibility to
disease by traditional varieties, especially to the Bean Golden Mosaic Virus (Mather, 2003).
Differences in yield due to fertilizer have not been found in survey data. Planting modern
varieties do increase expected income in regions where disease pressure is present due to their
disease resistance (Mather, 2003) and to drought resistance (Rosas, 2009; personal
communication).
Although modern varieties expected yield is higher due to disease resistance, Smeltekop
(2005) reported that bean farmers stated that they do not want to risk a low yield in a given year

13

due to terminal drought. Smeltekop defines terminal drought as “when rainfall ends and does not
resume over the life of the plant.” Smeltekop also reported that drought is the factor farmers
expressed to be most constraining to increase yields. The reason farmers pointed out that
traditional varieties outperform modern varieties in terminal drought years is that the maturity
period for the traditional variety is shorter by 10 days. These results depend upon the region of
the country where beans are being produced. Thus, farmers perceive they face an income-risk
tradeoff-- they can only obtain a higher expected income by increasing their exposure to risk. As
a result, they choose traditional varieties with lower yield variability over modern varieties with
higher yield variability. However, Rosas (2009, personal communication) stated that the most
recently released modern bean varieties (i.e., from 2007 and later) are more drought tolerant than
traditional varieties. Still, many farmers are still with the perception of older improved varieties.
2.2.2.1.

Higher income from modern varieties

In Honduras, traders discount the price they pay for modern varieties, due to their darker
red color (Martel, 1995; Mather, 2003; Pejuan, 2005). However, due to modern varieties’ higher
yield, it is thought that they generate a higher net return per hectare—compared to traditional
varieties. To assess this hypothesis, a partial budget analysis is presented below, in which one
hectare of a modern variety is substituted for a traditional variety (assuming all other costs are
the same). The information needed to build the partial budget (Table 3) is the following:


The price discount for modern varieties is 10-15% in the Postrera season (Martel,

1996; Mather, 2003; Pejuan, 2007).


Cost of modern variety seed and traditional variety grain are L. 33/kg and L.

10/kg, respectively. (Key informant interviews, 2007).

14



Modern varieties yield 28% more than traditional varieties (Mather, 2003),

assuming that the same quantity of fertilizer (i.e., 2.85 hundred-weight 12-24-12 or 18-46-0 per
ha) is applied on both types of varieties.


Honduran average bean yields of 750 kg/ha (INE, 2007).

Table 3. Partial budget analysis: change in income for substituting one hectare of a modern
bean variety for a traditional bean variety.
Increase
Increased income:
(750 kg/ha * 1.28) * L. 10/kg
L. 9,600
Reduced cost:
0
L.
0
Total increase
L. 9,600
Decrease
Reduced income:
750 kg/ha * (L. 10/kg * 1.10)
L. 8,250
Increased cost:
Seed cost (L. 33/kg – L. 10/kg) * 45.45 kg/ha
L. 1,045
Harvesting and threshing (750 kg *0.28)* L1.1/kg
L. 231
Total decrease
L. 9,526
Net change

L 74

Since the total change in income is L 74 (L. 9,600 – L. 9,526), it is profitable for a smallscale farmer to switch from growing a traditional to a modern bean varieties, if the farmer buys
new certified seed every planting season. However, if farmers amortize the cost of the seed over
three planting seasons, then the increased cost item changes to:
((L. 33/kg + L. 10/kg + L. 10/kg) * 45.45 kg/ha)/3 – L. 10/kg * 45.45 = 348
Now, given that the total change is L 771 (L. 9,600 – L. 8,829), it is even more profitable
to switch from planting a traditional to a modern variety. Note that the above analysis applies to
small-scale farmers. The budget for a large-scale farmer would be different—given that they
apply fertilizer at a higher rate per hectare and modern bean yield would be higher.

15

The change in variety package is also profitable for all villages, since the marginal cost of
applying fertilizer (L. 8.94/kg) is less than the marginal income from the additional product (i.e.,
a minimum of L. 9.22/kg).
2.2.2.2.

Perceived higher risk from modern varieties

Previously, it was indicated that risk affects choices. This section denotes farmers’
perceptions of modern and traditional varieties and how this affects their choice of which variety
to plant.
Smeltekop (2005) pointed out that modern varieties are less drought tolerant than
traditional varieties and require more days to mature, which predisposes them to the yield losses
due to terminal drought. Thus, it is hypothesized that while modern varieties have a higher yield
potential, they also have higher yield variability (i.e., a thicker lower end tails in their yield
distribution). This hypothesis assumes that for modern bean varieties, drought is a more
important constraint than disease, since modern varieties are more disease tolerant than
traditional varieties. If this is correct, the farmer needs to position him/herself on the incomerisk tradeoff line according to his/her risk preferences.
Also, as mentioned before, Rosas (2009, personal communication) pointed out that the
most recently released modern bean varieties are more drought tolerant than traditional varieties.
However, the hypothesis that the probability distribution of the yield of modern bean varieties
has a thicker tail than that of a traditional variety is probable in certain villages where some
traditional varieties are well established and perform well. This type of relative shapes in the
distribution was more an issue of the past (before 2005), when modern varieties had not been
bred and released for drought resistance. Nevertheless, it was evidenced in Smeltekop study
(2005) that farmers thought, and possibly, currently still think that the relationship persists.

16

Consequently, concerned about the possibility of drought, some farmers may continue to plant
traditional varieties in order to secure at least a minimum amount of production. Furthermore,
some farmers plant modern bean varieties that were released more than 15 years ago (1997)
when drought resistance had not yet been incorporated into modern varieties, which helps to
explain why some farmers still believe that the lower end of the tail of the yield probability
distribution is thicker for modern varieties, compared to traditional varieties.
It should be noted that the higher yield from modern varieties (i.e., in partial budget) is an
expected higher yield. This means that modern varieties not all years have a higher yield,
however, on average modern varieties have a higher yield compared to traditional varieties. The
main reason that modern varieties yield higher than traditional varieties in farmers’ fields is
disease resistance (Mather, 2003) and the main reason modern varieties do not yield higher than
traditional varieties in all years is terminal drought (Smeltekop, 2005). The latter depends where
in Honduras the beans are being produced. Thus, it makes sense for a risk neutral or low risk
averse farmer to choose the modern variety due to its expected higher yield, and it also makes
sense for a medium to a high risk averse farmer to choose a traditional variety because of
terminal drought years.
2.3. Risk Management
2.3.1. Risk management strategies
According to Ray (1981), farmers can manage agricultural risk by avoidance, prevention,
speculation, insurance, and self-insurance (Kurosaki, 1997). Avoidance refers to escaping those
risks that are foreseen, which can be largely avoided by moving to production areas where such
risky factors are not present (or at least less present) in order to ameliorate the risks. In other
literature, researchers refer to avoidance, not by escaping, but by confronting the problem—such

17

as when farmers plant a disease-resistant variety. Prevention refers to the reduction of risk by
using improved facilities and techniques and organization. Thirdly, insurance refers to how the
farmers mitigate the effect of risk by transferring the risk to someone willing to assume the risk
in return of a fee (insurance premium). Self-insurance refers to 1) the accumulation of funds or
capital; 2) by operating on a large scale when it reduces correlative risks; 3) pursuing activities
that reduce risk by employing risk-reducing inputs; or 4) by farmers diversifying their portfolio
of assets.
2.3.2. Risk management strategies used by Honduran farmers
Given the fact that small-scale farmers in Honduras are settled in small villages, with
limited access to transportation or capital to buy more land in other regions from farmers in those
regions, it is implausible for them to manage agricultural risk by avoidance and prevention, as
Ray (1981) defines it. Thus, farmers can only reduce risk through by planting disease- resistant
varieties and by purchasing insurance from an insurance company. Even when modern varieties
are disease resistant, the perceived income variation due to other factors (e.g., drought, heat
sensibility) makes farmers hesitant to switch from planting traditional bean varieties to planting
modern bean varieties. To a certain extent, farmers can manage risk themselves by planting
several crops, which allows them to diversify their portfolio of assets. The majority of smallscale bean farmers in Honduras plant beans together with corn in relay. Many of these farmers
also raise livestock and poultry. These two additional activities diversify farmers’ portfolio,
which reduces their income risk--possibly to the extent that they perceive the insurance premium
to be too high, compared to the risk transferred to the insurance company.

18

2.3.3. Risk attitudes and portfolio
Individuals’ risk attitudes are separated into three categories: (1) risk averse, (2) risk
neutral, and (3) risk loving. The income-risk tradeoff that results from the different technologies
used by farmers, in combination with their risk attitudes, will determine the portfolio of activities
that farmers choose. Generally, farmers are assumed to be risk averse. With evidence of credit
constraints, farmers will choose the portfolio that will maximize their expected utility. This
portfolio could be improved and yield a higher expected utility by either removing credit
constraints or removing risk from a farmer’s portfolio.
2.3.4. Diversifiable and covariate risk, and systemic and idiosyncratic risk
Farmers risk can be divided into diversifiable risk and covariate risk. Diversifiable risk
refers to the risk of a productive activity that can be mitigated by including more activities.
Covariate risk refers to risk that cannot be removed through diversification (e.g., pests, low
rainfall). If we assume a farmer faces covariate risk, we could further separate that risk into
systemic risk and idiosyncratic risk. The systemic risk refers to the risk that affects all other
farmers in the area as well, while idiosyncratic risk refers to the risk associated with the specific
farmer. Crop insurance would, in its best scenario, eliminate part of the covariate risk. The part
of that risk that it would eliminate would depend on the type of contract. For example, area yield
insurance would eliminate more of the systemic risk, while individual yield insurance would
eliminate part of both idiosyncratic risk and systemic risk.

19

2.4. Agricultural Insurance
2.4.1. Types of insurance
In developed countries, there are several types of insurance schemes offered to farmers,
including individual yield insurance, area yield insurance, individual revenue insurance (Ray,
1981), and weather index-based insurance (Hess and Syroka, 2005; The World Bank, 2005).
2.4.2. Adverse selection, moral hazard, and basis risk
Asymmetric information problems like moral hazard and adverse selection may cause the
insurance market to fail (Rothchild and Stiglitz, 1976; Just, R.; Calvin, L.; and Quiggin, J.; 1999).
Adverse selection occurs when the insurance company charges the same premium to farmers
with high and low risk (Williams, et.al.1993) due to the information asymmetry of insurance
companies not knowing who are the high risk and low risk farmers. Miranda (1991) asserts that
farmers will have better information about their outcome distributions and therefore buy
insurance, if the benefits outweigh the premiums. It follows that if a concentration of high risk
farmers purchases crop insurance but low risk farmers do not due to a relatively high premium,
the insurance company will either incur losses, increase their premiums and thereby create a
high-risk concentration of farmers, or try to correct the adverse selection through other contract
designs. Miranda (1991) and Williams et.al. (1993) also states that moral hazard occurs when a
farmer changes production practices that will increase his/her chances of being indemnized.
Several approaches in crop insurance contract designs are discussed in the literature that
addresses moral hazard and adverse selection problems.
When institutions, that regulate or design contracts, are in place to manage risk aversion
and credit, most of these problems are mitigated. Each type of insurance requires different types

20

7

of data , needs different structures in place in the country, and is subject to different problems in
varying degrees. For example, to establish an individual yield insurance scheme, the lender
needs information about the insured individual’s yield history (corroborated with some type of
proof). An advantage of this scheme is that farmers are not subject to basis risk (i.e., risk due to
the mismatch between when the farmer experience losses and when the farmer gets an indemnity
payment). However, individual insurance schemes are vulnerable to problems of moral hazard
and adverse selection. On the other hand, area yield insurance schemes do not need information
about individual yield history. Instead, they utilize a yield index of an area, which is the trigger
for indemnity payments. While area yield insurance has less of a problem with moral hazard and
adverse selection, it suffers from basis risk.
2.4.2.1.

Group insurance, moral hazard, and adverse selection

Pooling farmers to reduce transaction costs for purchasing insurance could introduce
hidden action and hidden information problems (e.g.,high risk farmers might act even more risky
when pooled with low risk farmers) (Armendariz and Morduch, 2005).
2.4.3. Weather index-based insurance
Recent research studies that have a promise for increasing insurance participation in
developing countries have focused on weather index-based insurance; for example, rain
insurance, where farmers receive an indemnity if rainfall falls below a certain level (World Bank,
2011; World Food Program, 2011). Some insurance policies also include excess rainfall triggers
since excess rainfall has become a problem in recent years. Weather index-based insurance

7

Data for designing insurance contracts are usually not available with the quality required.
Rainfall data is available for several rainfall stations. Nevertheless, time series data at the
municipality levels are not available and seldom individual farmers have historical yield data.
21

largely eliminates problems of moral hazard and adverse selection, which reduces the premiums
and increase participation--conditional that the index is highly correlated with farmers’ yields, so
that the basis risk is reduced to the maximum extent possible.
A weather index-based insurance contract should have a well-established system for
collecting data on weather-related variables. In addition, weather conditions affecting farmers
should not vary greatly from those at the weather station, so that the index and farmers losses are
highly correlated. The high correlation will reduce the basis risk.
2.4.4. Insurance and probability distributions
Insurance changes the income probability distributions by changing the possible
outcomes. This is because the outcomes, income, and/or expenses change when a farmer
purchases insurance and when the farmer is indemnized due to a peril. Due to the change in the
probability distribution, a farmer would make an optimal choice of type of variety because the
risk is transferred and would reduce the effect of risk on her/his choice.
2.5. Credit
2.5.1. Microcredit history in Honduras
Microcredit in Honduras is offered through small financial institutions called “savingand-credit rural cashiers”. These were established in 1983 by a FAO initiative to help women in
rural areas. Eighteen of 24 originally-established rural cashiers still function today (CEPAL,
1999).
These microfinance institutions extend loans to women. These women must be members
of the rural cashier, which they join by participating in a training program (CEPAL, 1999).

22

2.5.1.1.

Rural cashiers

In 2002, approximately 2,092 rural cashiers operated in Honduras. Thirty-one percent of
these rural cashiers extended loans of less than L. 5,000 and a similar percentage offered loans of
between L. 5,000 and L. 20,000. These rural cashiers require mortgage or pledge collateral.
8

Interest rates charges range from 3% to 5% monthly rates (Berdegue, 2000).
2.5.1.2.

Grammeen Bank

Fundacion Adelante is the Grameen’s Bank microfinance institution partner in Honduras.
This microfinance institution mainly serves women in rural areas
(http://www.grameenfoundation.org, 2007).
2.5.2. Use of credit and insurance
Implementation of a credit-insurance package for small-scale bean farmers is based on
the idea that peril insurance will deal with covariate risk and microcredit will provide access to
capital to obtain the modern variety package and also deal with the idiosyncratic risk.
Access to credit allows capital-constrained farmers to acquire cash to purchase modern
varieties and complementary inputs (Beke, 2011). Crop insurance reduces risk and may
potentially increase the farmers’ willingness to purchase modern variety package for risk averse
farmers (Gine and Yang, 2008). Thus, a successful credit-insurance package is expected to
increase farmers’ incomes through the purchase of modern varieties and complementary inputs.
With peril insurance, hidden action and hidden information are not much of a problem. As
pointed out earlier, when farmers are pooled to reduce transaction costs for purchasing insurance,
this could introduce hidden action and hidden information problems.

8

Monthly inflation rates in Honduras in 2007 were of 0.6% (Honduran Central Bank, 2007).
Thus the real monthly rates range between 2.39% and 4.37%.
23

2.6. Evaluating Risk
This section presents ways to evaluate farmers’ choices (i.e., type of variety package to
grow, and use of crop insurance and credit) within a context of risk exposure. One way is an
income-consumption model with the use of dynamic programming. Another is a stochastic
present value of wealth utility model (SPVWUM) derived from simulated outcomes in
multiperiod scenarios with quasi-permanent choices. The use of insurance and credit can be
incorporated in both types of models.
To justify the use of crop insurance for bean farmers, there is a need to first assess the
level of risk that farmers face with each choice of technology, including the tools that farmers
apply to mitigate risk. If farmers choose a traditional variety over a modern variety because of a
higher perceived risk, there is a chance that the choice can be reversed if insurance is purchased,
assuming that the farmer has access to purchase the modern variety. Both, the incomeconsumption utility model using dynamic programming and, the SPVWUM can evaluate that
kind of farmers’ choices. Stochastic events like farmers’ yields and prices are incorporated into
both kinds of models.
To run a stochastic income-consumption model or a stochastic wealth utility model where
farmers perception of yields are evaluated, farmers’ yield distributions are required, among
other things that are described in this chapter. In addition, even when objective data indicate that
farmers would be better off planting modern varieties, sometimes farmers’ perceptions of their
yield distributions lead them to believe that they are better off planting traditional varieties.
Therefore, elicitation of farmers’ subjective yield probabilities can help explain to what extent
farmers perceive that they face risk and thereby help explain farmers’ use of traditional over
modern varieties.

24

The yield distributions would also help explain the participation or lack thereof in
insurance schemes. Unfortunately, it is usually not possible to test for differences between
objective and subjective probability density functions of small-scale farmers in developing
countries like Honduras because objective probabilities are derived from historical data. This
would require yield data for at least 15 years, but such data are not available because farmers in
Honduras do not usually maintain records of past yields.
Existing insurance policies can be evaluated by comparing farmers’ income distributions
with and without insurance, and thereby determine the consistency of insurance premiums with
farmers’ level of risk. This is possible because insurance changes the income probability
distributions by changing the possible outcomes. The combination of income probability
distribution, outcomes of the distribution, the insured utility function, and the insurance design
with its premiums, determine the new risk faced by the insured when they purchase insurance. It
is the comparison between these risks (with and without insurance) that determine the extent of
risk transfer and the farmers’ perceived soundness of the insurance premium, relative to its risk
transfer.
2.6.1. Income consumption model
A dynamic model, where households maximize expected utility of consumption, is a
starting point for evaluating farmers’ decisions on obtaining credit and purchasing crop insurance.
Many studies have employed stochastic dynamic programming models, using the Bellman
Equation to analyze and optimize choices (Atwood et. al., 1996; Nyambane, 2005).
Studies that use dynamic programming for analysis have the objective of using the
Bellman equation to obtain a set of contingency paths that lead the choice maker to maximize
utility over time, as she/he responds to possible states of nature in each time period. The

25

different states of nature affect a farmer differently depending on which type of technology and
other mechanisms she/he uses to smooth consumption.
2.6.2. Income-consumption model studies
Many household models are framed so that farmers choose among actions to maximize
expected utility in their risky decision-making. Deaton (1991), who deviates from the traditional
life-cycle models by incorporating a liquidity constraint in the form of a borrowing restriction,
shows how consumers use savings to protect their consumption against bad draws. Deaton’s
main point is to explain household behavior when they are exposed to borrowing constraints. To
analyze household behavior, Deaton used programming and simulation based on Bellman’s
Equation of an infinite horizon expected utility maximizer.
A brief explanation of dynamic programming using the Bellman equation is required at
this point. The result of Bellman’s equation is a set of decision rules (or policy functions) that
indicate the optimal control (or action) to be taken at each time period and possible states of
nature. That is, the policy function tells the decision maker a complete set of contingent optimal
actions to take according to the unknown states of nature at each period where a decision needs
to be made. The state of nature is the realization of an uncertain condition that the decision
maker faces. For example, if a farmer wants to decide whether or not to purchase crop insurance
over several periods, the previously mentioned state of nature could be the amount of the rainfall
in the next growing season which is uncertain, but which is observed for the previous period;
then, the optimal choice of insurance could be a function of the amount of rainfall in the previous
period according to what the policy function dictates to do for that specific level of rainfall and
the possible outcomes related to that state of nature.

26

Carroll (1997) works with a model that describes saving behavior as a “buffer-stock”,
rather than a Life Cycle/Permanent Income Hypothesis model. The main contribution of Carrol
is that his model reconciles many behaviors, while other models seem to be able to explain just
one type of behavior (i.e., Carroll reconciles the consumption/income parallel and divergence).
Carroll uses simulation as a tool and applies it to a finite and infinite horizon expected utility
maximizer model. While previous studies assumed that the variance of incomes was constant,
Carroll incorporates a variable second-order variance of income. This model is said to be a close
substitute of Deaton’s model, since the model is very similar except that Carroll does not directly
impose the liquidity constraint and his transitory and permanent shocks are independent.
Osborne (2006) recently worked on comparing credit and insurance as substitutes. The main
contribution of Osborne’s article is that it contrasts the use of credit and insurance in agriculture
in developing areas. To analyze household behavior, Osborne used simulation--based on
Bellman’s Equation of an infinite horizon expected utility maximizer.
2.6.3. Utility of wealth models using simulation
Instead of using an income-consumption model where utility is obtained from
consumption, we could have a model with income and consumption where individuals obtain
utility from wealth. While utility of wealth can be also incorporated in a model using the
Bellman equation--like it is done in income-consumption models—using utility of wealth is
necessary to do if consumption and choices are held fixed for its scenarios in a multiperiod
setting. This way, the different choices yield different ending wealth for each period.
In simulation, choices are represented by the combination of different levels of choice
variables and then compared with a certain metric of interest. Sometimes optimization is
introduced in studies in a second stage (Berg, 2002).

27

Crop insurance alternatives and risk management practices have been analyzed using
simulations with predefined scenarios and choices (Jansen, 2012; Umarov, 2009). Thus, choices
of such as the choice of a technology package (i.e., improved technology package) within certain
representative scenarios (i.e., area planted, yields, prices, costs; with its predefined parameters),
and the alternatives of use of crop insurance and credit fits well into simulation-scenario-choice
9

analysis .
One metric of interest commonly used in insurance studies is expected utility or its
counterpart the certainty equivalent. The benefit with these metrics is that it incorporates
farmers risk aversion into the analysis.
The certainty equivalent of a risky outcome is the value amount that makes a decision
maker indifferent between accepting a sure but lower amount, and the higher expected but risky
amount (Usategui, 2009).
In a stochastic multiperiod frame, expected utility can be obtained in many ways (e.g.,
discounting certainty equivalents from each period, discounting expected utilities at each period,
and applying a utility function to net present values). The expected utility is used to obtain the
certainty equivalent (CE), since CE is also the inverse of the utility function at the expected
utility. A way to obtain the expected utility that appears to be appropriate --compared to several
other methods-- is the method where a set of net present values are obtained and later the utility
function is applied (Baucells and Sarin, 2007). Thus, to obtain the certainty equivalent, the
inverse of the utility function at the expected utility value with the latter method would be
appropriate.

9

In the business literature, scenarios are the base characteristics. Added to the scenario,
stochastic outcomes are added through simulation.
28

While the objective in dynamic programming is to identify the optimal policy functions
that dictate the actions to take depending on the state of nature, in simulation the actions are
fixed and the results over several periods of each action are compared to determine which action
is preferred. Farmers’ most common choice scenarios are compared to observe if farmers’
behavior is consistent to that of an expected utility maximizer.
2.6.4. Stochastic dominance (SD)
Another metric for comparing choices is the stochastic dominance criterion. This
criterion compares the distributions of the desired variable. For example, if we wanted to
compare two variety yields, the yield cumulative distribution functions (cdfs) would be
compared using the stochastic dominance criterion. Generally, there are two types of stochastic
10

dominance criterions used, first and second degree stochastic dominance . The three possible
results from comparing two distributions using first degree stochastic dominance (FDSD) are: 1)
distribution one (choice) dominates the distribution (choice) two in the first degree, 2)
distribution two (choice) dominates the distribution (choice) one in the first degree, and 3)
neither distribution (choice) dominates in the first degree. The same similar three results are
applicable for second degree stochastic dominance (SDSD), except that it is referred to as
dominance, or not, in the second degree. When one choice (i.e., a modern variety) dominates a
second choice (e.g., a traditional variety) in the first degree with respect to a certain variable (e.g.,
gross income), all individuals (e.g., farmers) that want more of the variable being analyzed (i.e.,
want more gross income) will pick the first choice (i.e., modern variety). When one choice

10

There are other types of stochastic dominance: third degree, fourth, …, nth degree.
29

stochastically dominates a second choice in the second degree, all risk averse individuals that
want more of the variable being analyzed will prefer the first choice.
The criteria for FDSD are the following:
For an increasing utility for x, choice F(x) dominates G(x) in the first degree if and only
if F (x)  G( x) for all x with at least one strict inequality, where F(x) and G(x) are the cdfs of
choices F( ) and G( ), respectively (Levy, 1973). Thus, FDSD compares the height of the cdfs of
both curves at all points, F(x) and G(x).
The criteria for SDSD are the following:
For an increasing and concave utility (i.e., wanting more of something and risk averse)
for x, choice F(x) dominates G(x) in the second degree if and only if:
x

x





 F ( x )   G( x) for all x with at least one strict inequality, where F(x) and G(x) were

previously defined (Levy, 1973). Thus, SDSD compares the area under the cdfs of both curves,
F(x) and G(x), at all points.
An advantage of the stochastic dominance criterion is that is free of a utility function.
That is, there is no need to assume a particular utility function. The disadvantage of this criterion
is that there is no single number to observe when comparing choices, compared to certainty
equivalent, and there are many cases where preference of one choice over another cannot be
determined for one particular type of farmer.
2.6.4.1.

Cumulative Density Functions and intuition to stochastic dominance

A cumulative density function (cdf), F(x), is a function that describes the probability that
a random variable “x” is below a specific value (Wilcox, 1996). That is, if “x” is the random
variable yield, in kg/ha, from a modern variety and the cdf specifies F(x=1,000)=0.294, it

30

describes that the probability that the modern variety yield will be below 1,000 kg/ha is 0.294,
the height of the function at 1,000 kg/ha in Figure 1.
Thus, FDSD is very intuitive-- if a choice, for an individual who always wants more of a
good (e.g., yield), always or almost always has a lower probability of being below a specific
value but never higher than another choice, the former choice will dominate the latter choice in
the first degree.
Figure 2 shows the hypothesized cdfs of yield (kg/ha) of a modern variety and a
traditional variety. No variety stochastically dominates the other, neither in the first nor in the
second degree. The modern variety does not stochastically dominate in the first degree because
the value of the cdf for the modern variety is not always equal or lower than the cdf of the
traditional variety. That is, the traditional variety cdf value at the low end of the distribution is
higher. Also, the area under the cdf for the modern variety is not always equal or lower than the
cdf of the traditional variety. However, the use of insurance with a modern variety might change
the revenue distribution and dominate the traditional variety revenue distribution in the second
degree (Figure 3).

31

Figure 1. Cumulative density function of yield (kg/ha) from a modern dry bean variety. For
interpretation of the references to color in this and all other figures, the reader is referred
to the electronic version of this dissertation.

Figure 2. Cumulative density functions of yield (kg/ha) from a modern and a traditional
dry bean variety.

32

The contribution of insurance is shown in Figure 3, where modern and traditional variety
gross income distributions without insurance do not dominate each other neither in first degree
nor in the second degree of stochastic dominance. However, the modern variety with insurance
clearly dominates the traditional variety in the second degree.

Figure 3. Cumulative density functions of gross income (L.) from a modern* and a
traditional** dry bean variety without insurance and a modern variety with insurance***.

*The modern variety without insurance is the irregular green dashed curve.
**The traditional variety without insurance is the blue smooth line.
***The modern variety with insurance is the regular red dashed curve.
2.6.5. Improved bean package
Improved variety packages include use of modern variety, fertilizer, and integrated pest
management practices. Manuals in Central America for bean farmers recommend use of modern
seed, control of pests by using insecticide, and applying fertilizer according to soil analysis.
Nevertheless, if a soil analysis is not available, Rosas (2003) recommends fertilizing with 65-130
kg/ha of formula (i.e., 18-46-0 or 12-24-12).

33

The use of a modern package is said to increase income through increased yields, even
though grain prices of modern varieties in Honduras are lower than that of traditional varieties
(Smeltekop, 2005). Nevertheless, these modern packages sometimes increase income variability.
Thus, research in the use of crop insurance to reduce income variability and increase adoption of
modern varieties is needed.
The availability of a credit-insurance package is hypothesized to encourage farmers to
adopt the improved bean production package, which includes certified seed, fertilizer (200 lbs of
18-46-0 or 12-24-12 at planting time and 100 lbs of urea before flowering), and timely pesticide
application. If credit is available but insurance is not, farmers may not adopt the full modern
bean production package, due to their risk aversion. Similarly, if insurance is available, but
credit is not, farmers may be unable to adopt the full modern bean production package, due to a
lack of capital. Thus, it is hypothesized that a credit-insurance package is a tool that will
encourage farmers to adopt the package and increase their income.
2.6.6. Insurance and credit
Honduran farmers are said to be constrained by lack of access to credit (Smeltekop,
2005), which limits their use of a modern variety packages. By helping farmers to mitigate their
liquidity constraint, the credit-insurance package will enable them to adopt the whole improved
bean production package. By reducing their risk exposure, the credit-insurance package will
encourage risk averse farmers to adopt the technology package.
Before establishing the conditions for a credit-insurance package, it is necessary to
analyze why farmers do not obtain credit alone or insurance alone when liquidity or risk aversion
is and is not a constraint. Reasons why farmers do not purchase insurance when risk aversion is
not a constraint include risk exposure, insurance premium, and information asymmetry. On the

34

other hand, small-scale farmers may not utilize credit due to a lack of collateral (liquidity) or
because of high transaction costs, both on the farmers’ side as well as the lending institutions’
side (Armendariz and Morduch, 2005).
The reason insurance markets exist is because of risk aversion (Nicholson, 1998). If
farmers were risk neutral or risk loving, which they usually are not, farmers in general would not
demand insurance. Thus, we need to know what farmers’ risk preferences are. In addition,
farmers are able to manage their income risk by creating a portfolio of assets. If the risk
associated with the assets in the portfolio is not highly correlated, the portfolio reduces income
risk—leaving only some covariate risk. In the farmers’ case, the portfolio is composed of
income from other crops, off-farm income, and accumulated assets that help reduce their income
risk and smooth their consumption. Important questions to answer include: 1) which of these
risk management tools (e.g., assets) are the most important?, 2) do farmers use their portfolio to
reduce risk, rather than purchase insurance?, 3) do existing insurance schemes transfer
significant amount of risk from the insured to the insurance company (i.e., to what extent does
insurance reduce the certainty equivalent (C.E.) or income coefficient of variation, compared to
when insurance is not present)?, and 4) for the majority of the population that the insurance
policy targets, does the insurance premium

11

reflect the risk that is transferred from the insured

to the insurer?
Sometimes financial institutions do not offer credit and farmers do not demand it because
of the transaction costs involved. Transaction costs incurred by the financial institution include
activities associated with handling the loan and reducing risk (e.g., “screening” clients, designing

11

If the information asymmetry is high the insurance markets may fail because then the
premium becomes too high/unattractive to farmers.
35

a contract, handling delinquent accounts, and liquidating collateral in case of default).
Transaction costs incurred by the farmers include the interest rate, fees, commission, transport
and food used to get to the financial institution (Guzman, 2006).
2.6.6.1.

Yield and revenue contracts

Insurance companies in Honduras do not offer revenue contracts. Rather, they only offer
yield contracts. A comparison between yield and revenue contracts is presented below to show
ways that farmers could reduce their risk with each of the insurance products, and how risk
would be reduced if revenue insurance was offered. The following section presents a summary
of yield and revenue contracts jargon and abbreviations. This will help define and structure the
jargon which will be used in the simulation chapter.
Abbreviations:
Y = realized yield
EY = expected yield (in a probabilistic sense)
Cov= Coverage (%)
YG = yield guarantee
Liab= Liability
Loss
IP = indemnity price (price paid in the indemnization of a loss, which is established when the
contract is signed)
IND = indemnity
One version of a standard yield contract includes a yield guarantee, loss, indemnity,
liability, and premium rate, as defined below:
Yield contract (one version)

36

YG = (Cov %/100) x EY
Loss = max(YG – Y,0)
IND = Loss x IP
Liab = YG x IP
Rate = average (IND)/Liability
As it is seen in the indemnity function, the yield contract covers any yield shortfall of the
actual yield below the covered percentage of the expected yield, priced at the indemnity price.
Parameters of these types of contracts are introduced into simulation-scenario-choice
analysis.
For a version of a standard revenue contract, the same contract definitions for a yield
contract are defined below:
Revenue contract (one version)
RG = YG x EP
Rev = Y x P
IND = max (RG – Rev, 0)
Liab = RG
Rate = average (IND)/liability
The revenue contract covers any revenue shortfall of the actual revenue below the
revenue guarantee, which is priced at a futures price contract.
The difference between yield and revenue insurance is that revenue insurance insures for
falling below a certain threshold of income, which is the product of yield and a price, while yield
insurance only insures for a shortfall on yield. That means that if price falls sharply and yield
remains approximately around its average value, a farmer will not be indemnized, while in the

37

revenue insurance farmers would be indemnized. Thus, the inclusion of the price variable into
the revenue insurance contract, which is allowed to vary according to the futures market price, is
the source of difference between the two contracts.
2.6.7. Subjective yield distributions
Subjective yield distributions have been used in studies where the distribution of yields of
individual farmers is not available. These subjective probabilities then serve to evaluate risky
decision making.
Francisco and Anderson (1972) argued that decision makers have a working familiarity
with the concept of probability and that subjective probabilities can be readily elicited for
important random variables. Another study by Grisley and Kellogg (1983) reports an experiment
in Thailand to elicit farmers’ subjective probability distributions with respect to price, yield, and
net income of selected crops. The elicitation procedure by Grisley and Kellogg (1983) was
similar to that of Francisco and Anderson with the important difference that farmers were offered
a monetary reward, if their stated expectations turned out to be accurate.
The “visual impact method” or “visual counter” method by Anderson, Dillon and
Hardacker was the method used by Grisley and Kellogg as well as Francisco and Anderson. The
“visual counter” consists in asking farmers about their minimum yield and maximum yield.
After recording both values, the range is divided into five intervals. Next, 25 counters are
distributed across the five intervals according to the farmers’ belief of the frequency of each
yield interval (Grisley and Kellogg, 1983). The result is a frequency distribution (i.e., histogram)
from which means, variances, and other statistics can be obtained.

38

2.7. Factors Affecting Adoption
According to Feder, Just, and Zilberman (1985), factors affecting adoption include risk,
credit, input supply, land tenure, social learning, human resource, farm size and information.
Feder et. al. present information on all those factors previously mentioned, while other studies
focus mainly on a few factors (i.e., risk and credit). Feder’s et.al.broader view of those factors is
appropriate to present, however researchers need to prioritize the factors to include in their study
due to funding constraints.
2.8. Simulation
With simulation, the stochastic aspect of life is introduced into models and provides a
representation of reality in order to comprehend how factors affect the system represented. It
provides insight on current and future behavior of agents.
2.8.1. Reasons for using simulation
Crawford (1982) points out to three reasons for using simulation. These are (1) when the
objective is to explore the functioning of a whole system, (2) to observe the effects by changing
key parameters, and (3) when the system is complex, dynamic, and interactive. The simulation
design proposed in this study draws, to a certain extent, from the three reasons stated above.
First, there are several constraints that characterize farmers’ technology adoption behavior and
therefore a comprehensive system should be represented. Second, to answer several of the
research questions, key parameters will be changed to observe the effects and predict agents’
future behavior and/or justify their current behavior. Third, farmers’ decision making, regarding
the decision to purchase insurance is dynamic.

39

2.8.2. Objective
Simulation will provide an opportunity to explore farmers’ behavior under different
scenarios. This will ultimately translate into a policy suggestions.
2.9. Chapter Summary
Poverty in Honduras is widespread. The introduction of modern bean varieties to farmers
is expected to help increase farmers income and reduce poverty. The use of crop insurance and
credit encourages farmers to use modern varieties, which on average have higher net returns.
Credit access enables farmers that are credit constrained to purchase inputs. Crop insurance
potentially reduces the downside yield risk and reduces income variance. The reduction in yield
downside risk from the modern variety increases the likelihood that a farmer will choose the
modern variety.
In Honduras, most of the country’s corn and beans are produced by small-scale farmers,
which highlight the importance of focusing on how to improve bean yields. Two of the most
important bean-producing departments in Honduras are El Paraiso and Olancho, where farmers
on average produce 2.12 and 2.02 hectares during the Postrera, the most important bean
production season. The national yield average in the Postrera season is of 702 kg/ha.
Studies have shown that modern varieties recently released in Honduras yield higher than
traditional varieties when compared under similar circumstances. Even when this has been
documented and advertised, a significant portion of farmers still plant traditional varieties. This
leads to ask the following four questions: 1) what are farmers perceptions of the probability
distributions of modern and traditional varieties? 2) what are the varieties characteristics that
lead to those differences in the distributions? 3) how important is the early-maturing

40

characteristic for bean varieties? 4) under what conditions does crop insurance motivates farmers
to change from traditional varieties to improved varieties?
In Honduras, only one crop insurance scheme is available to bean farmers, which is the
cost coverage insurance with yield adjustment. This insurance could potentially help smooth
farmers’ consumption under certain circumstances, but its demand will depend on farmers’ risk
preferences and income-risk tradeoff.
To farmers, the desirability of crop insurance depends on the farmers’ perception of bean
yield distributions, insurance premium, income-risk tradeoff, farmers risk preferences and their
use of risk management strategies. Depending on the combination of risk-reducing strategies,
insurance premiums might be considered too high for farmers’ willingness to pay to purchase the
insurance policy in return for the risk transfer from the farmer to the insurance company that the
policy offers. Nevertheless, insurance companies have to deal with information problems like
moral hazard and adverse selection which increases the premiums.
The previously mentioned factors that affect the desirability of crop insurance are used in
simulation-scenario-choice models under the form of a stochastic present value wealth utility
model (SPVWUM). The use of SPVWUM is to assess if farmers could potentially adopt (i.e.,
under which conditions) a modern variety package by using crop insurance together with credit.

41

3. CHAPTER III: Conceptual Model and Methods
This chapter first justifies the use of a stochastic present value of wealth utility model
(SPVWUM) to assess if a credit-insurance package will increase adoption of modern varieties.
It then presents the simulation model characteristics used in the @Risk general model. The
variables and parameters needed to run the model using Excel @Risk 5.7 add-in are specified, as
well as the data sources, data collection methods, and the data itself. Finally, it presents the
factors (i.e., technology choices and scenarios) to determine and the sequence in which these
should be compared to answer if the credit-insurance package will increase adoption of modern
varieties.
3.1. Why Simulation?
As noted in Chapter 2, some researchers have used stochastic dynamic programming to
evaluate dynamic credit-insurance decisions, including infinite and finite horizon optimization
models, and rely on this tool to evaluate farmers’ decisions. Some other researchers have used
multiperiod stochastic simulation (i.e., simulation of multiperiod sequences) to evaluate the
efficacy of insurance decision strategies.
Both simulation and dynamic programming can incorporate stochastic processes and
compare choices, but dynamic programming identifies an optimal set of contingency paths, while
simulation compares quasi-fixed choices through time to decide among the choices introduced.
That is, simulation compares a few selected time paths from the whole set of time paths to then
decide which of the paths to take.
However, a drawback of dynamic programming is that as more factors are entered into
the model to represent reality, the dimension of the problem increases geometrically (i.e.,

42

dimensionality problem) (Bellman and Dreyfus, 1962), which makes programming extremely
difficult. Consequently, this study uses simulation.
3.2. Simulation Model
The chosen model for this study is a stochastic present value of wealth utility model
(SPVWUM) derived from simulated outcomes in multiperiod scenarios with quasipermanent
choices. The stochastic simulation model was applied using Excel add-in @Risk from Palisade.
A utility model captures farmers risk aversion which is the reason for their use of crop
insurance and one of the reasons farmers do not adopt technologies like a modern variety
package. Wealth instead of consumption is the source of utility in this model because
consumption and choices are held fixed or quasi-fixed for the simulation scenarios in a
multiperiod setting, and the stochastic outcomes are the driving force along the time path. In
addition, its multiperiod setting allows for evaluation of shocks, and decisions with outcomes
accrue over time.
Simulation is used, as previously mentioned, for its direct way of comparing among
choices; and the stochastic aspect of the model is suitable for representing risky choices. A utility
of present value of wealth approach was taken to not violate the principle of time preference of
money. Finally, the metric of comparison between choices is the certainty equivalent for being
this a currency value and have practical meaning.
3.3. Specifics of the SPVWUM
The SPVWUM is a 20 year simulation model. The model starts in year one with an
initial level of wealth. Consumption and stochastic production outcomes change the ending
wealth in that year, which then becomes the initial wealth in the following year. The stochastic
streams of marginal changes in wealth, which is the difference between terminal wealth of

43

continuous years, are discounted at the appropriate discount rate. The resulting present values
of stochastic streams are then introduced into an isoelastic

12

utility function with a constant

relative risk aversion coefficient. Finally, the metric of comparison, the certainty equivalent, is
estimated. This process is performed for different technology choices and scenarios given the
village the farmer is located at. The expected utility model is specified below in equation 1:

T MW cv

it
EU (W )  U icv WI  
t

i 1
t 1(1  r )

N

where




 Equation 1.


WI = initial wealth
Uicv= utility for random stream draw i, of choice c, and village v.
MW

cv
it=Marginal wealth for random stream draw i at time t,

of choice c, and village v.
r = is the time preference discount rate
Farmer choices are a combination of types of technology package (e.g.,traditional variety
and low inputs, modern variety and high inputs) and options of the credit-insurance package (e.g.,
insurance only, insurance and credit). Scenarios are a combination of parameters of the creditinsurance package (e.g., insurance coverage level, premium loading factor), and parameters of
the farmer (i.e., risk aversion, time preference). Both farmer choices and scenarios are evaluated
at different villages, where the village has an effect on production (e.g., yield) and marketing (i.e.,
price) parameters.

12

The isoelastic utility function is used to have a constant elasticity of substitution throughout
the range of the function. It also implies that the elasticity of substitution decreases with respect
to risk aversion.
44

The initial wealth, or initial capital, includes the value of farmers’ assets: land,
infrastructure, machinery, and farm animals. The marginal wealth is the increase or decrease in
wealth due to the contribution of gross income from the bean and corn activities. The marginal
wealth for random stream draw i, at time t, choice of c, and given village v derive from
consumption and production activities from different choices and villages. As mentioned above,
the choices farmers have are a combination of technology categories and the use of crop
insurance and/or credit for the bean enterprise (i.e., not for corn). The combinations of
technology categories and the use of crop insurance and/or credit are presented in Table 4. The
metric, certainty equivalent, which will be presented in each of the cells in tables similar to Table
4, will be compared among the choices of interest. Note that not all combinations in Table 4 will
be compared because not all combinations are applicable from the standpoint of this research.
Table 5 shows the acronyms for each of the combinations of technology levels, and
actions taken by farmers in terms of obtaining credit and crop insurance. This table will be
useful when comparing each of the choices within the cases of risk aversion researched in this
study. The two cases of interest in this research where the credit-insurance package could be
beneficial are: 1) farmers not constrained by capital but who are driven by risk aversion to
purchase traditional varieties instead of modern varieties because of the latter failure in drought
situations (i.e., risk averse and not capital constrained), and 2) farmers constrained by capital and
driven by risk aversion to not access credit to purchase a modern variety and complementary
inputs in case of a bad event happening and not being able to repay the loan (i.e., risk averse and
capital constrained).

45

Table 4. Choices (i.e., c=1, …,11) resulting from a combination of technology categories and use of credit and/or insurance that
will be compared on the metrics certainty equivalent and insolvency percentage, within each village.
Technology Category/credit and insurance

Choice number

Non-mechanized+traditional variety+low input use+no credit+no insurance
1
Non-mechanized+traditional variety+high input use+no credit+no insurance
2
Non-mechanized+modern variety+low input use+no credit+no insurance
3
Non-mechanized+modern variety+high input use+no credit+no insurance
4
Non-mechanized+modern variety+high input use+no credit+with insurance
5
Non-mechanized+modern variety+high input use+with credit+with insurance (switch*)
6
Mechanized+traditional variety+low input use+no credit+no insurance
7
Mechanized+traditional variety+high input use+no credit+no insurance
8
Mechanized+modern variety+high input use+no credit+no insurance
9
Mechanized+modern variety+high input use+no credit+with insurance
10
Mechanized+modern variety+high input use+with credit+with insurance (switch)
11
*Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation, modern
variety, high input) to the base technology category (i.e., non-mechanized land preparation, traditional variety, low input). The base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

46

13

Table 5. Acronyms of choices (i.e., c=1, …,11) resulting from a combination of technology categories and use of credit and/or
insurance that will be compared on the metrics certainty equivalent and insolvency percentage, within each village.
Technology Category/credit and insurance

Acronym

Non-mechanized+traditional variety+low input use+no credit+no insurance
Non-Mech+TV+LoIn+NCr+NIn
Non-mechanized+traditional variety+high input use+no credit+no insurance
Non-Mech+TV+HiIn+NCr+NIn
Non-mechanized+modern variety+low input use+no credit+no insurance
Non-Mech+MV+LoIn+NCr+NIn
Non-mechanized+modern variety+high input use+no credit+no insurance
Non-Mech+MV+HiIn+NCr+NIn
Non-mechanized+modern variety+high input use+no credit+with insurance
Non-Mech+MV+HiIn+NCr+WIn
Non-mechanized+modern variety+high input use+with credit+with insurance (switch*)
Non-Mech+MV+HiIn+WCr+WIn
Mechanized+traditional variety+low input use+no credit+no insurance
Mech+TV+LoIn+NCr+NIn
Mechanized+traditional variety+high input use+no credit+no insurance
Mech+TV+HiIn+NCr+NIn
Mechanized+modern variety+high input use+no credit+no insurance
Mech+MV+HiIn+NCr+NIn
Mechanized+modern variety+high input use+no credit+with insurance
Mech+MV+HiIn+NCr+WIn
Mechanized+modern variety+high input use+with credit+with insurance (switch)
Mech+MV+HiIn+WCr+WIn
*Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation, modern
variety, high input) to the base technology category (i.e., non-mechanized land preparation, traditional variety, low input). The base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

13

These acronyms are for technology categories and use of credit and insurance. The first section is for type of land preparation,
Non-Mechanized or Mechanized. The second section is for the type of variety used: traditional or modern variety. The third section
is for the intensity of input use: low input or high input used. The fourth section is for the use of credit: no credit or with credit. The
fifth section is for the use of insurance: no insurance or with insurance.
47

Table 6 shows the relevant comparisons of the certainty equivalents (from Table 4), to
assess if the credit-insurance package is beneficial. Other comparisons (i.e., gains from
additional inputs) are discussed in the end of this chapter.
Table 6. Proper certainty equivalent comparisons of choices for each of the cases of interest
where a credit-insurance package could potentially be beneficial.
Case
Proper Comparison
Risk averse and not capital constrained
Non-Mech+TV+LoIn+NCr+NIn
vs.
Non-Mech+MV+HiIn+NCr+WIn
(i.e., choice 1 vs. choice 5 in Table 4) and
Mech+TV+LoIn+NCr+NIn
vs.
Mech+MV+HiIn+NCr+WIn
(i.e., choice 7 vs. choice 10 in Table 4)
Non-Mech+TV+LoIn+NCr+NIn
vs.
Non-Mech+MV+HiIn+WCr+WIn
(i.e., choice 1 vs. choice 6 in Table 4) and

Risk averse and capital constrained

Mech+TV+LoIn+NCr+Nin
vs.
Mech+MV+HiIn+WCr+WIn
(i.e., choice 7 vs. choice 11 in Table 4)

The base technology category is the use of a traditional variety with low input (i.e., low
fertilizer) for each type of land preparation system. Nevertheless, when comparing each choice,
only income and cost change-- the scenario (i.e., village parameters, and technology parameters
besides those related to income and cost) remains the same.
The certainty equivalent will provide the lower but guaranteed amount of wealth that will
make the farmer indifferent to taking that lower but guaranteed amount compared to taking a
higher expected but uncertain wealth.
Another element to note is the switch to the base technology category which is related to
what was previously named quasi-fixed choice. In two cases, the choices in the simulation are
48

not completely fixed through time: 1) if the farmer becomes insolvent and therefore stops
farming to do off-farm activities and earn the amount equal to the fixed consumption, and 2)
when a capital-constrained farmer becomes ineligible for credit, which is needed to buy modern
seed and inputs (i.e., modern seed package), is forced to switch from the modern variety package
to the base technology category (i.e., traditional low input). That is, the initial choice, for the
farmers who with credit are able to buy the modern variety package in a specific year, is the
modern variety package, but the choice is changed to the base category because of credit
ineligibility.
Marginal wealth from in each period is the difference in ending wealth between periods.
It is also each year´s gross income, defined here as revenue minus variable costs. The only
source of variation here is in wealth due to chances in production and not consumption. Gross
incomes are used instead of ending period wealth to avoid double counting. Marginal wealth can
be defined as follows in Equation 2.:

MWitcv  I itcv  Eitcv

Equation 2.

cv

where I it=income from draw i, at time t, for choice c and village v.
cv

E it= expenses from draw i, at time t, for choice c and village v.
It is necessary to clarify that the only source of randomness in income is income from
crops (i.e., stochastic yields and prices for beans and corn). Consequently, since farmers
expenses depend on yield (i.e., harvesting costs), expenses are also stochastic; although it is also
the only source of randomness in this variable.
Income from draw i, at time t, for choice c and village v is defined as follows in Equation
3:

49

Iitcv  GI itmv  GI itcv  GIOS v
Where GI

Equation 3.

mv
it =is corn gross income from draw i, time t, and village v.

cv

GI it = bean gross income from draw i, time t, and village v, and choice of bean
technology c, c=1,…,14.
GIOS= gross income from other sources (i.e., off-farm income, net remittances, and
income from other crops, cattle, or business ) in village v.
Expenses from draw i, at time t, for choice c and village v is defined as follows in
Equation 4.:

Eitcv  Citm  Citc  HHE

Equation 4.

m

Where C it= corn cash costs (i.e., family labor and land is not accounted for) for draw i,
at time t.
c

C it= bean cash costs (i.e., family labor and land is not accounted for) for draw i, at time t,
choice of bean technology c, c=1,…,14.
HHE= household expenses (i.e., food, clothing, education).
Specifics of corn and bean income and costs are specified in the “@Risk general setup”
section. However, simulations in this study are modeled as a steady state. That is, prices and
yields have the same mean through the 20 year period. It also does not facilitate investment,
which means that growth is not incorporated into the model.
3.4. @Risk General Setup
The general @Risk setup is composed of random variables, parameters (constants) and a
correlation matrix that correlate each of the random variables with each other. The random

50

variables are a function of the estimated or assumed parameters. The actual values of the
parameters, statistics of variables, and correlations are presented in Chapter 4.
The random variables introduced in @Risk are prices and yields for corn and beans,
where distributions are conditional on choices and villages. All other elements in equations 1
through 4 are estimated parameters for each of the choices and villages. The correlation matrix
contains the pairwise correlations between all random variables (i.e., prices and yields).
Prices and yields for corn and beans in @Risk are defined by probability distribution
parameters established for each choice and village. The established yield parameters initially
would have come entirely from farmers’ elicitation of yield distributions, but as will be
explained in Chapter 4, these were rescaled taking into consideration expert opinion. The expert
opinion helped shape the scale and variance, of yield distributions for each of the choices and
villages. Price distributions come from secondary data (i.e., price coefficient of variation comes
from SIMPAH prices) adjusted to primary data characteristics (i.e., mean price of the village
from farmer survey).
Equation 1 in @Risk yields the expected utility for a certain choice and scenario within a
certain village. The inverse of the utility function is applied at the expected utility value to
obtain the certainty equivalent and then compare the choices. Certainty equivalent is used as a
metric of comparison instead of expected utility because it is in currency units and not in utils,
the unit of measure of utility, and provides a pragmatic version of the outcomes.
Marginal wealth values for choices with crop insurance for the bean enterprise include an
indemnity payment and an insurance premium. There is one insurance premium for each
technology category or choice with insurance in each village. The indemnity payment depends

51

on the loss function of the crop insurance contract. For this study, a yield contract is used;
therefore, the indemnity price multiplies the loss function to estimate the indemnity.
3.4.1. Utility function
The utility function used is an additive utility function with respect to marginal wealth
with a constant relative risk aversion coefficient. One of the challenges faced with incomeconsumption and utility models in general, is that liquidity and other factors are confounded with
the risk aversion effect. To mitigate this problem, several coefficients of risk aversion are
included in the model.
3.4.2. Quasi-fixed choices
Certain flexibility was given to the model to imitate certain common choices made by farmers
when they face certain situations. For example, when farmers reach a certain low threshold of
ending period wealth (insolvency), they would stop farming and obtain income from off-farm
work. The level of insolvency is established at 1.25 times the value of the house. Similar
thresholds were assumed to control when to obtain or stop obtaining credit and insurance. The
threshold that would make a farmer not eligible for credit and therefore need to switch back to
plan a traditional variety with low inputs is 1.75 times the value of the house. In Chapter 4, the
thresholds are described for each choice and village.
Since this study is using simulation instead of dynamic programming the need for credit
is assumed from the start of a choice in the simulation (i.e., Non-Mech+MV+HiIn+WCr+WIn)
and kept until the farmer is no longer eligible for credit, instead of obtaining credit only when it
is needed like it is done in dynamic programming models.

52

3.4.3. Monte Carlo simulation
To obtain farmers income probability distributions, a Monte Carlo simulation procedure
was used. Running the Monte Carlo simulation procedure requires estimates of the probability
distribution parameters of corn and bean yields and prices to estimate the probability distribution
of revenue. The income probability distributions are then used to obtain the cumulative
distribution function of income with and without insurance, and thus, order the uncertain choices.
The increasing cdf format is presented because it is easy to observe the probability of observing
an outcome below a certain value, which is related to the purpose of this study of reducing
downside risk though crop insurance. Also, it is easier to observe if one cdf stochastically
dominates another cdf.
3.4.4. Random variables, correlation matrix and parameters
The bean and corn yields were obtained through a mix of elicitation of farmer subjective
yield distributions, secondary data, and expert opinion due to problems in the elicitation results.
The visual impact method described in Chapter 2 was used to elicit the yield distributions.
Corn and bean price probability distributions are assumed to be lognormal; parameters
were estimated from secondary price data and adjusted for each village and technology (i.e.,
variety type) choices (i.e., due to price differentials in bean variety prices).
A correlation matrix is specified by using secondary time-series data for bean and corn
yields and corn and bean prices to obtain the off diagonal terms. This is estimated by obtaining
the correlations of the residuals of the regressions of prices and yields on time. The regression
on time is required to detrend the regression and obtain the actual correlations.
The parameters used in the models are the following: farmers’ discount rate and
intertemporal time preference, initial wealth, corn and bean area planted, household expenses,

53

investment according to insurance, insurance premium, bean indemnity price, insurance
coverage, premium loading factor, off-farm income, and income of other activities.
Cost has a fixed part and a random part. The fixed component was estimated from the
survey data and presented in the parameters section as base fixed cost. The randomness is due to
harvesting costs that depends on the stochastic yield.
3.4.4.1.

Elicitation of subjective probabilities

The elicitation of subjective probabilities is the source to obtain the yield probability
distributions of beans and corn. Other sources of randomness in the simulation model are the bean and
corn prices, and bean and corn costs because bean and corn costs (i.e., harvesting costs) are dependent
on yields.
The visual impact method was used to estimate the probability distribution, and within
itself, the distribution parameters (i.e., the means and variances).
3.4.4.1.1.

Visual Impact Method

The visual impact method was implemented by having an enumerator ask the farmers to
indicate their maximum and minimum corn and bean yields, independently. Then, the
enumerator divided the range into five intervals and asked the farmer to distribute 25 counters
among the intervals--according to the farmers’ expectation regarding the occurrence of each
interval (Table 7 and 8). These counters, along the intervals, are used to estimate the yield
distribution parameters (i.e., means and variances).
Tables 7 and 8 show examples of the elicitation of probability yield distributions of bean
and corn using the visual impact method. Table 7 shows that the minimum yield stated by the
farmer is zero hundred weight (cw) and the maximum yield is 27 cw. The range is divided by
five, the number of intervals, and yields 5.4. The decimal part of the ratio (i.e., 0.4*5) was

54

distributed to the lowest and highest intervals, and left the rounded value (i.e., 5 cw) to the
middle intervals. Thus, a width of 5 cw was assigned to the middle three intervals and 6 cw to
the lowest and highest interval. The same procedure was followed for corn in Table 8 to assign
the limits of the intervals.
Table 7. Example of a bean yield distribution, using the visual impact method.
if Min 0 cwt/ha ,
(27-0)/5 = 5.4 cwt for each interval; or 5 for each middle and 6
Max 27 cwt/ha.
for extremes
Range in Yield
Counters
Cumulative counters
0≤x<6
2
2
6 ≤ x < 11
4
6
11≤ x < 16
13
19
16 ≤ x < 21
4
23
21 ≤ x ≤27
2
25
Total
25
1 cwt= 100 pounds
Table 8. Example of a corn yield distribution, using the visual impact method.
if Min 0 cwt/ha ,
(64-0)/5 = 12.8 cwt for each interval: or 12 for each middle and 14
Max 64 cwt/ha.
for extremes
Range in Yield
Counters
Cumulative counters
0 ≤ x < 14
2
2
14 ≤ x < 26
4
6
26 ≤ x < 38
13
19
38 ≤ x < 50
4
23
50 ≤ x < 64
2
25
Total
25
1 cwt= 100 pounds
To obtain expected utility of income, both a particular risk aversion function and a
coefficient of risk aversion are assumed

14

to represent a moderate and a high risk averse farmer.

14

One of the challenges faced with several income-consumption models is that liquidity and
other factors are confounded with the risk aversion effect. To mitigate this problem, several
coefficients of risk aversion and liquidity variables are included in the model
55

3.4.4.2.

Insurance and credit

The objective of using a simulation model is to compare bean farmers’ expected utility of
income with and without insurance and credit. Additionally, different insurance contract
parameters are used to evaluate the current insurance contracts and provide insights regarding
alternative contracts that could be offered by insurance companies.
3.4.4.3.

Insurance contract parameters

To represent the yield insurance contract for the bean enterprise used in the model,
several insurance contract parameters need to be established. The insurance contract parameters
are the following: bean indemnity price, bean expected yield, coverage, loading factor, and
premium. The model does not include any insurance contract for the corn enterprise.
The bean indemnity price established by insurance companies in Honduras was
used. The indemnity price is 89% of the expected harvesting period price. Two coverages and
two loading factors were used to evaluate premiums paid in Honduras. The first coverage (75%)
15

is from a standard yield insurance contract in the US . The other is based on the coverage
offered to bean farmers in Honduras (i.e., 45%).
Premiums used in the model depend on the choices presented previously, as well as on
the village. The estimation of premium is shown in the next section.
3.4.4.3.1.

Premium evaluation

Honduran contracts will be evaluated according to the premiums and coverages.
Additionally, the premiums will be evaluated by estimating the perceived actuarially fair
premium and compare it to the actual premium.

15

This coverage was selected to have a normative example that could dictate the goal for
insurance contracts in Honduras.
56

To obtain the perceived actuarially fair premium to the farmer, the following loss
equation will be used:

E[ IND]  E[ Loss ] * IP  E[max[(YG  Y )(1  copay),0]] * IP
The actual premium was obtained from key informant interviews with insurance
companies’ staff.
3.5. Data Collection
The sources of data are from key informant interviews, a farmer survey, and secondary
data.
3.5.1. Key informant interviews
Key informant interviews were done with staff of insurance companies, banks and other
financial institutions, and scientists. As will be explained in Chapter 4, due to the problems
encountered in the data collection from the survey, many parameters were obtained from key
informant interviews. The parameters that were obtained from key informant interviews are the
following: yield differences between villages for corn and beans, bean yield response to fertilizer,
16

bean minimum yields , interests paid on loans (i.e., base for discount rate used in the model),
and coverage, premiums, and liability of Honduran bean insurance contracts.
3.5.1.1.

Insurance companies

Staff of insurance companies were contacted to obtain information on the contracts
offered at the study sites (e.g., the trigger mechanisms, the premiums) to learn about the policies
and existing arrangements between the insurance company and farmers. Staff from all six crop
insurance companies in Honduras, which are either licensed or on their way to be licensed, were

16

Bean minimum yields are the bean yields expected in worst case scenario in the village of
interest.
57

contacted and interviewed. These companies are the following: (1) ProAgro, (2) Seguros
Interamericana, (3) Seguros HSBC, (4) Seguros Atlantida, (5) Seguros Equidad, and (6) Seguros
Continental.
3.5.1.2.

Banks and other financial institutions

Key informants from banks and micro-credit agencies that service the region were
interviewed to learn about the conditions and services rendered, and to determine their interest in
creating a credit-insurance package and what would the conditions be under which they would
offer such a product.

The staff from the following banks was contacted: Banco Atlantida,

Banco de Occidente, BAMER, HSBC, BANADESA, BAC, and Banco Azteca. BAC and Banco
Azteca did not provide credit to agriculture, so their staff was not interviewed. Also, a staff
member from one of the most influential microcredit organizations in Honduras, ODEF, was
interviewed.
3.5.1.3.

Scientists

Scientists from Zamorano were interviewed to obtain information on farmers’ bean yields,
practices, and modern variety adoption. The information collected from the insurance companies,
banks, and scientists are reported in the following sections.
3.5.2. Farmer survey
Two hundred forty-five farmers were interviewed for this study. The sampling method
and farmer data is presented in the next sections.
3.5.2.1.

Farmer sample selection

Eight villages were selected where insurance companies in Honduras were selling bean
insurance policies and were perceived by researchers to have the most important types of risks
(e.g., drought, excess rain, and high temperatures) for bean farmers. An insurance company

58

provided a list of 34 villages in two of the most productive bean departments in Honduras,
Francisco Morazan and El Paraiso. Each of these villages had at least 20 insured farmers. The
17

villages were then categorized by elevation (< 500, 500-1000, and >1,000 m.a.s.l. ) and by
18

precipitation (< 300, 300-600, and >600 mm ). The villages were randomly selected within
each stratum. One village which had very similar characteristics to a village already chosen was
dropped and replaced with one with very different characteristics to obtain a wide range in
elevation and precipitation. One village was selected randomly within the low and high
19

elevation strata, and six villages were selected from the middle elevation strata . The selected
villages are the following: San Pedro Alauca, Chirinos, Arauli, La Cienega, Guaimaca, Talanga,
Villa de San Francisco, and Sabaneta de Valle de Angeles.
In these villages, a list of farmers who had most likely not purchased crop insurance was
added to the list of farmers who had. Fifteen farmers who had bought crop insurance and 15
who had most likely not were selected for the sample. Later in the field, it was realized that
many farmers were afraid to admit they had bought crop insurance because this was a requisite
for a loan they had obtained and many of them were delinquent. Thus, the final sample did not
balance as previously planned.
3.5.2.2.

Farmer data collection

The survey did not collect time series data on farmers’ yields and prices for beans nor
corn. Rather, it elicited from farmers their subjective probabilities regarding yields; and
collected data for the previous year on off-farm income (e.g., paid labor, net remittances),

17
18
19

m.a.s.l. meters above sea level.
m.m. millimeters of rainfall.
This study is a case study, so the arbitrary selection is not of importance.
59

household expenditures (i.e., food, clothing, education), use of varieties (bean and corn), use and
cost of inputs and labor, use of credit, and general socioeconomic characteristics (household (hh)
size, education and age of hh head, assets (land owned, machinery, draft animals, range animals).
The following procedure was performed to obtain an approximate sample size to use.
First, a standard deviation of the yield variable, which is of most interest, is needed. Thus, the
cross farmer and village coefficient of variation for modern and traditional varieties for the
Postrera season (Table 9) was obtained from Mather (2003). The sample size formula derived
from the margin of error definition is the following:
2

N = (tα/2* σ/E) , where t is the value from the t-distribution at a significance level α (e.g.,
5%), multiplied by the standard deviation, σ, divided by the margin of error.
Table 9. Mean bean yields and coefficient of variation for Postrera season 1999 and 2000,
Honduras.
Postrera 2000
Postrera 1999
Yield (kg/ha)
CV (%)
Yield (kg/ha)
Variety
537
93
791
Modern
446
80
612
Traditional
CV= coefficient of variation (standard deviation divided by the mean).
Source: Mather (2001) p.33

CV (%)
53
63

The margin of error used to calculate the sample size is the amount of kg of beans (134
kg) that is equal in value to the extra cost of fertilizer and extra cost of seed needed to change
from a traditional variety technology package to a modern variety technology package.
Assuming the cost per hectare of switching from a traditional variety package to a modern
variety package is approximately the cost of 90 kg of fertilizer (L. 993), plus the extra cost of
seed (i.e., amortized by 3 planting periods) (L. 348). The cost of L. 1,341translated to kilograms
of dry beans is 134 kg. The following total sample size is obtained:
2

N = (1.96 * 500/134) = 53
60

The coefficient of variation in Mather’s (2003) is not corrected for village effect so the
sample size for each village should be lower. Due to research budget allowing and desire to
include a wide range of places, eight villages were surveyed with approximately 30 farmers per
village were surveyed.
3.5.2.3.

Farmer data

A farmer survey was used to collect data to estimate income and expenses, farmers’
capital, and socioeconomic characteristics (i.e., age, education). Selling and consumption prices
were valued equally.
3.5.2.3.1.

Income

The elicited yield distributions were obtained as well as price data on bean and corn sold.
Additionally, other sources of income were collected (i.e., net remittances, off-farm income, and
income from other activities).
3.5.2.3.2.

Capital

To obtain the farmers level of wealth, data on house value, land value, machinery, and
animals owned were collected.
3.5.2.3.3.

Costs

The production costs (i.e., input costs, labor costs, mechanization costs, harvesting costs),
and marketing costs (i.e., transportation costs) were obtained by directly surveying farmers
regarding the amounts and prices of inputs used. Details on how production cost data were
obtained are explained below.
Input expenditures were obtained by multiplying quantities used, and prices paid by the
farmers. These quantities and prices were collected using the survey instrument.

61

Labor-days per activity and type of labor were obtained through the survey instrument.
Labor days were categorized into family labor and hired labor. Women and men family labordays were valued equivalently, as opposed to valuing women labor-days less than a man’s laborday as it is done in some studies.
The tractor activities were created in “hours” units instead of “days” units. The total
number of hours worked for the tractor activities was calculated by multiplying the number of
tractors times the number of hours worked. Similar to the labor-days, two general categories
were created for the unlikely event when two or more tractors worked different number of hours;
one for hired tractors and another for owned tractors. The total number of hours was recorded in
each of these general categories.
Threshing is an activity that had two ways of calculating its cost. One way of calculating
the cost was using labor-days and the other was with tractor-hours, depending on which type of
threshing activity was done. For the threshing tractor hours, an estimate of the number of hours
per unit of harvest weight was estimated; these hours per unit of weight were then multiplied by
the total harvest weight to obtain the total number of tractor hours.
Since some farmers prepared land and planted on the same day, these two activities are
reported jointly and summarized in one category named “preparation and planting.”
Similar to changes made to the inputs’ physical quantity outliers, labor-day outliers were
also replaced with frontier values. Frontier values were limit values set that a farmer would
unlikely surpass, as considered by the author and key informants.
Once the labor-day quantities for each activity were calculated, the labor expense was
calculated by multiplying the wage (reported by the farmer) times the labor-days.

62

The simulation requires households’ cash flow per year. Since family labor is, generally,
a non-cash expense, its expense was subtracted from the total labor expense. The difference was
named “residual labor expense,” and was the one used in the simulation.
3.5.3. Secondary data
Secondary data was obtained from the Honduran Market Information System for
Agricultural Products (SIMPAH) and the National Statistics Institute (INE). SIMPAH provided
price historical data for beans and corn, while INE provided historical data on bean and corn
national yields. These data on prices and yields were used to estimate correlation coefficients
between them.
3.6. Factors to Compare
Rather than directly testing hypotheses (i.e., value and direction of parameters estimated as
it is done with regression), this analysis simulates farmers’ actions to see if they are sensible,
given their circumstances. In other words, are farmers’ actions consistent with the hypotheses
that there is no difference in absolute value of simulated outcomes derived from different farmers’
actions, compared to the usual positive testing (i.e., there is no statistical difference in observed
outcomes derived from different farmers’ actions). Farmers’ circumstances and environments
are simulated and compared in these environments to match the principal factors affecting bean
farmers’ yields (e.g., rainfall, elevation, disease), perils (e.g., drought, excess rain), technology
(e.g., low, high). However, due to a small sample size or because some combinations do not rise
naturally (i.e., high elevation, drought and high technology levels), only some of the scenarios
were compared. Thus, this study is a case study--it does not provide results that can be
inferred/generalized to the population. Rather, the different factors (i.e., elevation, rainfall) that

63

lead to the various farmer environments became subject of target when selecting the villages for
the study.
As the literature on technology adoption cites, the lumpiness of a technology packages (i.e.,
many technology changes in one package instead of one small technology change at a time) is
sometimes the reason why farmers do not adopt a technology. For that reason, this study makes
comparisons of choices when one item is changed at a time, as well as when the whole
technology package is adopted. More specifically, it compares the farmers’ certainty equivalent
from a base choice to when just the type of variety is changed, or if only a higher amount of
inputs is used. Moreover, it also compares that base choice to when all changes simultaneously:
type of variety and amount of input use (land preparation was kept constant).
To complement the metric of certainty equivalent, stochastic dominance criterion is
applied to order choices. Stochastic dominance is used as a method to order uncertain choices to
be able to communicate more efficiently with other disciplines.
3.7. Certainty Equivalents Comparison of Technology Choices
The initial hypothesis stated that small-scale farmers do not plant modern varieties due to
the perception of these varieties risk and also because some farmers were unable to buy the seed
and other inputs without credit. Thus, there are two types of constraints to test indirectly. One is
a hard constraint (i.e., farmer cannot do an action because of the constraint) where we assume
farmers need cash to acquire inputs; and the other constraint (i.e., farmer is unwilling to perform
an action) is to observe if farmers are driven to use traditional varieties by risk aversion and fear
of low yield events. To evaluate these assertions, several comparisons of certainty equivalents
are made according to the two cases of interest in this research: risk averse and not capital

64

constrained and risk averse and capital constrained, which are discussed in the following two
sections. Additionally, gains from additional inputs are presented in the last section.
3.7.1. Certainty equivalent comparison for risk averse and not capital constrained
To observe if the credit-insurance package is beneficial for a risk averse and not capital
constrained farmer, the proper comparisons of certainty equivalents (CE) are the following: 1)
compare the CE between the choice Non-Mech+TV+LoIn+NCr+NIn (see Table 10 for acronym
codes) and the choice Non-Mech+MV+HiIn+NCr+WIn (i.e., compare CE from choices 1 and 5
from Table 4), given that the farmer prepares land with only labor; 2) compare the CE between
the choice Mech+TV+LoIn+NCr+NIn and the choice Mech+MV+HiIn+NCr+WIn (i.e.,
compare CE from choices 7 and 10 from Table 4), given that the farmer prepares land with
mechanization.
Table 10. Acronyms of choices related to technology packages to compare for a risk averse
and not capital constrained farmer.
Acronym
Term
Non-Mech+TV+LoIn+NCr+NIn
Non-mechanized+traditional variety+low input use+no
credit+no insurance
Non-Mech+MV+HiIn+NCr+WIn
Non-mechanized+modern variety+high input use+no
credit+with insurance
Mech+TV+LoIn+NCr+NIn
Mechanized+traditional variety+low input use+no
credit+no insurance
Mech+MV+HiIn+NCr+WIn
Mechanized+modern variety+high input use+no
credit+with insurance
For a farmer with non-mechanized land preparation and mechanized land preparation,
the comparison between choices Non-Mech+TV+LoIn+NCr+NIn vs. NonMech+MV+HiIn+NCr+WIn , and Mech+TV+LoIn+NCr+NIn vs. Mech+MV+HiIn+NCr+WIn ,
respectively, are appropriate. They are appropriate because a risk averse farmer might not want
to risk a lower gross income from a modern variety package in case of a bad event, thus the
farmer uses the traditional variety package (e.g., traditional variety and low fertilizer). However,

65

by purchasing crop insurance the farmer potentially would be willing to buy the complete
modern variety package (e.g., modern seed and high fertilizer). Even with crop insurance it is
not certain that the farmer will prefer the modern variety package because it depends on the
farmers’ level of risk aversion, insurance contract parameters, and gross income parameters from
each of the technology packages.
3.7.2. Certainty equivalent comparison for risk averse and capital constrained
The proper comparisons of certainty equivalents (CE) to observe if the credit-insurance
package is beneficial for a risk averse and capital constrained farmer are the following: 1)
compare the CE between the choice Non-Mech+TV+LoIn+NCr+NIn (see Table 11 for acronym
codes) and the choice Non-Mech+MV+HiIn+WCr+WIn (i.e., compare CE from choices 1 and 6
from Table 4), given that the farmer prepares land with only labor (i.e., non-mechanized); 2)
compare the CE between the choice Mech+TV+LoIn+NCr+NIn and the choice
Mech+MV+HiIn+WCr+WIn (i.e., compare CE from choices 7 and 11 from Table 4), given that
the farmer prepares land with mechanization.
Table 11. Acronyms of choices related to technology packages to compare for a risk averse
and capital constrained farmer.
Acronym
Term
NonNon-mechanized+traditional variety+low input use+no
Mech+TV+LoIn+NCr+NIn
credit+no insurance
NonNon-mechanized+modern variety+high input use+with
Mech+MV+HiIn+WCr+WIn
credit+with insurance
Mech+TV+LoIn+NCr+NIn
Mechanized+traditional variety+low input use+no credit+no
insurance
Mech+MV+HiIn+WCr+WIn
Mechanized+modern variety+high input use+with
credit+with insurance
For a farmer with non-mechanized land preparation and mechanized land preparation,
the comparison between choices Non-Mech+TV+LoIn+NCr+NIn vs. NonMech+MV+HiIn+WCr+WIn , and Mech+TV+LoIn+NCr+NIn vs.

66

Mech+MV+HiIn+WCr+WIn , respectively, are appropriate. They are appropriate because a
capital constrained risk averse farmer is unable to purchase the modern variety package, and if
credit would be made available, even then the farmer might not want to use the modern variety
package because of the risk a lower gross income and not being able to repay the loan in case of
a bad event. Each technology category is assumed to last for the whole 20 years except in the
case where ineligibility for credit is reached, in which case the farmer returns to the base
category. The ineligibility threshold is valued at 1.75 times the house value. Base category for
non-mechanized technologies is non-mechanized traditional low input, and for mechanized
technologies is mechanized traditional low input. In addition, if insolvency is reached after
having switched to the base category, it is assumed that the farmer keeps working off-farm and
stays at the insolvency threshold for the remaining years. The insolvency threshold is valued at
1.25 times the house value.
In summary, without the credit-insurance package some farmers would probably select
the traditional variety package and with the credit-insurance package, the farmer might use the
modern variety package. Again, it is not certain that the farmer will prefer the modern variety
package even with the credit-insurance package because it depends on the farmers’ level of risk
aversion, credit and insurance contract parameters, and gross income parameters from each of
the technology packages. Thus, both credit and crop insurance need to be available to risk averse
and capital constrained farmers. Credit is needed for purchasing the modern seed and inputs, and
the crop insurance is needed to be able to repay the loan in case of a bad event.
3.7.3. Gains from additional inputs
Traditional and modern varieties respond to fertilizer. Similar to the comparison of a
modern variety to a traditional variety, using a high level of fertilizer produces a lower minimum

67

yield and lower gross income than a low fertilizer level when a bad event (i.e., drought) occurs.
Nevertheless, for both types of bean varieties a high level of fertilizer is expected to have a
higher gross benefit than a low level of fertilizer in. This tradeoff of having a higher net benefit
at the cost of a lower minimum net benefit (i.e., the expected minimum net benefit when a bad
event occurs) when applying high level of fertilizer, creates a similar response in risk averse
farmers like when comparing modern and traditional varieties. However, the effect of the levels
of fertilizer on risk averse farmers might not be of the same magnitude because the difference in
yield between minimum yields (i.e., the lowest expected yield when bad events occur) for the
low and high fertilizer level is lower than the difference between the minimum yields of the
modern and traditional varieties.
To observe the gains from additional inputs for a risk averse and not capital constrained
farmer, the proper comparisons of certainty equivalents (CE) are the following: 1) compare the
CE between the choice Non-Mech+TV+LoIn+NCr+NIn (see Table 12 for acronym codes) and
the choice Non-Mech+TV+HiIn+NCr+NIn (i.e., compare CE from choices 1 and 2 from Table
4), given that the farmer prepares land with only labor (i.e., non-mechanized); 2) compare the CE
between the choice Non-Mech+MV+LoIn+NCr+NIn and the choice NonMech+MV+HiIn+NCr+NIn (i.e., compare CE from choices 3 and 4 from Table 4), given that the
farmer prepares land with only labor; and 3) compare the CE between the choice
Mech+MV+LoIn+NCr+NIn and the choice Mech+MV+HiIn+NCr+NIn (i.e., compare CE from
choices 7 and 8 from Table 4), given that the farmer prepares land with mechanization.

68

Table 12. Acronyms of choices related to level of fertilizer to compare for a risk averse and
not capital constrained farmer.
Acronym
Term
Non-mechanized+traditional variety+low input use+no
Non-Mech+TV+LoIn+NCr+NIn
credit+no insurance
Non-Mech+TV+HiIn+NCr+NIn
Non-mechanized+traditional variety+high input use+no
credit+no insurance
Non-Mech+MV+LoIn+NCr+NIn
Non-mechanized+modern variety+low input use+no
credit+no insurance
Non-Mech+MV+HiIn+NCr+NIn
Non-mechanized+modern variety+high input use+no
credit+no insurance
Mechanized+traditional variety+low input use+no
Mech+TV+LoIn+NCr+NIn
credit+no insurance
Mech+TV+HiIn+NCr+NIn
Mechanized+traditional variety+high input use+no
credit+no insurance
The comparisons for fertilizer levels above are appropriate for farmers in the same region
because each of the three comparisons is a movement from a low fertilizer level to a high
fertilizer level, given a type of variety and land preparation system. However, this assumes that
the native fertility of the soil for farmers in the same region is similar.
3.8. Chapter Summary
The study uses a simulation model instead of dynamic programming because simulation
modeling allows a straight forward comparison of choices and villages. The Excel add-in
@Risk was used to estimate stochastic present value of wealth utility model (SPVWUM) to
generate a Montecarlo simulation process. SPVWUM integrates income and expenses to
generate the marginal wealth along the multiperiod model.
Yield and price variables, along with parameters of wealth and time preferences,
insurance contract characteristics, other sources of income, and household expenses were used in
the SPVWUM. The yield distributions used in the study were obtained through a mix of
elicitation of farmer subjective yield distributions, secondary data, and expert opinion due to
problems in the elicitation results, as will be explained in the next chapter. Yield and prices were

69

modeled as a steady state having the same means throughout the 20 year period of the simulation.
Growth was not incorporated into the model.
Certainty equivalent was the metric used for comparison between farmers choices.
Data were collected through key informant interviews with staff of insurance companies,
banks, and scientists and a farmer survey. Additionally, secondary data from INE and SIMPAH
was collected to obtain several simulation parameters.
Finally, cases of interest to determine if the credit-insurance package is beneficial were
presented along with the proper comparisons of the certainty equivalents. The cases of interest
are: 1) risk averse and not capital constrained farmers and 2) risk averse and capital constrained
farmers. Additionally to the cases of interest, several comparisons for gains from additional
inputs are presented.

70

4. CHAPTER IV: Data Analysis and Information Collection
The following chapter summarizes the information collected from key informant
interviews, survey data and secondary data collected mostly in 2007. The key informants
included personnel from insurance companies, banks, microcredit institutions and scientists. The
farmer survey was used to collect information to estimate key parameters used in the stochastic
present value of wealth utility model (SPVWUM). Finally, the chapter summarizes the
parameters used for the different runs of the model for final comparisons between choices.
4.1. Key Informant Interviews
4.1.1. Crop insurance history in Honduras
After hurricane Mitch in 1998, industrial firms like Demasah and Alianza, who usually
has financed farmers, saw the necessity to insure crops (Erazo, 2007; personal communication).
These firms had built relationships with firms in Mexico and asked ProAgro, the largest
insurance company in Mexico, to help design, establish, and develop insurance programs to
protect against extreme weather events (ProAgro website, 2007). ProAgro started insuring
farmers first in Guatemala and Honduras, and later included El Salvador and Nicaragua
(ProAgro website=http://www.proagroseguros.com.mx/historia.html).
The first insurance company to offer crop insurance in Honduras was Seguros El Ahorro
Hondureño along with ProAgro Honduras as a reinsurer. El Ahorro Hondureño is currently part
20

of HSBC .

20

El Ahorro Hondureño merged with Bancahsa (Bancahsa bought El Ahorro Hondureño) in
July 2000 (http://reseau.crdi.ca) to form a new bank called BGA (Banco Grupo El Ahorro
Hondureño). BGA was later bought by HSBC in July 2006 (http://www.hsbc.com).
71

4.1.2. Crop insurance rules and regulations
Crop insurance in Honduras is regulated by the Insurance Superintendence (IS), which is
a subdivision of the Superintendence of Banks and Insurance (SBI).
The IS works with funds provided by the Central Bank of Honduras as well as fees
charged to insurance companies, which come from a percentage of the premiums paid to the
insurance companies. Insurance companies obtain their insurance policy approved by presenting
to the IS a proposal of an insurance policy. The IS personnel revise the policies and suggest
and/or require changes in the policy. Finally, a resolution is made. The required time from a
submitted policy to when the submitter receives an answer is 30 days. Nevertheless, when new
products with its own complexities are submitted to the IS, the time required for actual resolution
is much greater than 30 days. An insurance company that does not abide by these rules and
regulations is sanctioned pecuniarily and/or legally.
Crop insurance in Honduras has been compulsory from 2002 to the present (except for
one year) at the Honduran development bank, BANADESA. It has remained at private banks’
discretion to make it compulsory or not for their clients.
All crop insurance contracts offered in Honduras do not take quality of produce into
account. Quantity is the only factor included for indemnization.
A crucial part of the indemnization is the yield adjustment. Yield adjustment is done by a
contractor of the crop insurance company, which might lead to a conflict of interests. In any
event, a farmer can formally complain about a yield adjustment done by an insurance company
contractor by filing a complaint to the Chamber of Commerce and either conciliate or go to
arbitration. In the conciliation, the insured and insurer would try to come to an agreement but it
has no legal standing. The unsettled differences would be determined by the arbitration. If the

72

parties go to arbitration, disagreements will be settled by a committee of arbitrors, which are
uneven in number. The decision taken by the arbitration committee is usually final. In rare
circumstances, there is nullity recourse or reposition recourse.
4.1.3. Insurance companies
The Insurance Superintendence has extended crop insurance permits to the following
crop insurance companies: (1) Seguros Atlantida, (2) Seguros HSBC and (3) Interamericana de
Seguros (currently, Seguros Ficohsa). Nevertheless, Seguros HSBC and Interamericana de
Seguros did not insure directly but operated as a fronting company for ProAgro, a Mexican
insurance company. Seguros Equidad was about to get licensed and Seguros Continental was
about to start the process to obtain a license in October 2006.
4.1.3.1.

ProAgro

ProAgro is a firm operating in Honduras, which does not operate in Honduras as an
insurance company per se, but it acts as an insurance agent, while HSBC does the fronting for
them. Fronting is the process by which these insurance companies do all the legal insurance
paperwork for a fee, but transfer all the risk to the reinsurer ProAgro. Nevertheless, ProAgro is
in charge of advertising, setting insurance premiums, assessing risk, and explaining the contracts
to clients. Interamericana de Seguros has an arrangement with ProAgro similar to HSBC, except
that risk is shared.
ProAgro in Honduras started insuring corn and sorghum farmers in 2001. In 2002, crop
insurance was promoted together with the “Monedero Agricola.” The Monedero Agricola was a
program that used to extend credit to farmers that was implemented from 2002-2006. Credit was
extended by BANADESA by issuing credit cards, where the products that the farmer could buy
were limited to agricultural inputs.

73

ProAgro offers five types of contracts: (1) Investment Cost Coverage with Yield
Adjustment (ICCYA) (2) Yield Insurance with Production Cost Adjustment (YIPCA), (3) Tree
Crop Insurance (4) Direct Damage Scheme, and (5) Compensatory Insurance Scheme. The
insurance contract offered to bean growers is the ICCYA . The trigger in this type of contract is
given by three distinct and separate steps: (1) the peril insured for needs to occur, (2) notification
from the insured to the insurer about a loss, and (3) damages to the crop need to be related to the
peril.
The indemnity formula for the contract of ICCYA is the following:

E[ IND]  E[ Loss ] * IP  E[max[(YG  Y )(1  copay),0]] * IP
The indemnity formula is applied once the steps for the trigger are fulfilled.
ProAgro, working through its fronting collaborators, charge bean insurance premiums
from 8% to 10% of the insured amount depending on the amount of land insured and region of
the country. Insurance price

21

for a kg of bean have ranged from L. 8.8/kg to L. 9.9/kg. The

copay schedule for different perils or mix of perils insured by ProAgro is shown in Table 13.
Table 13. Copay (% of loss) schedule for type of peril insured.
Type or mix of perils insured
Wind
Excess rain
Flood
Low temperature
Wind and excess rain
Hail
Drought
Drought, excess rain and wind
Source: Pejuan Bean Key Informant Interview, 2007.

21

Percentage
10
10
15
15
10-15
20
30
30

Insurance price is the price at which production is value for indemnity purposes; also called
indemnity price.
74

Yield adjustors are employees of ProAgro, but since ProAgro is not the company
extending the contract, it acts as an independent company exercising the yield adjustment.
ProAgro insures individual bean farmers in a region that farm less than 100 km away
from the insurer headquarter, if there are at least 7 ha to be insured. If the place is farther than
100 km and less than 200 km away, ProAgro would insure individual farmers, if the farmer
insures at least 14 ha. For a collective policy (i.e., a policy with more than one beneficiary),
ProAgro insures farmers with at least 0.35 ha, but where the sum of the hectares is greater than
21 ha, as long as they are located less than 200 km away. Nevertheless, if a farmer wanting to
insure a crop with a land area no less than 0.7 ha and it is near a place where there are already
other insured farmers, ProAgro would sell the contract to the farmer.
ProAgro’s crop insurance contractual procedure, done through its fronting agents, is the
following: (1) filling of application form, (2) risk assessment, (3) risk acceptance, (4) quote, (5)
subscription visit, (6) policy emission, (7) payment of premium, (8) follow-up visit, (9) alert of
peril, (10) yield adjustment, and (11) indemnization.
ProAgro also extends contracts to groups of farmers, which works the same way as the
individual contract where the insurer only deals with one person for payment purposes, except
that claims and indemnizations are done individually by each member in of the group.
4.1.3.2.

HSBC

Seguros HSBC does fronting for ProAgro. The only function of Seguros HSBC is to do
the paperwork for ProAgro. This means that even though farmers sign the contract emitted by
Seguros HSBC, it is ProAgro who assesses the risk and establishes the contract parameters.
Farmers need to notify ProAgro for any loss due to a peril instead of Seguros HSBC.

75

4.1.3.3.

Seguros Interamericana

Seguros Interamericana

22

is the other company doing fronting for ProAgro. This

company made available crop insurance policies in 2005. Nevertheless, it had not sold any bean
insurance policy up until October 2007, at the time of the interview. Seguros Interamericana
offers two types of contracts: (1) ICCYA, and (2) YIPCA. The difference between the relation
of HSBC and ProAgro, and Seguros Interamericana and ProAgro, is that Seguros Interamericana
does share some risk with ProAgro, while HSBC does not bear any risk; and farmers can notify
either Seguros Interamericana or ProAgro about a loss due to a peril. By October 2007, Seguros
Interamericana had not sold any group policies, but were considering it.
4.1.3.4.

Seguros Equidad

In October of 2006, Seguros Equidad was on the approval process to become a certified
crop insurance company. It had not offered any crop insurance policy, but instead had trials in
farmers’ fields. Seguros Equidad offered weather-based index crop insurance policies for one
time only in 2008. The company had the support of the World Bank, which had the objective to
include small-scale and medium-scale farmers. Seguros Equidad has been the only insurance
company who has offered weather-based index crop insurance. Seguros Equidad made alliances
in 2008 with meteorological stations to obtain and publicize, in the period of the contract, the
index to farmers, which would include the trigger of the contract. The first crop covered as a
tryout by Seguros Equidad was corn, but they are planning to expand to other crops. The
insurance contract to be offered by Seguros Equidad is that of production cost coverage.
Premiums and insurance price were not established at the time of the interview. The insurance

22

Seguros Interamericana is now Seguros Ficohsa.
76

contract was planned to be for individual or groups of farmers. The purpose of offering crop
insurance is to insure those farmers who request credit, and the indemnization would be paid
directly to the financial institution instead of to the farmer.
4.1.3.5.

Seguros Atlantida

Seguros Atlantida started designing crop insurance contracts in 2002 and started
operations in 2003. Seguros Atlantida offers three types of contracts: (1) ICCYA, (2) YIPCA,
and (3) Tree Crop Insurance. The insurance contract offered to bean growers is the ICCYA.
The trigger in this type of contract is given by the same three distinct and separate steps as it is
for PROAGRO: (1) the peril insured for needs to occur, (2) notification from the insured to the
insurer about a loss, and (3) damages to the crop needs to be related to the peril.
The indemnity formula for ICCYA is the same as for PROAGRO.
Crop insurance is compulsory for some crops for credit seekers from Banco Atlantida.
Seguros Atlantida charges a premium between 6 and 12 % depending on the perils covered and
the estimated risk of the farmer. For instance, farmers with irrigation systems have been charged
a premium of 6% and those without irrigation 8%. The copay for perils that do not include
drought is 15% of the loss. If drought is included in the perils covered, even if it is the only peril
insured, the copay is that of 20% of the loss.
Seguros Atlantida started operating with premiums and insurance prices set by the
reinsurer, based on the experience of those companies in other countries. With the passage of
time and experience, Seguros Atlantida has been readjusting their premiums.
Seguros Atlantida extended bean insurance for the first time in 2004. For an ICCYA type
of contract, three terms need to be specified: (1) the insured amount (2) the bean insurance price

77

per weight measure at time of buying the policy (indemnity price), and (3) the copay according
to the peril(s) insured.
Seguros Atlantida does extend contracts to groups of farmers, which works the same way
as the individual contract, except that the insurer only deals with one person for payment
purposes, but claims and indemnizations are done individually by each member in the group.
Seguros Atlantida insures bean farmers in a region that are less than 100 km away from
the insurer headquarter, if there are at least 18 ha insured. If the place is farther away than 100
km, there must be at least 35 ha. Nevertheless, if a farmer wanting to insure a crop with a land
area no less than 0.7 ha and farms near a place where there are already other farmers insured,
Seguros Atlantida would sell the contract to the farmer.
Seguros Atlantida has two reinsurers, Swiss RE and MAPFRE RE. Among other things,
the reinsurer trains the yield adjustment personnel. Yield adjustors are not Seguros Atlantida’s
employees, but are contracted by Seguros Atlantida. There have been a few complaints due to
the yield adjustments, but none have transcended to arbitration.
The crop insurance contractual procedure is the following: (1) filling of application form,
(2) payment of premium, (3) hand out of contract copy, (4) alert of peril, (5) yield adjustment,
and (6) indemnization.
4.1.4. Credit and microcredit
4.1.4.1.

Credit

Several banks offer credit to farmers in Honduras. Some of the most important banks
involved in lending to the agricultural sector are listed in the following section.

78

4.1.4.1.1.

BANADESA

BANADESA, the agricultural development bank of Honduras, extends loans to smallscale farmers. The types of loans extended are fiduciary, with a pledge, or with mortgage
collateral. The fiduciary loan requires two guarantors; the pledge guarantees a loan in the
amount of 40-50% of its appraised value; and the mortgage guarantees a loan in the amount of
60% of its appraised value. Loans are at the minimum amount in the order of L. 15,000 and at
maximum at L. 300,000. Loans are conditional on crop insurance and the farmer can buy crop
insurance from any of the crop insurance companies.
Annual interest rates are of 10%. The payback of the loan is after six months in one lump
sum. BANADESA has arrangements with insurance companies that if the farmer purchases crop
insurance the premium would be deducted from the loan and paid to the insurance company, and
in case of forfeit due to a peril, the insurance company would pay the bank and the farmer is
exonerated from the debt.
BANADESA does not have any arrangements with rural cashiers

23

for providing

microcredit to farmers. Rural cashiers need to handle their loans efficiently, which could be
from funds of BANADESA, since the risk falls on them and not on the borrowers. BANADESA
extends loans to rural cashiers in the range of L. 15,000 to L. 150,000 for the first two loans and
up to L. 300,000 from the third loan and onward. Life insurance is a requisite for all borrowers.

23

Rural cashiers are organized legal entities with the purpose of providing credit to small-scale
farmers.
79

4.1.4.1.2.
BAMER

24

BAMER

generally requires loans with one type of collateral, a mortgage. Guarantors,

usually a requisite in the fiduciary type of loan, are a complement of mortgage. Pledges are
restricted to a few set of assets and are also a complement to the mortgage. Mortgage collateral
needs to be an urban property and serves as a guarantee for 70% of its appraised value. Pledges
are also guarantee for 70% of their appraised value.
25

The minimum amount of a loan is L. 944,500 . The maximum loan covered by the
Small and Medium Firm division of BAMER is of L. 18,890,000. Prior to 2006, loans were
given for a minimum of L. 10,000. This big difference has been due in part because BAMER is
not financing the agricultural sector unless it’s a known client with guarantees. Interest rates
charged are from 12-15% with mortgage collateral, 15-17% with a pledge, and 19% if a
fiduciary guarantee is provided. BAMER does require crop insurance to extend a loan. The crop
insurance must cover drought, wind, and excess rain. Insurance can be bought from any
insurance company in Honduras. The payback of the loan is after nine months in one lump sum.
If a borrower is late with his/her payments, the following steps are taken: (1) call or notify the
client, (2) accept or reject justification, (3) readjust loan terms or sent to judicial department. If
farmers are late due to catastrophic events, the loan terms are adjusted as follows: (1) extend the
time period to pay, and (2) lower interest rates. BAMER does not have any arrangements with
rural cashiers nor do they provide microcredit.

24
25

Currently, BAMER is the Central American Bank (BAC).
Exchange rate is L. 19.03/US$.
80

4.1.4.1.3.

Banco Ficohsa

Banco Ficohsa generally requires a mortgage as collateral for its loans. The fiduciary type
of loan needs a reputable guarantor. Pledges need to be easily liquidable. Mortgage collateral
needs to be an urban property and serves as a guarantee for 70% of the appraised value. Pledges
are a guarantee for 50% of its appraised value.
The minimum amount of a loan is L. 25,000. The maximum loan handled by the credit
committee is of L. 5,000,000.
Interest rates charged are from 10-14%. Interest rates are greater as the time to repay
increases and if repayment is in a lump sum instead of several installments.
Banco Ficohsa does require crop insurance to extend a loan. The crop insurance must
cover drought and excess rain. Insurance can be bought from any insurance company in
Honduras. If a borrower is late with his/her payments, the following steps are taken: (1) call or
notify the client, (2) accept or reject justification, (3) readjust loan terms or sent to bank’s
judicial department. If farmers are late due to catastrophic events, the loan terms adjusted are the
following: (1) extend the time period to pay, and (2) lower interest rates.
Loans are usually not extended for a bean investment only, but it needs to be
accompanied with another investment due to the high risk Banco Ficohsa perceives as being
associated with the grain sector. Banco Ficohsa does not have any arrangements with rural
cashiers nor do they provide microcredit.
4.1.4.1.4.

Banco HSBC

HSBC generally does not extend loans for growing grains, including beans. It only
extends these loans to clients who have a long credit history with them. HBSC’s policy is a
result of high risk and governmental decrees that have led to a behavior by borrowers of not

81

repaying loans. For those clients to whom HSBC extends loans for beans, HSBC requires
mortgage and pledges as collateral. Fiduciary collateral are rarely approved. The mortgage
collateral needs to be an urban asset and serves as guarantee for 70% of the appraised value. The
pledge serves as guarantee for 30% of its appraised value.
The minimum loan amount for small-scale farmers is L. 100,000. The maximum, within
the small-scale qualification, is of L. 1,130,000.
Interest rates charged range from 14-18%, with no reduction in the interest rate due to
crop insurance.
HSBC may or may not require the farmer to obtain crop insurance. The farmer may
choose his own insurance company.
When farmers are late with their payments and no arrangements have been made between
the bank and the borrower after 60 days, the account falls under administrative delinquency and
several notices are sent to the farmer. After 90 days, delinquency accounts go to judicial
department and collaterals are liquidated.
HSBC does not have any arrangements with rural cashiers or groups of farmers for
extending loans, nor do they provide microcredit.
4.1.4.1.5.

Banco de Occidente

Banco de Occidente (BO) provides conventional credit to farmers, to microcredit
institutions, and to farmers’ groups.
BO requires mortgage or fiduciary collateral for a loan. Pledges are accepted as
complement of mortgage collateral. Mortgage collateral can be an urban or rural property and
serves as guarantee for 70% of the appraised value. Fiduciary collaterals need to present proof
of monthly income three times higher than the monthly financial installments to repay the loan.

82

The minimum amount of a loan is L. 100,000. The maximum loan is of L. 1,200,000 to a
fairly new client and up to L. 3,000,000 to a client with credit history.
Interest rates charged range from 12-16%, with no reduction in the interest rate due to
crop insurance. Borrowers with credit history are charged 12% and those without or with very
little credit history are charged interest rates from14-16%.
Crop insurance is not a requisite for obtaining a loan. When farmers are late with their
payments and no arrangements have been made between the bank and the borrower, the account
goes to the recuperations department. Nevertheless, flexibility is usually given in the form of
longer time period to repay the loan.
BO extends loans to farmer groups and microcredit institutions through conventional
credit and no arrangements or special considerations are given to these borrowers. The farmer
group and microcredit institution bears the risk when these extend loans to individual farmers.
BO does not provide microcredit.
4.1.4.1.6.

Banco Atlantida

Banco Atlantida (BA) generally does not extend loans for growing grains, including
beans. It only extends these loans to clients who have a long history with them. For those bean
farmers that BA does extend loans to, BA requires mortgage or fiduciary collateral. The
mortgage collateral could be an urban or rural asset and serves as guarantee for 50% of the
appraised value.
The minimum amount of a loan for a small-scale farmer is L. 10,000. The maximum,
within the small-scale qualification, is L. 3,000,000.
Interest rates charged vary according to the borrower’s credit history and if they have
insurance or not. A new client with insurance is charged 16%, while those without insurance are

83

charged 18%. A client with credit history and crop insurance would be charged 13%, while
those without insurance could be charged 16%. For bean farmers who have a long credit history
with the bank, crop insurance is required and it needs to be from Seguros Atlantida. Other bean
farmers would not get credit.
When farmers are late with their payments and no arrangements have been made between
the bank and the borrower, after 90 days the account falls under administrative delinquency and
several notices are sent to the farmer. After 180 days, delinquency accounts go to the judicial
department and collaterals are liquidated. BA does not have any arrangements with rural
cashiers or groups of farmers for extending loans, nor do they provide microcredit.
4.1.4.2.

Microcredit

There are around 21 microcredit institutions in Honduras that are associated with the
REDMICROH (Honduran Net of Microfinance Institutions).

REDMICROH is a non-profit

civil association that coordinates activities among its members to promote microfinance projects.
Key informants identified ODEF (Feminine

26

Entrepreneurial Development Organization) as

one of the most successful and largest microfinance institution in Honduras.
4.1.4.2.1.

ODEF

Staff from ODEF indicated that microcredit was not extended to bean related activities
when this was the only activity of the farmer. ODEF would only extend credit to the farmer if
activities besides agriculture, with less risk, are involved. ODEF staff also noted that microcredit
institutions in Honduras, in general, followed the same policies of not lending for agricultural
activities only. ODEF extends microcredit loans to individual farmers or to groups of farmers.

26

Even though the organization was created to be directed to women, their clientele constitutes
of 62.8% women and 37.2% men.
84

Even though the loan is extended to the whole group, the loan is handled on an individual basis.
Individual farmers who seek credit need to have either a pledge or mortgage collateral (against
the idea of microcredit). Group of farmers need to be consolidated as a legal entity and present
pledges as collateral to receive loans.
The minimum amount of a loan extended to either an individual or an individual in a
group is of L. 1,000. The maximum amount of a loan extended to an individual is L. 100,000;
and to an individual as part of a group the maximum is of L. 5,000 (it could go up to L. 20,000
with group credit history).
Crop insurance is not a requirement for obtaining loans because farmers receiving loans
would be involved in another activity not related to agriculture.
ODEF charges an annual interest rate of 24%. The strategies ODEF uses to recover the
loans include the following: (1) meticulous group selection, (2) financial advisor visits, (3) threat
to liquidate collateral, and (4) ask other members to pay for a defaulting member. If payments
are late due to an unexpected natural disaster, ODEF extends the repayment period of the loan,
although there has not been such a case in the past.
4.2. Farmer Survey
A total of 246 farmers were interviewed in 2007. The farmer survey collected data
needed to estimate the stochastic present value of wealth utility model (SPVWUM), including
farmers elicited yield distributions, crop production costs, income from other sources,
consumption, and farmers’ capital assets. Additional data collected included farmers’ use of
credit and crop insurance, as well as how farmers coped with disaster.

85

4.2.1. Farmers’ elicited yield distributions analysis
The key variable driving the model was intended to be the farmers elicited yield
distributions. However, when analyzing these data, some problems with the shape of the
distributions became apparent. The following section describes the problem with the yield
distribution elicitation.
4.2.1.1.

Enumerator bias

The first step after obtaining the farmer data was to generate graphs of the farmer elicited
distributions. However, these graphs had unexpected shapes (e.g.,uniform distributions, U
shaped, etc.), which led to suspicion of enumerator bias. Therefore, an analysis of variance of
the elicited yield distribution was conducted for traditional and modern varieties separately,
where the enumerator was entered as a variable. More specifically, the model below (Equation 5)
was applied for each type of variety:
Pr obijk    Inti  Vill j  Enumk  eijk

Equation 5

where Probijk= probability of interval i, at village j, and enumerator k
µ= mean probability across intervals, villages, and enumerators
Inti = interval i
Villj= village j
Enumk=Enumerator k
eijk= error term
Results showed (Tables 58 and 59 (Appendices)) an enumerator effect for modern
and traditional varieties. These results indicate enumerator bias (i.e., the shape of the distribution
varied by enumerator).
Equation 5 did not control for other factors that could explain the shape of the probability
distribution, which is needed to test appropriately if there is an enumerator effect. Therefore,
other variables were introduced into the model to control for these other factors that affect the
86

yield probability distribution. The other factors that might affect the yield distribution were
hypothesized to be capital, labor, inputs, and managerial techniques. Thus, to control for these
factors, the following variables, which serve as proxy variables, were introduced into the model
(i.e., Equation 5): type of land preparation system, and type of technology package. Due to the
few degrees of freedom, those variables were included in the Equation 5 sequentially.
“Type of land preparation system” was the first variable to be introduced in the analysis
of variance of Equation 5 and “type of technology package” followed. Even when additional
variables were included in the model (i.e., including the land preparation, or type of technology)
the enumerator effect remained (Tables 60 and 61 (Appendices)). To mitigate the enumerator
bias effect, farmers’ responses from enumerators with the highest percentage of uniform
distributions (for beans and corn) were removed from the analysis (Table 14).
After removing the responses of farmers interviewed by enumerator five (i.e., the worst
case of bias), the enumerator effect still persisted (Tables 62 and 63 (Appendices)). Only the
results for traditional varieties are used in the previous analysis because modern varieties had too
few observations for a reliable result. Modern varieties were assumed to have the same results in
the analysis of variance as traditional varieties did due to few degrees of freedom; thus only the
traditional variety analysis of variance generated a trustful output.

87

Table 14. Number and percentage of uniform probability distributions that were elicited by
enumerators.
Enumerator
# of even
Total # of
% of even
distributions
distributions
distributions
0
74
0.0
1
0
48
0.0
2
7
66
10.6
3
0
66
0.0
4
18
83
21.7
5
0
75
0.0
6
0
49
0.0
7
3
16
18.8
8
0
15
0.0
9
3
46
6.5
10
0
8
0.0
11
1
18
5.6
12
32
564
Total
Source: Pejuan Bean Farmer Survey, 2007.
Furthermore, farmers were asked the name of the varieties they grew and to classify them
as either a modern or traditional variety. Some farmers wrongly classified the type of variety.
The farmers with these mistakes were dropped from the sample because it creates uncertainty
when comparing traditional and modern varieties.
4.2.1.2.

Incomplete distributions

After partially correcting for enumerator bias (i.e., discarding the responses from
enumerator 5), an analysis showed two general forms of farmers’ elicited probability
distributions: one named half distribution (Figure 4) and the other named more than half
distribution (Figure 5). Half distribution was an elicited distribution that showed either a
probability density distribution that was increasing or was increasing with a final plateau-like
shape without a decreasing section. A more than half distribution was an elicited distribution
that showed first an increasing and then a decreasing section of the probability density
distribution. Other improper types of shapes were excluded from further analysis (e.g., uniform,
U shaped).
88

Figure 4. Half distribution (pdf) classification found in 2007 farmer survey, Honduras 2007.
Probability
density

Probability
density

Yield

Yield

Figure 5. More than half distribution (pdf) classification found in 2007 farmer survey,
Honduras 2007.
Probability
density

Yield
At this point, it is noted that, few farmers had half distributions for both modern and
traditional varieties. Similarly, few farmers had more than half distributions for both modern
and traditional varieties. This situation reduces the power of the comparison analysis, due to the
fact that traditional and modern varieties could only be analyzed across farmers instead of within
farmers.
It is hypothesized that those farmers with half and more than half distributions
misunderstood the elicitation question and instead of stating the maximum yield, they stated the
average or most likely yield.

89

Several t-tests were performed to evaluate the hypothesis that the half distributions
resulted from farmers reporting their most likely yield instead of the maximum yield as requested
in the questionnaire. Likewise, several t-tests for the more than half distributions were
performed to test if farmers had in fact understood the question.
The test for half distribution compared the maximum, as well as the midpoint of the 5

th

interval from the elicited distribution, to the farmer self-reported mean. There was no significant
difference between the farmers’ reported mean and the fifth interval midpoint for the traditional
variety half distributions, and there was a significant difference from the maximum reported
(Table 64 (Appendices)). Similarly, for the half distribution of modern varieties, the reported
mean was higher than the midpoint of the fifth interval, and there was no significant difference
from the maximum reported. These results support the hypothesis that farmers with half
distributions misunderstood the question and reported maximum yield instead of the most likely
yield.
th

th

The test for more than half distributions compared the 4 and 5 intervals midpoint to
the self-reported mean. The modern variety’s more than half distribution reported mean was
significantly different than the fourth interval’s midpoint and not different from the fifth interval
midpoint and the maximum. Contrary to the hypothesis, for the traditional variety’s more than
half distributions, there were significant differences between the reported mean and the midpoint
of the fourth interval, and also to the midpoint of the fifth interval. While some findings do not
support the hypothesis, most findings do support the hypothesis that half and more than half
distributions resulted from farmers misunderstanding the question.

90

4.2.1.3.

Villages and farmer categories

The final village groupings and farmer categories (i.e., technology package, and land
preparation systems) were created through a sequence of steps. First, villages with similar
estimated interval probability parameters and similar in elevation were pooled (Tables 65 and 66
(Appendices)). As a result, the eight sampled villages were reduced to three groups of villages
(i.e., Chirinos and Arauli; Talanga and Guaimaca; and Villa de San Francisco and Sabaneta de
Valle de Angeles), and two villages that were not pooled (i.e., San Pedro Alauca and La
Cienega).
Second, the technology package included four categories: modern varieties with low
input use, modern variety with high input use (i.e., modern variety package), traditional varieties
with low input use (i.e., traditional variety package), and traditional varieties with high input use.
Low and high input use was defined according to the rates of fertilizer. The threshold used to
separate low and high rates of fertilizer is 130 kg/ha of the combined amount of all chemical
fertilizers used by the bean farmer. This threshold is similar to the rates recommended by Rosas
(1996). High input use farmers are those that use amounts of fertilizer greater than or equal to
130 kg/ha of the combination of compound fertilizer (e.g.,12-24-12, 18-46-0, etc.) and urea.
The reason fertilizer was the only criterion used to define the use of input was because
most farmers in the sample used insecticides. Since all farmers used insecticides and their
application rate depended on the incidence of the pest attack, insecticide use would not provide a
meaningful way to define the technology package.
Finally, land preparation method was also used to categorize farmers. Farmers use two
types of land preparation system: non-mechanized, and mechanized, which included animal and
tractor traction. The non-mechanized category included all households that prepared land with

91

only manual labor. The mechanized category included those households that used oxen and/or
tractor. Thus, the combination of the three factors (i.e., land preparation system, type of variety,
and input use) result in eight categories, for which seven

27

categories are of interest.

In summary, the model includes seven technology categories of bean farmers for each of
the village groupings (i.e., Alauca; Arauli and Chirinos; Talanga and Guaimaca; La Cienega; and
Villa de San Francisco and Sabaneta de Valle de Angeles). The seven categories are the
following: 1) non-mechanized+traditional variety+low input use (Non-Mech+TV+LoIn); 2)
non-mechanized+traditional variety+high input use (Non-Mech+TV+HiIn); 3) nonmechanized+modern variety+low input use (Non-Mech+MV+LoIn); 4) nonmechanized+modern variety+high input use (Non-Mech+MV+HiIn); 5) mechanized+traditional
variety+low input use (Mech+TV+LoIn); 6) mechanized+traditional variety+high input use
(Mech+TV+HiIn); and,7) mechanized+modern variety+high input use (Mech+MV+HiIn). The
certainty equivalents for certain contrasting technology categories (e.g., Non-Mech+TV+LoIn
vs. Non-Mech+MV+HiIn) within each village will be compared. Note that these categories only
applied to the farmers’ bean production since the survey did not collect data on fertilizer use on
corn. Thus, corn technology categories only included four categories. The four corn technology
categories are: 1) non-mechanized+traditional variety, 2) non-mechanized+modern variety, 3)
mechanized+traditional variety, and 4) mechanized+modern variety. However, the SPVWUM
assumed yield and corn production costs for all runs to be for traditional corn variety grown at a
non-mechanized land preparation system.

27

The category mechanized land preparation+modern variety+low input (from a full factorial
combination of the three factors) is not included in the model because it is not likely that farmers
use this combination.
92

4.2.1.4.

Completing the distributions

Several efforts were done to complete the partial yield distributions (i.e., half and more
than half distributions) obtained from the farmer survey data, and obtain the final shapes of these
distributions. These efforts are specified in the Appendices section “Completing the
distributions.” However, in summary, the half and more than half distributions were completed
using several methods. Completed half distributions were not used because these did not
provide a sense of skewness. One method used to complete the more than half distributions, out
of two methods used, gave as a result, the parameters for an empirical distribution, which would
later serve to fit a smoother distribution.
4.2.1.5.

Fitted distribution and distribution parameters

Once all elicited probability distributions were completed for each of the farmers with
more than half distributions within each village, an empirical distribution for each technology
level were established. All the parameters were available

28

to estimate the empirical distribution

from the completed distributions. A smoother probability distribution was fitted to the empirical
distribution and the parameters were obtained.
The Beta distribution gave the best fit to the empirical distribution in terms of Chi-square
goodness-of-fit test compared to the Weibull, Pearson, Inverse Gauss, Log Normal, and other
distributions. The Beta distribution was then selected over the empirical distribution due to the
smoothness properties that it provides for the analysis.

28

The minimum, maximum, and probabilities for each interval give shape to an empirical
distribution.
93

The Beta parameters from the fitted Beta distribution to the more than half distribution
were obtained, however, from only one village, which was the village with several farmers with
more than half distributions.
Once the Beta distribution parameters (i.e., α, β), which give the distribution its shape,
were established, distributions were rescaled

29

for each of the categories using their assumed

values of minimum and mean yields,which are discussed on how they were obtained in later
sections of this chapter. Other parameters (i.e., coefficient of variation) of the yield probability
distributions were examined to observe if these were sensible. It was seen that these
distributions did not follow the distributions that were expected in terms of coefficient of
variation (CV), so the Beta distribution parameters were finally determined by setting certain
restrictions: 1) the distribution needed to have assumed minimum and mean parameters for each
technology level discussed in later sections, 2) the parameters have to be as close as possible
from the previous Beta parameters obtained, and 3) have a CV of around 30-40% due to
estimation of yield CVs with time series data (discussed in later sections). Additionally, all
villages kept the same shape parameters.
4.2.1.6.

Rescaling

There is a need to rescale because the Beta parameters obtained from fitting the Beta
distribution to the empirical distribution would provide a mean and variance parameter in a scale
from zero to one. Thus, the rescaling converts the mean and variance parameters from a scale

29

The Beta distributions need rescaling from its original size, since the original distributions
give a sense of shape and relationships of central tendency and variance parameter are in a scale
from zero to one.
94

from zero to one to the actual means and variances assumed for each of the technology
categories.
The parameters used to rescale the yield distributions for each of the technology
categories are the following: minimum yield values (min), mean yield values (ȗ), and the Beta
distribution parameters (α, β). These parameters characterize the whole yield probability
distribution. The formula that summarizes the rescaling of the distribution and relationship
among these parameters is the following:
ȗ = min + (max-min) α/ (α+β)
To adjust the Beta distribution parameters to fit the central tendency (i.e., mean) and
dispersion (i.e., CV) parameters, the assumed minimum and mean values of each distribution
were kept constant and the Beta distribution parameters were varied until satisfying the set of
restrictions. All maximum values obtained by satisfying the restrictions were in a reasonable
range.
As will be noted in the next section, the mean yield values of the bean and corn yield
distributions used in the SPVWUM were adjusted from the farmer survey mean values due to
general agreement that the yields were lower than normal in that year. In addition, it will be
made clear differences between technology levels and villages were obtained from a mix of
farmer elicited yield distributions, secondary data, and expert opinion.
4.2.2. Farmer crop revenues and costs
Crop revenues in the SPVWUM were based from the elicited yield distributions, which
were later modified, and from a price distribution of corn and beans adjusted to the mean values
farmers received at their villages.

95

The mean yield values from farmer reported yield data of previous years for both corn
and beans would have provided a first check on average yields of the elicited yield distributions,
if it had been a normal year and if the yield distributions elicitation had generated reasonable
shapes. Since neither of these conditions was true, the actual bean and corn yields used in the
SPVWUM were based on a combination of information. Differences in yield due to type of
variety came from Mather (2003) because there were few observations for modern varieties in
the farmer survey. Shape parameters were established as explained previously. Mean yield
values were established from secondary information. Yield differences from technology and
villages were obtained from primary and secondary information and expert opinion. Differences
in yields due to type of land preparation were obtained from differences observed between nonmechanized and mechanized land preparation among the farmers surveyed. Differences in yields
due to fertilizer were initially obtained from differences observed between low and high fertilizer
use among the farmers surveyed and later were adjusted relying on expert opinion. Differences
in yields among villages and minimum yields were based on expert opinion. For the base
village (i.e., La Cienega), bean mean yields were based on secondary data and corn mean yields
were based on survey yield data.
Production costs for the two crops included in the SPVWUM came from the farmer
survey and adjusted for differences between technologies, which are explained in the following
section.
4.2.2.1.

Corn revenues and costs

Many families that grow beans in Honduras also plant corn (i.e., 300,000 families). This
makes it necessary to think the bean farmers’ portfolio, as a minimum, include corn.

96

4.2.2.1.1.

Corn yields

Two and a half times as many farmers planted corn in the Primera season compared to the
Postrera season. Also, the corn area planted in the Primera, as reported by the National
Statistical Institute, was four times the area planted in the Postrera season. Corn yields, as
mentioned earlier, were estimated for four technology categories. Farmers’ corn yields in the
Primera season, for each of the four technology categories, are reported in Table 15. Due to the
high right skewness of the data, trimmed means

30

are reported.

Table 15. Trimmed mean corn yields (kg/ha) in Primera by type of land preparation and
type of variety.
Trimmed
Category
Mean (0.25)
N
Non-mechanized+traditional variety
1,252
30
Non-mechanized+modern variety
1,226
55
Mechanized+traditional variety
1,644
21
Mechanized+modern variety
1,302
36
Total
142
Source: Pejuan Bean Farmer Survey, 2007.
Table 15 above shows the mean yield of modern and traditional corn varieties under
different land preparation systems.
Due to a higher proportion of farmers planting corn in the Primera, it is assumed in the
SPVWUM that farmers planted in the Primera. Differences in corn yields due to land preparation
system were estimated by regressing corn yields on dummy variables of villages, type of variety,
and land preparation system (Table 67 (Appendices)), however, these were not used in the

30

A trimmed mean is a mean where a percentage or a proportion (α) of the total number of
values is discarded from the smallest and largest observations of the ordered values (Hoaglin
et.al., 1983), and the mean is computed from the remaining observations. The proportion
selected to discard depends on the tradeoff of robustness and efficiency. As more data is
discarded or trimmed, the more robust is the estimate but the efficiency decreases.
97

SPVWUM because it was assumed only a traditional corn variety was used in the portfolio for
all farmers. The difference in corn yield for farms using non-mechanized land preparation versus
mechanized land preparation is 152 kg/ha.
To get a sense of the corn yield risk farmers face, the coefficient of variation (CV) of
yields of corn was estimated with a time series of corn yields aggregated at the regional level. It
resulted in 9% CV. Based on farmers elicited yield distributions and experiences seen at other
studies the yield CV was established at 32% for the SPVWUM.
In the SPVWUM, differences among villages in corn yields were based on data provided
31

32

by expert opinion . Using these data, corn yields were regressed on rainfall , elevation above
sea level, and level of fertilizer to estimate the coefficients needed to obtain differences in corn
yields among villages (Table 68 (Appendices)). The yield differences between all other villages,
compared to La Cienega, are presented in Table 16.
Table 16. Corn yield differences (kg/ha) between La Cienega and every other pooled village.
Village
Corn Yield Differences between village and base
264
Alauca
238
Arauli & Chirinos
83
Guaimaca & Talanga
0
La Cienega
-7
Sabaneta & Villa de San Francisco
Source: Pejuan Farmer Survey, 2007.

Corn yields were not adjusted because the values reported by the sample of farmers were
similar to national corn yields and these two crops in the SPVWUM are planted in different
seasons. A yield of 1,252 kg/ha (Table 15) was used for yield in the base village, La Cienega.

31

Agronomist with 18 years of experience growing or supervising corn and bean plots.

32

Rainfall square was not included to the corn regression due to the levels of rainfall of the
villages are best explained just with the linear coefficient.
98

4.2.2.1.2.

Corn prices

The average price of corn that farmers receive in the Primera harvest period averaged L.
3.98/kg. Similarly to the procedure done to obtain differences among villages, price was
regressed on dummy variables for villages and type of variety. As expected, there were no
differences between variety types. Nevertheless, only one price was used because of the
assumption made of the use of only traditional varieties.
The results from the regression to adjust the corn prices for each village are presented in
Table 17. In Table 17, the difference in corn prices reflects how far the village was from a main
city. That is, the farther away the village is from a main selling city, the lower the price.
Nevertheless, for Alauca, the price is high even though it is close to the department capital. This
might actually reflect prices of a net buyer village.
Table 17. Corn prices in Primera by pooled villages.
Villages
Alauca
Arauli and Chirinos
Talanga and Guaimaca
La Cienega
Villa de San Francisco and Sabaneta de Valle de Angeles
Source: Pejuan Farmer Survey, 2007.

Price (L. /kg)
5.08
3.73
5.44
4.91
4.54

To get a sense of the corn price risk farmers face, the coefficient of variation (CV) of
prices of corn was estimated with a time series of corn prices at a market in the capital city of
Tegucigalpa. It resulted in 16% CV. It was then adjusted up to 20% due to differences from the
market level and at the farm level, and because of experiences in other studies.
Corn prices used in the SPVWUM are based on mean selling prices farmers obtained in
each of the villages. Also, the price distribution used in the SPVWUM assumed a log-normal
distribution with the previously estimated 20% coefficient of variation. Corn and traditional and
modern bean price variety correlations were obtained from historical price Honduran
99

Agricultural Product Information System (“Sistema de Información de Mercados de Productos
Agricolas de Honduras” (SIMPAH)). Correlation was estimated at 0.52 and 0.45, respectively.
4.2.2.1.3.

Corn production costs

Corn production costs (mean) for each of the four corn farmer categories are presented in
Table 18. These costs are cash expenses, so that the return on corn is return to family labor, land,
and management; this is also gross income (i.e., revenue minus variable costs (except for family
labor in this study)).
Table 18. Trimmed mean corn costs (L./ha) in Primera by type of land preparation and
type of variety.
Trimmed
Category
Mean (0.25)
N
Non-mechanized+traditional variety
5,303
30
Non-mechanized+modern variety
4,791
55
Mechanized+traditional variety
6,115
21
Mechanized+modern variety
7,067
35
Total
141
Source: Pejuan Farmer Survey, 2007.
Since all of the SPVWUM scenarios assumed that farmers planted a traditional corn
variety, the corresponding yield, as well as the costs of growing a traditional variety, is used in
the model. The corn production costs used in SPVWUM consist of a base cost, which is 77% of
the cost of non-mechanized land preparation+traditional variety category (Table 18), plus a yield
adjusted cost which depends on the yield. The 77% is approximately the percentage of cost
which is incurred during the first month after planting. Thus, if a farmer observes a bad season
and decides not to use the usual amount of inputs, he/she has already incurred 77% of the
production costs. Additionally, harvesting costs dependent on yields are included in the model.

100

4.2.2.2.

Bean revenues and costs

4.2.2.2.1.

Bean yields

Contrary to corn, two and a half times as many farmers planted beans in the Postrera
compared to the Primera. Also, the bean area planted in the Postrera, as reported by the National
Statistical Institute (INE), is more than twice the area planted in the Primera season. Farmers’
bean trimmed mean yields by land preparation system, type of variety, and technology level are
shown in Table 19.
Table 19. Trimmed mean bean yields (kg/ha) by type of land preparation, type of variety
and technology applied.
Trimmed
Category
Mean (0.25)
N
Non-mechanized+traditional variety+low input use
380
36
Non-mechanized+traditional variety+high input use
654
60
Non-mechanized+modern variety+low input use
241
8
Non-mechanized+modern variety+high input use
682
4
Mechanized+traditional variety+low input use
321
52
Mechanized+traditional variety+high input use
537
27
Mechanized+modern variety+low input use
130
3
Mechanized+modern variety+high input use
931
6
Total
196
Source: Pejuan Bean Farmer Survey, 2007.
Table 19 above was mostly used to estimate the response of traditional bean varieties to
fertilizer levels under different land preparation systems. However, for modern varieties, there
are too few observations to make a similar inference.
The yields of modern and traditional bean varieties given above are for informational
purposes only. Given that there is insufficient data on yields for modern bean varieties and that
the national bean yield is above those values, it was necessary to use secondary data to adjust
those values for the simulations.
To get a sense of the bean yield risk farmers face, the coefficient of variation (CV) of
yields of bean was estimated with a time series of bean yields aggregated at the regional level. It
101

resulted in 14% CV. Based on farmers elicited yield distributions and experiences seen at other
studies the yield CV was established at a range from 32% to 34% for the SPVWUM depending
on the technology.
Due to a higher proportion of farmers planting beans in the Postrera, in the SPVWUM, it
was assumed that farmers planted in the Postrera. As it was done for corn, differences in bean
yields due to land preparation system were estimated by regressing bean yields on dummy
variables for villages, type of variety, and land preparation system (Table 67 (Appendices)).
This analysis showed a difference in bean yield between non-mechanized land preparation and
mechanized land preparation to be 65 kg/ha.
Bean yield differences between modern and traditional varieties are not captured well by
the farmer survey, since few farmers planted modern varieties. Mather (2003) reported a 28%
difference between modern and traditional varieties which translates to a 200 kg/ha difference
when using the national average bean yield for the Postrera season (i.e., 713 kg/ha * 0.28 = 200
kg/ha).
Yield differences due to low and high fertilizer levels were adjusted from a 172 kg/ha
(i.e., Table 67 (Appendices)) to a 200 kg/ha difference. This adjustment was made due to expert
opinion expressing that bean yields normally respond to fertilizer in a one-to-one ratio (i.e., the
fertilizer use difference between low and high fertilizer was 200 kg/ha).
For the SPVWUM, differences in bean yields among villages were obtained, like in corn,
from data provided by experts. Using these data, bean yields were regressed on rainfall, as well
as rainfall squared, elevation above sea level, and level of fertilizer. This analysis provided the
coefficients needed to estimate differences in bean yields among villages (Table 68

102

(Appendices)). Table 20 presents bean yield differences between all other villages, compared to
La Cienega.
Table 20. Bean yield differences (kg/ha) between La Cienega and other villages.
Village
Bean Yield Differences between village and base
Alauca
-17
Arauli & Chirinos
-3
Guaimaca & Talanga
190
La Cienega (base)
0
Sabaneta & Villa de San Francisco
36
Source: Pejuan Farmer Survey, 2007.
It was assumed that the yield in the base village La Cienega for a nonmechanized+traditional variety+low input farmer (i.e., 856 kg/ha) was 20% above the national
average yield for the Postrera season (i.e., 713 kg/ha, Table 21). As mentioned before, this 20%
yield adjustment is justified because these villages are more specialized in beans than the average
village in Honduras. The mean yields in the other villages were estimated by adding the yield
differences between these villages and village La Cienega (Table 20) to the latter village
assumed yield. La Cienega was selected as the base village because the average elevation and
precipitation in these villages in El Paraiso and Francisco Morazan departments matches this
village--making it a representative village for these departments.
Table 21. Average national bean yields (kg/ha) for period 1990-2007.
Crop
Season
Yield
N
Beans
Primera
630
18
Beans
Postrera
713
18
Source: Honduras. Instituto Nacional de Estadistica (INE), Encuesta Agropecuaria
Básica, May 2007.
The minimum yields used in the SPVWUM for each of the technology categories
represent the riskiness from a bad event. A lower minimum yield represents a riskier technology
category. Table 22 presents the expected minimum yields for different bean categories.

103

Table 22. Expected minimum bean yields by category.
Category
Non-mechanized+traditional variety+low input use
Non-mechanized+traditional variety+high input use
Non-mechanized+modern variety+low input use
Non-mechanized+modern variety+high input use
Mechanized+traditional variety+low input use
Mechanized+traditional variety+high input use
Mechanized+modern variety+low input use
Mechanized+modern variety+high input use
Source: Pejuan Bean Farmer Survey, 2007 and expert opinion.

Average Minimum
kg/ha
145
130
15
10
140
130
10
5

The expected minimum values (kg/ha) and differences among these minimum values in
each of the technology categories were made from judgment based on minimum yields from the
farmer survey, assumptions from literature, and from expert opinion. Low yields are usually due
to drought and therefore the expected differences among technologies rely mostly on
precipitation.
Up until recent years, a traditional bean variety was expected to have a higher minimum
yield than a modern variety because of its higher drought tolerance. With the release of new
modern varieties, this is no longer the case. Nevertheless, farmers might not be aware of this and
still think the distribution of the modern varieties has a thicker lower tail. Table 22 assumes that
farmers still perceive that a modern variety has a lower minimum yield vis-à-vis traditional
varieties. It also assumes that at a high fertilization level, the minimum yield is lower than at the
low fertilization level. This is due to the combination of low precipitation and a high fertilizer
level which might burn the plant and result in lower yields. During drought, high fertilizer
increases the likelihood of chemical damage to the plant resulting in lower yields.
Additionally, non-mechanized preparation of the land leaves more mulch on the ground
than does mechanized land preparation, which helps maintain moisture in the soil and therefore
result in higher minimum yields.
104

However, for runs of the SPVWUM in @Risk, the actual minimum yield values used
were lower than those in Table 22. The minimum yield values set to run @Risk where not only
lower than those in Table 22, but many were even negative so that the actual minimum outcomes
33

from simulation runs, would result in the minimum values from Table 22 . These values will
be presented in the parameters section at the end of this chapter. This practice might be
unorthodox, however, this is a practice done by simulators when values resulting from the
simulation are not close to the values set for the distribution.
4.2.2.2.2.

Bean prices

Similar to corn, bean prices were regressed on dummy variables for villages and on type
of variety. However, only the mean prices for the traditional variety in each village were taken
and the differences among villages were adjusted. The bean prices for each village are presented
in Table 23.
Table 23. Bean prices (L. / kg) in Postrera by pooled villages and type of variety.
Type of Variety
Villages
Traditional
Modern*
11.00
9.90
Alauca
10.24
9.22
Arauli and Chirinos
10.72
9.65
Talanga and Guaimaca
12.99
11.69
La Cienega
11.17
10.05
Villa de San Francisco and Sabaneta de Valle de Angeles
*Modern variety prices are 10% lower than traditional varieties
Source: Pejuan Bean Farmer Survey, 2007.
As was the case for corn prices, there were no differences found between variety types.
However, this is because of the few observations for the price of modern varieties and their high

33

Since the minimum yield values observed from the preliminary runs in @Risk where much
higher than the values in Table 22, the minimum yield values were then reduced until they
matched the values in Table 22. This reduction lead to negative minimum points for minimum
yields but no actual negative values for yields because of truncation of yields at zero: Yield used
in simulation=max(yield from random number generator , 0).
105

variance. Since price differentials between modern and traditional bean varieties have been well
documented (Martel, 1996; Mather, 2003; Pejuan, 2005), a 10% price difference was assumed
for the SPVWUM (Table 23).
Also, the bean price distribution used in the SPVWUM assumed a log-normal distribution
and a 20% coefficient of variation.
To get a sense of the bean price risk farmers face, the coefficient of variation (CV) of
prices of beans was estimated with a time series of bean prices at a market in the capital city of
Tegucigalpa. It resulted in 19% CV. It was then adjusted up to 20% due to differences from the
market level and at the farm level, and because of experiences in other studies.
4.2.2.2.3.

Bean costs

Bean production costs are also part of the SPVWUM. These costs are subtracted from
revenues to obtain returns to family labor, land, capital, and management defined in this study as
gross income. Bean costs in the farmer survey are divided into input costs and activity
expenditure costs. Activity expenditure costs included costs of machinery used.
The general treatment of information obtained on inputs from the farmer survey was as
follows. If the price of the input was missing, it was replaced with the mean price of the input in
the villages or of a closely related input. Out of 1,077 observations, a total of 68 mean prices
were introduced. Also, input physical quantity outliers were replaced with frontier values (i.e.,
three standard deviations from the mean of each input value). Outliers were defined as unlikely
values that, probably, were either due to a misunderstanding of a question or a data entry error.
4.2.2.2.4.

Input use and costs

Bean inputs include seed, fertilizer, herbicide, fungicide, insecticide and other pesticides.
The inputs farmers used across the study areas varied greatly. Farmers mentioned 19 different

106

bean varieties (Table 24). Also, farmers used 9 different types of fertilizer, 16 brands of
herbicide, 11 brands of insecticide, and 9 brands of fungicide. Within the brands farmers used
different formulations of the product. The different bean varieties used by farmers, as well as
most commonly used brands of pesticides, are presented in the following sections.
4.2.2.2.5.

Varieties and seeding rate

Farmers used a diversity of bean varieties (Table 24). Nevertheless, only a few were
widely used in most villages. The overall mean seed rate was 42kg/ha.
Table 24. Bean varieties* used by farmers, Honduras 2007.
Variety Name
Amadeus**
Re Tinto***
Tio Canela**
Renegrido***
Arbolito***
Rosita***
Catracho****
Rural****
Chato***
Tinto***
Cuarenteño***
Vaina Blanca***
Danli****
Vaina Roja ***
Marciano**
Vaina Rosada ***
Paraisito***
Zamorano****
Payomo***
*Some varieties may have the same genetic profile but a different name.
**Modern variety
***Traditional variety
****Improved variety released before 1990, thus classified as a traditional variety.
4.2.2.2.6.

Fertilizer use

Farmers used mainly compound fertilizer. That is, they used types of fertilizer that would
provide at least two of the major nutrients (i.e., N, P, K). In addition, a few farmers used organic
matter and/or foliar fertilizer. The compound fertilizers types farmers used are presented in
Table 25, together with urea. The most popular fertilizer formula is 12-24-12-- 42% of farmers
applied it at an average rate of 142 kg/ha and paid an average price of L. 7.44/kg. The second
most popular fertilizer was 18-46-0-- 30% of farmers applied it at an average rate of 102 kg/ha
and paid an average price of L. 8.00/kg. The third most popular fertilizer was urea-- 5% of
107

farmers applied it at a rate of 78 kg/ha and paid an average price of L. 6.98/kg. Some farmers
used more than one type of fertilizer (e.g., 12-24-12 and urea).
Table 25. Fertilizer used by farmers, Honduras 2007.
Fertilizer formula
12-24-12
18-46-0
Urea (45-0-0)
20-20-20
15-15-15
Source: Pejuan Bean Farmer Survey, 2007.
4.2.2.2.7.

Herbicide use

Herbicide is widely used by farmers. The herbicides farmers used are presented in Table
26. The most popular herbicide was Gramoxone-- 69% of farmers applied it at an average rate
of 3.73 l/ha and paid an average price of L. 98.74/l. The second most popular herbicide was
Fusilade-- 13% of farmers applied it at an average rate of 1.24 l/ha and paid an average price of
L. 422/l. The third most popular herbicide was 2-4,D-- 5% of farmers applied it at a rate of 1.93
l/ha and paid an average price of L. 88/l. The fourth most popular herbicide was Glyphosate--3%
of farmers applied it at a rate of 2.57 l/ha and paid an average price of L. 136/l. Some farmers
used more than one type of herbicide (i.e., Gramoxone and 2-4,D).
4.2.2.2.8.

Fungicide use

Fungicide is used by a small percentage of farmers. The fungicides farmers used are
presented in Table 27. The most popular fungicide was Dithane-- 8% of farmers applied it at an
average rate of 2.07kg/ha and paid an average price of L. 131/kg. The second most popular
fungicide was Antracol-- 6% of farmers applied it at an average rate of 0.8 kg/ha and paid an
average price of L. 158/kg.

108

Table 26. Herbicides used by farmers, Honduras 2007.
Gramoxone
Rimax
Bayta
Basagran
Fusilade
Flin
Glyphosate
Basta
Rodoy
Remaxato
Boa
Riandi
Flet
Gesaprin
2-4 D
Rescate
Source: Pejuan Bean Farmer Survey, 2007.
Table 27. Fungicides used by farmers, Honduras 2007.
Dithane
Curatan
Antracol
Bravo
Mancozeb
Bayfolan
Aeroback
Batalla
Manzate
Source: Pejuan Bean Farmer survey, 2007.
4.2.2.2.9.

Insecticide use

Insecticide is as widely used as herbicide is. The insecticides farmers use are presented
in Table 28. The most popular insecticide was Folidol-- 34% of farmers applied it at an average
rate of 1.4 l/ha and paid an average price of L. 168/l. The second most popular insecticide was
Thiodan-- 14% of farmers applied it at an average rate of 1.41 l/ha and paid an average price of L.
174/l. The third most popular insecticide was Tamaron-- 12% of farmers applied it at a rate of
1.30 l/ha and paid an average price of L. 199/l. Several farmers use more than one type of
insecticide in the growing season.
Table 28. Insecticides used by farmers, Honduras 2007.
Insecticide
Insecticide
Folidol
Pirineta
Thiodan
Cipermetrina
Tamaron
Karate
Monarca
Muralla
Tambo
Decis
Source: Pejuan Bean Farmer Survey, 2007.

109

4.2.2.2.10.

Input expenditures

Input expenditures (L./ha) for each of the eight technology categories or choices are
presented in Table 29. Input expenditures for non-mechanized land preparation were,
unexpectedly, larger compared to mechanized land preparation system (Table 29). The
differences in input expenditures between the non-mechanized and mechanized land preparation
systems can be explained because of the difference in area between types of farmers (i.e.,
mechanized farms were generally larger), which is sometimes observed in empirical research.
Also, farmers using the non-mechanized land preparation system and traditional varieties with
high inputs have greater input expenditures than those using non-mechanized land preparation
system with the modern package. This result is not reliable due to few observations for farmers
using modern varieties. Thus, the input expenditure values for the modern variety categories
were modified for inclusion in the SPVWUM. This modification used the traditional variety as a
base, then the additional cost of the seed, and harvesting costs were included.
Table 29. Trimmed mean input expenditures (L./ha) by type of land preparation, type of
variety and technology applied.
Trimmed
Category
Mean (0.25)
N
Non-mechanized+traditional variety+low input use
1,466
36
Non-mechanized+traditional variety+high input use
3,475
60
Non-mechanized+modern variety+low input use
1,390
8
Non-mechanized+modern variety+high input use
2,744
4
Mechanized+traditional variety+low input use
1,399
53
Mechanized+traditional variety+high input use
2,649
28
Mechanized+modern variety+low input use
1,491
3
Mechanized+modern variety+high input use
2,262
6
Total
198
Source: Pejuan Bean Farmer Survey, 2007.
It is sometimes observed that farmers with larger areas of crop land use fewer inputs per
hectare compared to those who have smaller farms because of economies of scale or just because
they are being more efficient. To observe if higher input expenditures are related to area in the
110

survey data, input expenditures were regressed on area planted to bean by type of land
preparation system. The results show that farmers with less land have higher input expenditures
per hectare, probably because of unnecessary use of inputs.
A higher cost is expected for the modern variety category due to the higher price of a
modern seed variety. Thus, the modern variety input expenditure categories used the traditional
34

input expenditure cost as a base in the SPVWUM, added an annualized cost of modern seed
and reduced the cost of fungicides due to higher disease resistance of modern varieties.
4.2.2.2.11.

Partial budgets for variety change in each village

Table 3 in Chapter 2 showed a general partial budget analysis for the change in variety.
This showed that, in general, the change was profitable. However, prices and yields change
across the villages and it is important to see if the change in variety in each village is profitable,
assuming the prices, yields, and costs estimated for each village. The net change per hectare by
the change in variety from a traditional variety to a modern variety for La Cienega, Talanga and
Guaimaca, Arauli and Chirinos, Villa de San Francisco and Sabaneta de Valle de Angeles, and
Alauca are L. 741, L. 286, L. 442, L. 495, and L. 537, respectively (Tables 69 through 73 in
Appendices). Change in variety is profitable in each village.
4.2.2.2.12.

Activity expenditures

Activity expenditures included labor expenditures and mechanized activities. Labor
expenditures included manual labor for land preparation, planting, fertilizing, applying pesticides,
harvesting, threshing, and transportation of harvest. Mechanized activities include mechanized
plowing, planting, and threshing. As mentioned before, expenditures only include cash costs, so

34

Farmers use saved-grain as seed for several growing seasons.
111

family labor is not included. Thus, gross income is returns to capital, family labor, land, and
management. Activity expenditures, in this setting of returns to capital, family labor, land, and
management, can also be called residual activity expenditures. Farmers’ residual activity
expenditures are shown in Table 30.
Table 30. Trimmed mean residual activity expenditures (L./ha) by type of land preparation,
type of variety and technology applied.
Trimmed
Category
Mean (0.25)
N
Non-mechanized+traditional variety+low input use
4,077
36
Non-mechanized+traditional variety+high input use
4,027
60
Non-mechanized+modern variety+low input use
2,178
8
Non-mechanized+modern variety+high input use
7,487
4
Mechanized+traditional variety+low input use
3,535
53
Mechanized+traditional variety+high input use
4,306
28
Mechanized+modern variety+low input use
4,723
3
Mechanized+modern variety+high input use
4,213
6
Total
198
Source: Pejuan Bean Farmer Survey, 2007.
Residual activity expenditures were lower for the mechanized land preparation system
compared to the non-mechanized land preparation system, given that a traditional variety and
low inputs were used. Conversely, residual activity expenditures were higher for the mechanized
land preparation system compared to non-mechanized land preparation system, given that a
traditional variety and high inputs were used. These higher expenditures for the mechanized
land preparation system is due to a higher proportion of cash to non-cash expenditures for the
mechanized land preparation system, compared to the non-mechanized land preparation system.
Residual activity expenditures were added to the input expenditures to obtain the total
residual expenditures. Similar to corn, 77% of the total residual expenditures were used as a
base expenditure and then a yield adjusted cost is added to the total residual expenditures.
The total residual expenditures are subtracted from bean revenues to obtain gross income
(i.e., return to capital, family labor, land, and management).
112

4.2.2.2.13.

Area planted to beans and corn

In the SPVWUM, the area planted to bean and corn is assumed to be the same, given that
they are planted in different seasons. The area used in the SPVWUM is 1.9 hectares.
4.2.3. Gross income from other sources
Farmers have other sources of income in addition to the corn and bean revenue presented
before. Farmers engage in off-farm labor, plant other crops, and have other small businesses,
and net remittances. During the survey, farmers’ estimates of these gross income sources were
collected for the previous year and used to contribute (i.e., gross income from crop production is
also needed) to estimating the marginal wealth for all years (i.e., it was assumed that all income
from other sources was kept constant through the years).
In theory, net income should equal consumption expenditures. However, in all villages
there was a gap between total gross income--income from other sources added to mean gross
income from crop production—and consumption expenditures (i.e., consumption expenditures
were larger than gross income), probably because some farmers failed to report income from
other sources.
To address this problem, data for consumption, as well as data for gross income from
other sources, were pooled across villages with similar values for consumption and gross income
from other sources. These pooling created three groups of villages. After pooling, the gross
income from other sources for Alauca; Arauli and Chirinos, Talanga and Guaimaca; and Villa de
San Francisco and Sabaneta are L. 5,718; L. 10,478; and L. 20,039, respectively. Pooling did not
solve the problem of the gap. Due to this, gross income from other sources was adjusted to fill
the gap between gross income and consumption expenditures. The gap was filled by taking the

113

gross mean income approximately 3% above the consumption expenditure level across the
villages.
4.2.3.1.

Remittances

While net remittances are already included in the income from other sources section, it is
important to note that remittances were received by just a few farmers in each village. However,
the mean values were considerably high. While no farmers in Alauca received positive net
remittances, for Arauli and Chirinos, Talanga and Guaimaca, and La Cienega mean net
remittances were L. 24,698 (N=18). Villa de San Francisco and Sabaneta de Valle de Angeles
had a mean net remittance of L. 25,500 (N=7). Most of the farmers who received remittances
were in the mechanized land preparation+traditional variety+high fertilizer category (N=11).
4.2.4. Consumption
As previously described, consumption data were pooled similarly to the gross income from other
sources. Farmers’ food, clothing, and education expenditures for Alauca; Arauli and Chirinos,
Talanga and Guaimaca; and Villa de San Francisco and Sabaneta were L. 19,350; L. 31,845, and
L. 32,517, respectively.
4.2.5. Farmer assets
Farmers’ assets will account for the farmers’ initial wealth. In later years, changes in
wealth will be observed and computed for the SPVWUM. Farmers´ assets included land,
infrastructure, machinery, and farm animals. Farmers’ value of capital assets is shown in Table
31.

114

Table 31. Trimmed mean value (L.) of farmers’ capital assets by type of land preparation,
type of variety and technology applied.
Trimmed
Category
Mean (0.25)
N
Non-mechanized+traditional variety+low input use
108,325
36
Non-mechanized+traditional variety+high input use
139,467
60
Non-mechanized+modern variety+low input use
76,775
8
Non-mechanized+modern variety+high input use
65,995
4
Mechanized+traditional variety+low input use
182,564
53
Mechanized+traditional variety+high input use
293,966
28
Mechanized+modern variety+low input use
376,536
3
Mechanized+modern variety+high input use
504,100
6
Source: Pejuan Bean Farmer Survey, 2007.
4.2.6. Farmer use of credit
At one point in time during the 2001 to 2006 period, 40% of the farmers (N=193)
obtained credit from a formal institution (bank, rural cashier, NGO). The other 60% of the
farmers reported several reasons for not obtaining credit including the following: disliking
borrowing (42%), not having collateral (17%), banks denying credit (13%), risk of losing
collateral (12%), having alternative lenders, having high interest rates, and it being a burdensome
process.
Seventy-three percent of the farmers who did not borrow had not heard of crop insurance.
All farmers who did not borrow also did not buy crop insurance. From those who did borrow, 79%
had heard of crop insurance. Sixty-eight percent of those who borrowed bought crop insurance.
4.2.7. Farmer use of crop insurance
Farmers who had bought crop insurance accounted for about one-fifth of the sample (i.e.,
after removing the observations by enumerator 5 and farmers confusing traditional and modern
varieties). Farmer awareness and use of crop insurance is shown in Table 32. This shows that a
high percentage (i.e., 50%) of farmers still is not aware of a service available like crop insurance.

115

Table 32. Farmer awareness and use of crop insurance.
Number of farmers
Heard of crop
*
**
Ever bought crop insurance ?
insurance ?
No
Yes
99
0
No
50
42
Yes
*
Five missing values.
**

Three missing values.
Source: Pejuan Bean Farmer Survey, 2007.
Farmers who actually bought crop insurance, reported several reasons why they
purchased it (Table 33). The majority reported some version of the desire to insure their
investment. However, more than a third bought it because it was bundled with the loan. Farmers
who had heard of crop insurance, but had not bought it reported several reasons why they had not
bought crop insurance (Table 34). The main reason was “not knowing how crop insurance
works.”
Table 33. Percent of farmers quoting different reasons for buying bean insurance.
*
Reason for buying bean insurance
(%)
Bundled with loan
For indemnization
To automatically repay loan in case of crop loss
For indemnity and automatic repay of loan in case of crop loss
To insure investment
*
Two missing values; N=40
Source: Pejuan Bean Farmer Survey, 2007.

37.5
42.5
7.5
5
7.5

Table 34. Percent of farmers quoting different reasons for not buying bean insurance.
*
Reason for buying bean insurance
(%)
Lacking money
Not interested
Does not know the procedure/benefits of bean insurance
Just crops a small area
I have not obtained a loan
*

16 missing values/non responses; N=34

Source: Pejuan Bean Farmer survey, 2007.

116

18
32
44
3
3

4.2.8. Coverages in Honduras
In Honduras, the insurance coverages, which are the liability amounts (i.e.,
reimbursement to the farmer in case of total loss) as a percentage of the mean yield, are
approximately 43%. This coverage was estimated by: 1) dividing the liability (i.e., L. 5,000/ha)
by the indemnity price (i.e., L8.87/kg), which results in the insured amount of beans (in kg/ha),
and 2) dividing the insured amount of beans (in kg/ha) by the average yield (i.e., 1,319 kg/ha,
assuming the insurance contract is offered to a farmer using the modern variety package and who
uses the mechanized land preparation system). The coverage level offered is low. However,
Honduran crop insurance services are new and most insurance coverages in developed countries
started at low levels. The low levels are to insure an amount approximately equal to bean farmers
working capital, which is to assure farmers would be able to grow a new crop in case of a bad
event happening.
The SPVWUM uses coverage of both 45% and 75%. The latter coverage is a standard
U.S. coverage. The latter coverage is also used to observe if an improvement in the coverage
level will induce farmers to adopt modern varieties.
Given those coverages, the loading factors

35

where unusually high (i.e., 3 to 5 depending

on the technology category). Thus, a more reasonable loading factor (i.e., 1.4) was used for the
SPVWUM. This loading factor was used to observe if an improvement in the crop insurance
contract would encourage farmers to adopt the modern variety package.

35

The loading factor multiplies the actuarially fair premium (i.e., the premium equal to the
expected losses) and the product of the multiplication is the actual premium paid by the farmer.
117

4.2.9. Farmer coping with disaster
Forty percent of the farmers reported that they had never experienced a production level
low enough to not be able to cover the years’ worth of normal expenses (food, clothing,
education). The farmers who suffered an extreme shortage experienced it with different
frequency (Table 35).
Table 35. Farmers experiencing shortage of income different percent of time.
Percent of time experiencing shortages
Number of farmers Cumulative Percent
Less than or equal to 20%
48
42
Greater than 20% and less than or equal to 40%
32
70
Greater than 40% and less than or equal to 60%
22
87
Greater than 60% and less than or equal to 80%
10
97
Greater than 80%
3
100
*
2 missing values/non responses; N=115
Source: Pejuan Bean Farmer Survey, 2007.
The 60% of farmers that have experienced shortcomes need to deal with the effects of
perils and have different ways of coping with its shocks that impact their incomes. Farmers cope
with income shortages by mainly relying in working in another job besides their crop enterprise
(50%), selling their assets (24%), and asking for a loan from friends, family or other person
(14%) (Table 74 in Appendices).
4.3. Parameters used for simulation
Table 4 in Chapter 3 showed the different technology categories or choices that the
farmer could make. The sequence followed in making certainty equivalents comparisons is
important according to the question to be answered. The following section presents a summary
of the parameters used in the SPVWUM that are associated with different technologies (i.e., type
of variety, amount of fertilizer applied, land preparation system), level of yields, prices, and costs,
as well as the levels of wealth, consumption, and income from other sources for each village.
For this summary of the parameters used for the SPVWUM, it is necessary to note that the

118

certainty equivalents (as shown in Table 4) that will be computed for the five villages, will have
values for two levels of coverage (i.e., 45% and 75%), two rates of time preference (i.e., discount
rate) levels (i.e., 7% and 14%), and two loading factors (i.e., 1.0 and 1.4). The summary of
parameters used to run the SPVWUM for each of the villages are presented in Tables 36 to 40.
Many of the parameters are constant across several categories in each run of the
SPVWUM, in order to be able to make leveled comparisons and observe if a change in just one
aspect of the technology would make a difference to the farmer, or if a change of technology in a
bundle would make a difference. This way, confounding effects are avoided.

119

Table 36. Parameters* used in @Risk for the stochastic present value of wealth utility model for Alauca farmers, Honduras.
Parameter across
Technology category
technology categories
NMTVLI NMTVHI
NMMVLI NMMVHI MTVLI MTVHI
MMVHI
Bean mean yield value (kg/ha)
839
1,039
1,039
1,239
903
1,103
1,303
Bean price (L./kg)
11.00
11.00
9.90
9.90
11.00
11.00
9.90
Bean price standard deviation
2.20
2.20
1.98
1.98
2.20
2.20
1.98
Bean minimum yield value (kg/ha)
70
20
-72
-125
66
20
-172
(L./kg)
Bean base cost (L./ha)
4,269
5,777
4,558
6,065
3,800
5,356
5,644
Corn mean yield value (kg/ha)
1,516
1,516
1,516
1,516
1,516
1,516
1,516
Corn price (L./kg)
5.08
5.08
5.08
5.08
5.08
5.08
5.08
Corn price standard deviation
1.02
1.02
1.02
1.02
1.02
1.02
1.02
Consumption (L.)
21,314
21,314
21,314
21,314
21,314
21,314
21,314
(L./kg)
Income from other sources (L.)
11,218
11,218
11,218
11,218
11,218
11,218
11,218
Corn minimum yield value (kg/ha)
45
45
45
45
45
45
45
Corn base cost (L./ha)
4,083
4,083
4,083
4,083
4,083
4,083
4,083
Initial wealth (L.)
108,325
108,325
108,325
108,325
108,325
108,325
108,325
Bean and corn area (ha)
1.9
1.9
1.9
1.9
1.9
1.9
1.9
Bean and corn alpha distribution
5.08
5.08
5.08
5.08
5.08
5.08
5.08
Bean and corn beta distribution
4.50
4.50
4.50
4.50
4.50
4.50
4.50
parameter
House value (L.)
43,690
43,690
43,690
43,690
43,690
43,690
43,690
parameter
*For each technology category and village, two discount rates (i.e., 0.07 and 0.14) and two coefficients of relative risk aversion
(i.e., 2 and 4) were applied. For those choices that include insurance, two loading factors (i.e., 1 and 1.4) and two coverage levels
(i.e., 0.45 and 0.75) were used. The bank semi-annual interest rate used was of 0.08. NMTVLI= Non-Mech+TV+LoIn; NMTVHI=
Non-Mech+TV+HiIn; NMMVLI= Non-Mech+MV+LoIn; NMMVHI= Non-Mech+MV+HiIn; MTVLI=Mech+TV+LoIn;
MTVHI=Mech+TV+HiIn; MMVLI=Mech+MV+LoIn; MMVHI=Mech+MV+HiIn

120

Table 37. Parameters* used in @Risk for the stochastic present value of wealth utility model for Arauli and Chirinos farmers,
Honduras.
Parameter across
Technology category
technology categories
NMTVLI NMTVHI
NMMVLI NMMVHI MTVLI MTVHI
MMVHI
Bean mean yield value (kg/ha)
851
1,051
1,051
1,251
917
1,117
1,317
Bean price (L./kg)
10.24
10.24
9.22
9.22
10.24
10.24
9.22
Bean price standard deviation
2.05
2.05
1.84
1.84
2.05
2.05
1.84
Bean minimum yield value (kg/ha)
70
20
-72
-125
66
20
-172
(L./kg)
Bean base cost (L./ha)
4,269
5,777
4,558
6,065
3,800
5,356
5,644
Corn mean yield value (kg/ha)
1,490
1,490
1,490
1,490
1,490
1,490
1,490
Corn price (L./kg)
3.73
3.73
3.73
3.73
3.73
3.73
3.73
Corn price standard deviation
0.75
0.75
0.75
0.75
0.75
0.75
0.75
Consumption (L.)
34,019
34,019
34,019
34,019
34,019
34,019
34,019
(L./kg)
Income from other sources (L.)
27,978
27,978
27,978
27,978
27,978
27,978
27,978
Corn minimum yield value (kg/ha)
45
45
45
45
45
45
45
Corn base cost (L./ha)
4,083
4,083
4,083
4,083
4,083
4,083
4,083
Initial wealth (L.)
108,325
108,325
108,325
108,325
108,325
108,325
108,325
Bean and corn area (ha)
1.9
1.9
1.9
1.9
1.9
1.9
1.9
Bean and corn alpha distribution
5.08
5.08
5.08
5.08
5.08
5.08
5.08
Bean and corn beta distribution
4.50
4.50
4.50
4.50
4.50
4.50
4.50
parameter
House value (L.)
43,690
43,690
43,690
43,690
43,690
43,690
43,690
parameter
*For each technology category and village, two discount rates (i.e., 0.07 and 0.14) and two coefficients of relative risk aversion
(i.e., 2 and 4) were applied. For those choices that include insurance, two loading factors (i.e., 1 and 1.4) and two coverage levels
(i.e., 0.45 and 0.75) were used. The bank semi-annual interest rate used was of 0.08. NMTVLI= Non-Mech+TV+LoIn; NMTVHI=
Non-Mech+TV+HiIn; NMMVLI= Non-Mech+MV+LoIn; NMMVHI= Non-Mech+MV+HiIn; MTVLI=Mech+TV+LoIn;
MTVHI=Mech+TV+HiIn; MMVLI=Mech+MV+LoIn; MMVHI=Mech+MV+HiIn

121

Table 38. Parameters* used in @Risk for the stochastic present value of wealth utility model for Talanga and Guaimaca
farmers, Honduras.
Parameter across
Technology category
technology categories
NMTVLI NMTVHI
NMMVLI NMMVHI MTVLI MTVHI
MMVHI
Bean mean yield value (kg/ha)
1,046
1,246
1,246
1,446
1,109
1,309
1,509
Bean price (L./kg)
10.72
10.72
9.65
9.65
10.72
10.72
9.65
Bean price standard deviation
2.14
2.14
1.93
1.93
2.14
2.14
1.93
Bean minimum yield value (kg/ha)
70
20
-72
-125
66
20
-172
(L./kg)
Bean base cost (L./ha)
4,269
5,777
4,558
6,065
3,800
5,356
5,644
Corn mean yield value (kg/ha)
1,335
1,335
1,335
1,335
1,335
1,335
1,335
Corn price (L./kg)
5.44
5.44
5.44
5.44
5.44
5.44
5.44
Corn price standard deviation
1.09
1.09
1.09
1.09
1.09
1.09
1.09
Consumption (L.)
34,019
34,019
34,019
34,019
34,019
34,019
34,019
(L./kg)
Income from other sources (L.)
21,978
21,978
21,978
21,978
21,978
21,978
21,978
Corn minimum yield value (kg/ha)
45
45
45
45
45
45
45
Corn base cost (L./ha)
4,083
4,083
4,083
4,083
4,083
4,083
4,083
Initial wealth (L.)
108,325
108,325
108,325
108,325
108,325
108,325
108,325
Bean and corn area (ha)
1.9
1.9
1.9
1.9
1.9
1.9
1.9
Bean and corn alpha distribution
5.08
5.08
5.08
5.08
5.08
5.08
5.08
Bean and corn beta distribution
4.50
4.50
4.50
4.50
4.50
4.50
4.50
parameter
House value (L.)
43,690
43,690
43,690
43,690
43,690
43,690
43,690
parameter
*For each technology category and village, two discount rates (i.e., 0.07 and 0.14) and two coefficients of relative risk aversion
(i.e., 2 and 4) were applied. For those choices that include insurance, two loading factors (i.e., 1 and 1.4) and two coverage levels
(i.e., 0.45 and 0.75) were used. The bank semi-annual interest rate used was of 0.08. NMTVLI= Non-Mech+TV+LoIn; NMTVHI=
Non-Mech+TV+HiIn; NMMVLI= Non-Mech+MV+LoIn; NMMVHI= Non-Mech+MV+HiIn; MTVLI=Mech+TV+LoIn;
MTVHI=Mech+TV+HiIn; MMVLI=Mech+MV+LoIn; MMVHI=Mech+MV+HiIn

122

Table 39. Parameters* used in @Risk for the stochastic present value of wealth utility model for La Cienega farmers,
Honduras.
Parameter across
Technology category
technology categories
NMTVLI NMTVHI
NMMVLI NMMVHI MTVLI MTVHI
MMVHI
Bean mean yield value (kg/ha)
856
1,056
1,056
1,256
919
1,119
1,319
Bean price (L./kg)
12.99
12.99
11.69
11.69
12.99
12.99
11.69
Bean price standard deviation
2.60
2.60
2.34
2.34
2.60
2.60
2.34
Bean minimum yield value (kg/ha)
70
20
-72
-125
66
20
-172
(L./kg)
Bean base cost (L./ha)
4,269
5,777
4,558
6,065
3,800
5,356
5,644
Corn mean yield value (kg/ha)
1,252
1,252
1,252
1,252
1,252
1,252
1,252
Corn price (L./kg)
4.91
4.91
4.91
4.91
4.91
4.91
4.91
Corn price standard deviation
0.98
0.98
0.98
0.98
0.98
0.98
0.98
Consumption (L.)
34,019
34,019
34,019
34,019
34,019
34,019
34,019
(L./kg)
Income from other sources (L.)
22,978
22,978
22,978
22,978
22,978
22,978
22,978
Corn minimum yield value (kg/ha)
45
45
45
45
45
45
45
Corn base cost (L./ha)
4,083
4,083
4,083
4,083
4,083
4,083
4,083
Initial wealth (L.)
108,325
108,325
108,325
108,325
108,325
108,325
108,325
Bean and corn area (ha)
1.9
1.9
1.9
1.9
1.9
1.9
1.9
Bean and corn alpha distribution
5.08
5.08
5.08
5.08
5.08
5.08
5.08
Bean and corn beta distribution
4.50
4.50
4.50
4.50
4.50
4.50
4.50
parameter
House value (L.)
43,690
43,690
43,690
43,690
43,690
43,690
43,690
parameter
*For each technology category and village, two discount rates (i.e., 0.07 and 0.14) and two coefficients of relative risk aversion
(i.e., 2 and 4) were applied. For those choices that include insurance, two loading factors (i.e., 1 and 1.4) and two coverage levels
(i.e., 0.45 and 0.75) were used. The bank semi-annual interest rate used was of 0.08. NMTVLI= Non-Mech+TV+LoIn; NMTVHI=
Non-Mech+TV+HiIn; NMMVLI= Non-Mech+MV+LoIn; NMMVHI= Non-Mech+MV+HiIn; MTVLI=Mech+TV+LoIn;
MTVHI=Mech+TV+HiIn; MMVLI=Mech+MV+LoIn; MMVHI=Mech+MV+HiIn

123

Table 40. Parameters* used in @Risk for the stochastic present value of wealth utility model for Villa de San Francisco and
Sabaneta de Valle de Angeles farmers, Honduras.
Parameter across
Technology category
technology categories
NMTVLI NMTVHI
NMMVLI NMMVHI MTVLI MTVHI
MMVHI
Bean mean yield value (kg/ha)
891
1,091
1,091
1,291
955
1,155
1,355
Bean price (L./kg)
11.17
11.17
10.05
10.05
11.17
11.17
10.05
Bean price standard deviation
2.23
2.23
2.01
2.01
2.23
2.23
2.01
Bean minimum yield value (kg/ha)
70
20
-72
-125
66
20
-172
(L./kg)
Bean base cost (L./ha)
4,269
5,777
4,558
6,065
3,800
5,356
5,644
Corn mean yield value (kg/ha)
1,245
1,245
1,245
1,245
1,245
1,245
1,245
Corn price (L./kg)
4.54
4.54
4.54
4.54
4.54
4.54
4.54
Corn price standard deviation
0.91
0.91
0.91
0.91
0.91
0.91
0.91
Consumption (L.)
32,633
32,633
32,633
32,633
32,633
32,633
32,633
(L./kg)
Income from other sources (L.)
24,539
24,539
24,539
24,539
24,539
24,539
24,539
Corn minimum yield value (kg/ha)
45
45
45
45
45
45
45
Corn base cost (L./ha)
4,083
4,083
4,083
4,083
4,083
4,083
4,083
Initial wealth (L.)
108,325
108,325
108,325
108,325
108,325
108,325
108,325
Bean and corn area (ha)
1.9
1.9
1.9
1.9
1.9
1.9
1.9
Bean and corn alpha distribution
5.08
5.08
5.08
5.08
5.08
5.08
5.08
Bean and corn beta distribution
4.50
4.50
4.50
4.50
4.50
4.50
4.50
parameter
House value (L.)
43,690
43,690
43,690
43,690
43,690
43,690
43,690
parameter
*For each technology category and village, two discount rates (i.e., 0.07 and 0.14) and two coefficients of relative risk aversion
(i.e., 2 and 4) were applied. For those choices that include insurance, two loading factors (i.e., 1 and 1.4) and two coverage levels
(i.e., 0.45 and 0.75) were used. The bank semi-annual interest rate used was of 0.08. NMTVLI= Non-Mech+TV+LoIn; NMTVHI=
Non-Mech+TV+HiIn; NMMVLI= Non-Mech+MV+LoIn; NMMVHI= Non-Mech+MV+HiIn; MTVLI=Mech+TV+LoIn;
MTVHI=Mech+TV+HiIn; MMVLI=Mech+MV+LoIn; MMVHI=Mech+MV+HiIn

124

4.4. Correlation Matrix
The correlations between prices and yields at different technology levels are presented in
Table 41. The correlation matrix shows that bean prices of modern and traditional bean varieties
are highly positively correlated (i.e., when the price of grain of a traditional bean variety
increases, the price of the grain of the modern bean variety is expected to increase), while the
correlation between corn and bean prices is medium positively correlated (i.e., correlations are
0.52 and 0.45). Also, prices and yields are not correlated for beans and corn.
Table 41. Correlation matrix of simulation random variables.
Random
1*
2
3
4
5
6
7
8
9
10
11
12
Variables
1
1.00 0.95 0.52 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
2
0.95 1.00 0.45 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
3
0.52 0.45 1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
4
0.00 0.00 0.00 1.00 0.92 0.88 0.88 0.92 0.88 0.88 0.83 0.42
5
0.00 0.00 0.00 0.92 1.00 0.88 0.88 0.88 0.92 0.83 0.88 0.42
6
0.00 0.00 0.00 0.88 0.88 1.00 0.92 0.88 0.83 0.92 0.88 0.42
7
0.00 0.00 0.00 0.88 0.88 0.92 1.00 0.83 0.88 0.88 0.92 0.42
8
0.00 0.00 0.00 0.92 0.88 0.88 0.83 1.00 0.92 0.92 0.88 0.42
9
0.00 0.00 0.00 0.88 0.92 0.83 0.88 0.92 1.00 0.88 0.92 0.42
10
0.00 0.00 0.00 0.88 0.83 0.92 0.88 0.92 0.88 1.00 0.92 0.42
11
0.00 0.00 0.00 0.83 0.88 0.88 0.92 0.88 0.92 0.92 1.00 0.42
12
0.00 0.00 0.00 0.42 0.42 0.42 0.42 0.42 0.42 0.42 0.42 1.00
*1=Traditional Bean Price, 2=Modern Bean Price, 3=Corn Output Price, 4=Bean Yield Nonmechanized Traditional Low Input, 5=Bean Yield Non-mechanized Traditional High Input,
6=Bean Yield Non-mechanized Modern Low Input, 7=Bean Yield Non-mechanized Modern
High Input, 8=Bean Yield Mechanical Traditional Low Input, 9=Bean Yield Mechanical
Traditional High Input, 10=Bean Yield Mechanical Modern Low Input, 11=Bean Yield
Mechanical Modern High Input, 12=Corn Yield Traditional

125

The purpose of the correlation matrix is to provide information on the strength of
association and direction of movement between two variables in the simulations, which also
serve as constraints in the simulation runs. Thus, if there is a price increase for the traditional
variety in the simulation, on average we would expect also a price increase for beans of a modern
variety, and not a random huge decrease. This avoids obtaining runs that would not likely
happen in a real life situation.
4.5. Analysis of Crop Insurance in a Portfolio Setting
Crop insurance seems an attractive alternative to reduce income risk when it is presented
as a sole alternative to a crop’s income. However, when crop insurance is used in a portfolio
setting, its impact on risk reduction might be reduced to a modest effect. The diminution of the
risk reduction effect is due to the portfolio’s risk mitigation effect. This portfolio risk mitigation
effect depends on the correlations between the assets included in the portfolio. The closer the
correlation of two assets is to negative one the greater the risk mitigation effect. However, even
with values close to zero, significant risk reduction can be obtained if the mix of assets is done
properly. The difference in the level of risk reduction of crop insurance alone and crop insurance
in a portfolio setting also depends on the percentage of that crop’s contribution to the total
portfolio.
The crops involved in this study’s model are beans and corn planted in two different
growing seasons. The correlation between beans and corn yields is of 0.42. In addition, off-farm
income and income from other activities are included in the portfolio. However, these are
constants so there is no correlation of this income with other variables. The percentage
contribution of bean income to the total yearly income in the portfolio for the base village La
Cienega is 86%. Given the medium correlation and high contribution of beans to the total

126

portfolio, insurance is expected to have medium risk mitigation when included in a farmer’s
portfolio. The results section for La Cienega and the rest of the villages will show if insurance
reduces risk.
The next section provides figures for cdfs that show the risk mitigation effects of crop
insurance in a regressive manner (i.e., risk is reduced in a decreasing way) as the type of
insurance contracts are presented. Revenue insurance contracts on bean activity as sole source
of income are presented first and ending with yield insurance contracts on beans in a portfolio.
4.6. Base Village Simulation and Cumulative Probability Distribution Functions
Cumulative probability distribution functions on revenue, gross income, and end of year
wealth will be presented in a regressive manner, a sequence that will show diminishing risk
reduction effects as the types of crop insurance contracts change for La Cienega; from revenue
insurance contracts to yield insurance contracts; from no copay to copay; from revenue streams
to gross income streams; and from sole enterprise (i.e., beans) to portfolio (i.e., beans, corn and
gross income from other sources).
Revenue insurance contracts are presented first, even though they are not offered in
Honduras, because these types of contracts completely eliminate downside risk for a certain
threshold. Thus, this type of contract would be what a farmer might have in mind as the perfect
contract.
4.6.1. La Cienega farmers distribution results for with and without insurance
Farmer cumulative probability distribution results (i.e., revenue, gross income, and
wealth distributions) for La Cienega will be presented to provide risk reduction comparisons for
different types of insurance contracts and income measures (i.e., revenue, gross income, wealth).
These distributions will be presented in the following sequence: 1) bean revenue cumulative

127

distribution functions-- without insurance and with insurance contracts (i.e., yield and revenue
insurance), 2) bean gross income cumulative distribution functions-- without insurance and with
insurance contracts (i.e., yield insurance), 3) total production gross income cumulative
distribution functions-- without insurance and with insurance contracts (i.e., yield insurance), and
4) end of year wealth cumulative distribution functions -- without insurance and with insurance
contracts (i.e., yield insurance). Within these sequence, insurance contracts with and without
copay will be compared, as well as high and low indemnity prices will be compared.
When distributions are presented in this way, it is easier to observe the risk reduction
effect of the different types of insurance contracts and of the portfolio. All the cumulative
distribution functions in this section are for a traditional variety, non-mechanized land
preparation system with low fertilizer use. Also, all insurance contracts specified in the next
section will be analyzed for actuarially fair premiums and insurance coverage used in many U.S.
contracts (i.e., 75%) to show the risk reducing effect that crop insurance could achieve, since
Honduran insurance contracts have coverages of around 45%, which is considered a low
coverage and would not likely show a risk reduction. Results of certainty equivalents and
insolvency parameters will be presented in the next chapter.
4.6.1.1.

Revenue distributions

The revenue crop insurance contract provides the highest risk reduction (i.e., downside
risk) to revenue streams. This contract indemnizes whenever revenue falls below a certain
threshold (e.g., L. 14,786 in Figure 6). Thus, revenue crop insurance contracts protect against
shortfalls in both price and yields, whenever the shortfall of either (i.e., price or yield) or both
have an impact large enough to make the revenue lower than the threshold. Yield insurance
contracts do not provide this drastic cut of downside revenues because price shortfalls are not

128

covered. Revenue insurance with an actuarially fair premium stochastically dominates in the
36

second degree the no insurance case as it is observed (and verified through tests with the data )
in Figure 6.
Once a copay policy is introduced in the revenue insurance contract, the risk is not
completely eliminated because the farmer covers some of the loss. This is observed in Figure 7.
The revenue insurance curve is no longer vertical but instead has a steep slope. Revenue
insurance without copay dominates the copay counterpart (i.e., 20% copay) in the second degree.

36

All the statements on stochastic dominance have been verified with the data. For verification
of second degree stochastic dominance, the data was ordered in ascending order and accumulated
for each series. The accumulated series were later subtracted from each other. If the series of
differences changes sign, there is no second degree stochastic dominance. If there is no sign
change, the series with the highest values dominates in the second degree. Similarly, for first
degree stochastic dominance the data was ordered in ascending order and each series was
subtracted from the other. If the series of differences changes sign, there is no first degree
stochastic dominance. If there is no sign change, the series with the highest values dominates in
the first degree.
129

Figure 6. Bean revenue (L./ha) cdfs for revenue insurance and no insurance; for Nonmechanized+traditional variety+low input use category, no copay, load=1.00, 75%
coverage, high indemnity price.

Figure 7. Bean revenue (L./ha) cdfs for revenue insurance with and without copay; for
Non-mechanized+traditional variety+low input use category, load=1.00, 75% coverage,
high indemnity price.

130

Yield insurance does not completely eliminate the risk of being below a certain revenue
level as revenue insurance does. This can be seen in Figure 8. The reason risk is not completely
eliminated is because income is also affected by price of the crop. Thus, there are times when
yield insurance has not been triggered and no indemnity has been paid, nevertheless, the price is
low enough that income is below the minimum income threshold used for the revenue insurance
case. In this case, there is no stochastic dominance in the second degree.
Figure 8. Bean revenue (L./ha) cdfs for revenue insurance and yield insurance; for Nonmechanized+traditional variety+low input use category, no copay, load=1.00, 75%
coverage, high indemnity price.

Figure 9 shows bean total revenue distributions with and without yield insurance. As
expected, the yield insurance distribution is below the no insurance case in the lower end of the
distribution, which means there is a lower probability of observing low values when insurance is
used. This is also seen in the minimum values. As we move along the revenue distribution
curve, the insurance curves cross due to the premium paid. We can also observe that since it is

131

an actuarially fair premium, the revenue means are the same. In this case, the insurance case
dominates the no insurance case in the second degree.
Figure 9. Bean revenue (L./ha) cdfs for yield insurance and no insurance; for Nonmechanized+traditional variety+low input use category, no copay, load=1.00, 75%
coverage, high indemnity price.

Also, when a copay policy is indicated in an insurance contract the farmer bears some of
the losses. This lower protection of an insurance contract with copay is observed in the yield
insurance contracts by the reduced area between the curves of a yield insurance contract with
copay versus no insurance (Figure 10), compared to the curves of a yield insurance contract and
no copay versus no insurance; and also between the yield insurance curves of bean revenues with
and without copay (Figure 11). In Figure 10, the insurance case dominates the no insurance case
in the second degree and in Figure 11, there is no dominance in the second degree.

132

Figure 10. Bean revenue (L./ha) cdfs for yield insurance and no insurance; for Nonmechanized+traditional variety+low input use category, with copay, load=1.00, 75%
coverage, high indemnity price.

Figure 11. Bean revenue (L./ha) cdfs for yield insurance with and without copay; for Nonmechanized+traditional variety+low input use category, load=1.00, 75% coverage, high
indemnity price.

133

When the indemnity price is lower than the expected price, the risk reduction by the use
of insurance is reduced. This is observed in Figure 12, where the high price revenue distribution
is below the low price revenue distribution in the lower end of the cdf.
Figure 12. Bean revenue (L./ha) cdfs for yield insurance with high and low indemnity price;
for Non-mechanized+traditional variety+low input use category, with copay, load=1.00, 75%
coverage.

4.6.1.2.

Gross income distributions

Gross income in this study is returns to capital, family labor, land, and management.
Gross income distributions in the following sections are first shown for the bean enterprise alone.
Later, these gross incomes become part of the marginal wealth (i.e., together with income from
other sources) in each year of the 20 year sequence in the SPVWUM, when beans are in the
portfolio.
Gross income distributions, compared to revenue distributions, are slightly different in
terms of shape but meaningfully different in terms of amounts. It can be observed in Figure 13
how far apart the gross income distribution is from the revenue distribution. The bean gross

134

income and revenue distribution curves are not exactly parallel sigmoid curves in Figure 13.
This is because the bean production costs are not fixed, but instead depend on the yield.
Risk reduction by crop insurance on bean gross income is similar to risk reduction on
bean revenues in terms of relative shapes of their distribution with and without insurance (i.e.,
but different in magnitudes). This is observed when comparing gross income distributions of no
insurance versus yield insurance, when insurance has a copay and high indemnity price (Figure
14), and low indemnity price (Figure 15). Both Figure 14 and Figure 15 show, as it did for
revenue distributions, that yield insurance dominates in the second degree the no insurance case.
Figure 13. Bean revenue (L./ha) and gross income (L./ha) cdfs with no insurance; for Nonmechanized+traditional variety+low input use category.

135

Figure 14. Bean gross income (L./ha) cdfs for yield insurance and no insurance; for Nonmechanized+traditional variety+low input use category, with copay, load=1.00, 75%
coverage, high indemnity price.

Figure 15. Bean gross income (L./ha) cdfs for yield insurance and no insurance; for Nonmechanized+traditional variety+low input use category, with copay, load=1.00, 75%
coverage, low indemnity price.

136

4.6.1.3.

Portfolio distributions

As mentioned before, crop insurance does not reduce risk in the same proportion when
there is only one crop, compared to when there are several crops. This is because of the
diversification effect or portfolio effect. The risk reduction will depend on the covariance
between the two crops and their share of income in the portfolio. To show this, first, the
difference in gross revenue magnitudes is shown in Figure 16, when only beans are present
compared to when beans and corn are in the portfolio. There is no stochastic dominance in the
second degree in Figure 16. Then, Figure 17 shows distributions of corn and beans in a portfolio
with and without yield insurance. The area between the curves in Figure 15 in the lower end of
these graphs is proportionally higher than the respective area between the curves in Figure 17.
This indicates risk reduction by crop insurance is lower in a portfolio, compared to when it is in a
sole enterprise. This is one of the reasons why farmers grow several crops. In Figure 17, there is
second degree stochastic dominance, although not in first degree.
Figure 16. Gross income cdfs (L./ha) for beans and total production (beans and corn); for
Non-mechanized+traditional variety+low input use category and no insurance.

137

Figure 17. Total production gross income (beans and corn) (L./ha) cdfs for bean yield
insurance and no insurance; for Non-mechanized+traditional variety+low input use
category, with copay, load=1.00, 75% coverage, low indemnity price.

Finally we move to the whole portfolio setting, which includes bean and corn gross
income, and gross income from other sources. The final wealth in year one is the sum of the
initial wealth and all the sources of gross incomes (Figure 18). Figure 18 shows the end of year
wealth for the first year for both purchasing and not purchasing yield insurance.
After observing the whole portfolio setting for the first year, it is helpful to observe the
dynamics through the years of the cdfs of the technology categories. Figure 19 shows the cdfs of
th

end of year wealth at the 20 year for both purchasing and not purchasing yield insurance.
Farmers’ wealth mean values increase as years increase (compared to Figure 18) due to wealth
accumulation--not all outcomes increase the value of wealth, however mean values do. The
lower end of the distribution presents an edge (i.e., a step) for both curves. This is due to the
farmers that go insolvent after 20 years. Thus, minimum values coincide, given that the
threshold to stop producing (insolvency) is the same for both distributions. The maximum value
138

for the insurance case, as expected, is lower than the no insurance case, due to the payment of the
premium. In Figure 18 the insurance case stochastically dominates the no insurance case, while
in Figure 19 it does not. The reason for the difference in stochastic dominance is the insolvency
inclusion in the simulation takes effect after several years impacting more on the insurance case.
Figure 18. End of first year wealth (L.) cdfs for yield insurance and no insurance; for Nonmechanized+traditional variety+low input use category, with copay, load=1.00, 75%
coverage, low indemnity price.

139

Figure 19. End of twentieth year wealth (L.) cdfs for yield insurance and no insurance; for
Non-mechanized+traditional variety+low input use category, with copay, load=1.00, 75%
coverage, low indemnity price.

4.6.1.4.

Chance of not meeting consumption expenditures

After running simulations with the whole portfolio, farmers in each village growing
traditional varieties would have different chances of not meeting their consumption expenditures.
Within each village the value would change if insurance was used on the traditional variety.
These chances were observed at each village and for the case of insurance and no insurance by
looking at the height of the cumulative probability density functions of accumulated wealth in
year one at the point of initial wealth minus half of the consumption expenditures. The chances
of not meeting consumption expenditures for the case of no insurance and for villages La
Cienega, Talanga and Guaimaca, Arauli and Chirinos, Alauca, and Villa de San Francisco and
Sabaneta de Valle de Angeles are 10.8%, 11.1%, 1.6%, 25.6%, and 6%, respectively. The
chances of not meeting consumption expenditures for the case of insurance and for villages La
Cienega, Talanga and Guaimaca, Arauli and Chirinos, Alauca, and Villa de San Francisco and

140

Sabaneta de Valle de Angeles are 10.4%, 10.6%, 0.8%, 25.6%, and 5.1%, respectively. Almost
in all villages farmers reduce the chance of not meeting consumption expenditures when using
insurance, except for Alauca. The average chance of not meeting consumption expenditures
reduces from 11.02% without insurance to 10.5% with insurance.
4.7. Chapter Summary
The study uses data collected from key informant interviews, survey data and secondary
data. The key informants interviewed were personnel from insurance companies, banks, and
microcredit institutions. The insurance companies personnel interviewed in the study were from
ProAgro, HSBC, Seguros Equidad and Seguros Atlantida. The banks and microcredit personnel
interviewed in the study were from BAMER, Ficohsa, HSBC, Banco de Occidente, Banco
Atlántida, and ODEF. Interest rates charged by banks ranged from 10% to 18%. The study used
a semi-annual interest rate of 8% and a rate of time preference of 7% and 14%, which includes a
middle value within the range of interest rate as well as a lower value to observe the effect of the
discount rate on farmers.
Crop insurance coverages were approximately 45% in Honduras and with high loading
factors that ranged from 3-5. Copays in the contracts offered by insurance companies in
Honduras are from 15% to 30% depending on the peril. The study used coverages of 45% and
75% to observe the effect of the coverage on farmers. The study also used actuarially fair
premiums as well as a load of 1.4, as it is thought to be a reasonable loading factor for insurance
companies to charge.
Enumerator bias was found on the subjective yield probability distributions of corn and
beans. Due to this, bean and corn yields were obtained through a mix of elicitation of farmer
subjective yield distributions, secondary data, and expert opinion. Time series data at the market

141

level in the capital city of Tegucigalpa revealed that beans have approximately a 19% CV, and
corn has a 16% CV, which were adjusted due to differences of CVs from the market level and
the farm level, and because of experiences seen at other studies. The adjustment made was for
bean farmers to face a coefficient of variation (CV) of 20% for both bean and corn price. Based
on farmers elicited yield distributions and experiences seen at other studies, farmers face a
perceived yield risk at a range of approximately 32-34% CV for beans depending on the bean
technology, and 32% for traditional corn varieties.
Consumption expenditures were obtained from the farmer survey. Income from other
sources was adjusted from the value obtained from the survey to a new amount which would be
enough to add, together with production gross income, to be 3% above the consumption
expenditures. All previous information was used to run the stochastic present value of wealth
utility model. In addition, farmers were asked how they coped with disaster. Farmers cope with
income shortages by mainly relying in working in another job besides their crop enterprise (50%),
selling their assets (24%), and asking for a loan from friends, family or other person (14%).
Graphs show how different types of insurance contracts work, specifically showing how
stochastic dominance is sometimes achieved by the insurance case compared to no insurance.
Finally, graphs also show how likely a farmer to not meet his /her consumption expenditures is.

142

5. CHAPTER V: Results and Analysis
This chapter presents the results to assess the benefits of a credit-insurance package for
bean farmers in Honduras.
5.1. Certainty Equivalent Ratios
The comparison analysis presented in this chapter consists of comparing the metrics
certainty equivalents (CE), certainty equivalent ratios (CER), and insolvency percentages (IP),
for each of the choices (i.e., combination of technology categories, and use of credit and
insurance) and villages. Similar to the probability distribution analysis, certainty equivalents and
insolvency percentages are presented for all villages. Then results are presented for each village
as well as comparisons of main aspects between villages.
Certainty equivalent ratios are introduced in this section to observe a percentage change
in CE for changes in variety package and/or credit-insurance package, for a certain type of
farmer. The CERs are ratios of CE for two levels of one factor, where the base level is placed in
the denominator. For example, the variety package CER is used to compare if the change in
variety package is perceived to be beneficial by the farmer. In this case, the traditional variety
package (base) is in the denominator and the modern variety package is in the numerator, for a
specific level of coverage, loading factor, CRRAC, and discount rate. However, to compare
across different levels of credit-insurance packages, the CERs are directly compared, ceteris
paribus. For example, farmers perceive a benefit for a certain level compared to another level of
coverage, loading factor, CRRAC, or discount rate, when the CER is higher for the former level.
These ratios are obtained by dividing specific certainty equivalents from Table 75 to Table 114
in the Appendices.

143

A specific example of a CER is the ratio of the CE of a moderately risk averse farmer
(CRRAC=2) from La Cienega using a non-mechanized land preparation system, using a modern
variety with high level of inputs, an actuarially fair insurance with a 75% coverage, a discount
rate of 7% and no credit, compared to the same farmer using a traditional variety with low level
of inputs, and without the use of insurance nor credit. The previous example is the CER for
change in variety package. The ratios are estimated for several types of credit-insurance
packages (i.e., combination of load, coverages, with and without credit) and farmer
characteristics (i.e., risk aversion, time preference rate (discount rate)). The base level within
each village and type of farmer (e.g., CRRAC=2 and discount rate of 7%) is the traditional
variety package (traditional variety, low level of input) with no credit and no insurance.
The input level certainty equivalent ratio is estimated by dividing the certainty equivalent
for the high input level by the low input level (i.e., base level) within the same type of variety
37

(i.e., traditional or modern variety), land preparation system , and village. The ratio is
estimated for several farmer characteristics (i.e., risk aversion, discount rate); however, this CER
is only estimated for the case of no insurance and no credit.
Pairwise comparisons are made between CERs across the levels of the characteristic of
interest for comparing insurance contract characteristics (i.e., coverages, loads), farmer
characteristics (i.e., discount rates, risk aversion), and credit need. For example, to compare the
effect of risk aversion on the perception if the change in modern variety package is beneficial,
the variety package certainty equivalent ratios would be compared across the two levels of risk
aversion.

37

Type of land preparation is not compared due to problems in the problem setup.
144

In the following sections, certainty equivalent ratios or comparisons between these ratios
are discussed to observe the effects of variety package, fertilizer, type of variety, insurance
contract characteristics, farmer characteristics, and use of credit. The end of the chapter
discusses the results for the two cases of interest previously described to asses if a credit
insurance package could increase the adoption of a modern variety package.
5.2. Effects of Variety Packages, Credit Insurance Packages, and Farmer
Characteristics.
The effects of variety packages, credit-insurance packages, and farmer characteristics are
discussed in a manner that shows the general trend of a change in the levels of each the packages
or characteristics. The discussion on the credit-insurance packages will provide insights into the
reasons why some combination of insurance characteristics and farmer characteristics do not
follow the general trend of increasing or decreasing CEs within the two cases of interest (i.e.,
risk averse farmer and not capital constrained, and risk averse farmer and capital constrained). It
will provide a discussion on interaction effects between those characteristics.
5.2.1. Modern variety package change
The effect of the change from a traditional variety to a modern variety package is
analyzed here only for the no credit and no insurance case. Throughout all villages, results show
(Tables 42-46) CERs equal to or higher than one for the change in variety package for
moderately risk averse farmers (i.e., CRRAC of 2). This means that there is a trend of increasing
CEs for moderately risk averse farmers in all villages when these farmers change from a
traditional variety package to a modern variety package, independent of discount rate.

145

Table 42. CE ratios of modern variety package over traditional variety package for each land preparation system and creditinsurance package in La Cienega.
No Credit
With Credit
Load
Land
Discount
Preparation
rate
CRRAC
1.0
1.4
1.0
1.4
No
System
(%)
Insurance
Coverage
Coverage
Coverage
Coverage
45%
75%
45%
75%
45%
75% 45% 75%
2

1.08

1.07

1.08

1.06

1.04

1.03

1.03

1.02

1.00

4

1.00

1.00

1.03

0.99

1.00

0.98

1.00

0.97

0.97

2

1.06

1.06

1.07

1.05

1.04

1.03

1.03

1.02

1.01

4

1.00

1.00

1.03

1.00

1.00

0.98

1.01

0.98

0.98

7
NonMechanized
14

146

Table 43. CE ratios of modern variety package over traditional variety package for each land preparation system and creditinsurance package in Talanga and Guaimaca.
No Credit
With Credit
Load
Land
Discount
Preparation
rate
CRRAC
1.0
1.4
1.0
1.4
No
System
(%)
Insurance
Coverage
Coverage
Coverage
Coverage
45%
75%
45%
75%
45%
75% 45% 75%
2

1.00

1.00

1.01

0.99

0.96

0.96

0.96

0.95

0.92

4

0.95

0.95

0.98

0.94

0.95

0.93

0.95

0.92

0.92

2

1.01

1.01

1.02

1.00

0.98

0.98

0.98

0.97

0.95

4

0.95

0.96

1.00

0.95

0.96

0.94

0.97

0.94

0.94

7
NonMechanized
14

147

Table 44. CE ratios of modern variety package over traditional variety package for each land preparation system and creditinsurance package in Alauca.
No Credit
With Credit
Load
Land
Discount
Preparation
rate
CRRAC
1.0
1.4
1.0
1.4
No
System
(%)
Insurance
Coverage
Coverage
Coverage
Coverage
45%
75%
45%
75%
45%
75% 45% 75%
2

1.03

1.03

1.04

1.02

1.00

0.98

0.99

0.97

0.95

4

0.96

0.97

1.00

0.96

0.97

0.94

0.97

0.94

0.93

2

1.03

1.03

1.05

1.02

1.01

1.00

1.00

0.99

0.97

4

0.97

0.98

1.02

0.97

0.98

0.96

0.98

0.95

0.95

7
NonMechanized
14

148

Table 45. CE ratios of modern variety package over traditional variety package for each land preparation system and creditinsurance package in Arauli and Chirinos.
No Credit
With Credit
Load
Land
Discount
Preparation
rate
CRRAC
1.0
1.4
1.0
1.4
No
System
(%)
Insurance
Coverage
Coverage
Coverage
Coverage
45%
75%
45%
75%
45%
75% 45% 75%
2

1.03

1.02

1.02

1.01

0.98

0.98

0.97

0.97

0.94

4

0.97

0.97

0.99

0.96

0.96

0.94

0.96

0.94

0.93

2

1.03

1.02

1.03

1.01

1.00

0.99

0.99

0.98

0.96

4

0.97

0.98

1.00

0.97

0.97

0.96

0.97

0.95

0.94

7
NonMechanized
14

149

Table 46. CE ratios of modern variety package over traditional variety package for each land preparation system and creditinsurance package in Villa de San Francisco and Sabaneta de Valle de Angeles.
No Credit
With Credit
Load
Land
Discount
Preparation
rate
CRRAC
1.0
1.4
1.0
1.4
No
System
(%)
Insurance
Coverage
Coverage
Coverage
Coverage
45%
75%
45%
75%
45%
75% 45% 75%
2

1.04

1.03

1.04

1.02

1.00

0.99

0.99

0.98

0.96

4

0.97

0.98

1.01

0.97

0.97

0.96

0.98

0.95

0.95

2

1.03

1.03

1.04

1.02

1.01

1.00

1.00

0.99

0.98

4

0.98

0.98

1.01

0.98

0.98

0.96

0.98

0.96

0.96

7
NonMechanized
14

150

However, the threshold set for a change in variety package to be considered beneficial is a CER
of 1.03. All moderately risk averse farmers consider it beneficial except for farmers in Talanga
and Guaimaca.
On the other hand, high risk averse farmers (i.e., CRRAC of 4) consistently have CERs
38

less than one

throughout all villages. The certainty equivalent ratios (CER) for high risk

averse farmers range from 0.95 to 0.99. These two results together show that there is a chance to
improve farmers wellbeing by introducing a modern variety package with a credit-insurance
package for risk averse farmers.
In general, the increase in CER indicates that there is an incentive to change to the
modern variety package when the increase is more than modest (i.e., greater than or equal to 3%).
A threshold of a CER of 1.03 for a change in variety package is reasonable because of the
transaction costs of changing to the new variety package are not included. However, the change
in the variety package bundled with a specific credit-insurance package depends on the type of
farmer case (i.e., risk averse and not capital constrained, and risk averse and capital constrained).
Results in general show that without the use of insurance, all high risk averse farmers
throughout the villages would not adopt the modern variety package because of the higher risk
involved in the modern package. However, most moderately risk averse farmers would adopt the
package without the use of insurance.
5.2.2. Use of insurance
The use of insurance improves only a few farmers’ wellbeing, especially insurance
contracts with high coverage (i.e., 75%) with an actuarially fair premium. For instance, all

38

Cells in La Cienega show a value of one due to rounding.
151

farmers that do not need credit for using a modern variety package purchasing an actuarially fair
insurance contract with a 75% coverage are at least as well off as with a modern variety package
without insurance in all villages, except for one case (i.e., in Arauli and Chirinos), and usually
better off. This is observed by looking at all CERs greater than or equal to the CERs with no
insurance. The modern variety package with an actuarially fair insurance contract of 45%
coverage makes many cases of farmers at least as well off. However, for several cases of
moderately risk averse farmers with low discount rate (i.e., 7%) and one case for high discount
rate (i.e., 14%), their wellbeing is reduced with that variety and insurance package. This is an
unusual result because an actuarial fair contract should always be preferred by a risk averse
farmer to not having the contract, irrespective of the time frame (i.e., one year vs. many years).
However, insolvency is possible in this simulations and this changes the expected result. When
insolvency is possible, indemnity payments are not always balanced with the premium paid since
those that go insolvent do not receive the indemnity, and makes the change in variety package
not attractive. In general, the higher coverage makes the change in variety possible because
indemnity payments are made more often, and these indemnity payments received compensate
for those that go insolvent, except for the case in Arauli and Chirinos. However, the higher
coverage makes the distribution have a lower variability.
In general, the increase in CER enough to justify a change in variety package with an
insurance contract with a subsidized premium happens in several cases in La Cienega, Alauca
and Villa de San Francisco and Sabaneta de Valle de Angeles, and only for some cases in other
villages.
In other ways, a small change from using a modern variety package to a modern variety
package with insurance, subsidized or not, might just be worth it for some farmers throughout all

152

villages since the change from one bundle to the other is sometimes just one percentage point
above one. However, the only change is the insurance, which might not carry much transaction
costs and one percent above unity might be enough for some farmers.
5.2.2.1.

Insurance premiums

CER for a variety package change varies according to the insurance premium level. As
expected, a subsidized insurance contract (i.e., an actuarially fair premium) increased the CE for
the variety package change more than the moderately priced insurance contract (i.e., insurance
contract with load of 1.4) in all villages, independent of use of credit, type of farmer and
coverage of the insurance contract. This is expected since the former insurance contract covers
the same risk at a lower cost for the farmer. This is observed by higher CERs for the subsidized
premium compared to the moderately priced insurance premium.
When insurance is not actuarially fair (i.e., load greater than 1), it would not be unsusual
to observe that the risk transfer is not enough to compensate farmers, who do not need credit, for
the higher cost of the insurance. This is observed in CERs, of a variety package change with
insurance contracts with load of 1.4 and 75% coverage, being mostly lower than CERs of a
variety package change without insurance. Furthermore, for farmers in all villages is usually the
case that insurance does not compensate the risk reduction for the higher cost, except for three
cases (i.e., for a high risk averse farmer in Talanga and Guaimaca with a discount rate of 14%;
for a high risk averse farmer in Alauca with both discount rates). The cases where insurance does
not compensate farmers are the cases where farmers perceive the insurance contract to be too
costly for the combination of the risk the insurance transfers and the risk aversion of the farmer.
The exceptions listed above, where insurance still compensate farmers’ risk transfer, are due to
farmers high risk aversion.

153

5.2.2.2.

Coverages

CERs for a variety package change vary according to the insurance coverage level.
Farmers in villages find the change in variety packages appealing, depending on the coverages
and the premiums. That is, for the specific case of an actuarially fair premium, a farmer would
want to fully insure (i.e., have full coverage (100%)) and would reduce its coverage with a less
than actuarially fair premium. For example, the CERs for cases of actuarially fair premiums and
coverages of 75% are at least as high as the 45% coverage insurance contract counterparts, and
usually higher, except for one case where farmers that need credit in Arauli and Chirinos.
However, most of the cases of not actuarially fair insurance contracts have CERS higher in the
45% coverage contract compared to the 75% coverage, except for some high risk averse farmers
in several villages (e.g., La Cienega). However, this does not mean that the lower coverage will
always be preferred to the high coverage level, but it will depend on the level of the load and of
the coverage. The result seen for moderately risk averse is the opposite for high risk averse
farmers. The result is reversed due to the need of a higher coverage because of a combination of
a modern variety package being riskier than the traditional variety package and the higher risk
aversion.
5.2.3. Input level
The input level is analyzed by observing the ratios of CEs of high input use over low
input use within the same type of farmer (i.e., same discount rate and CRRAC), same variety,
and without insurance and credit. The input level refers specifically to fertilizer use. In the
Cienega, the CERs for the change in input use level indicate that the change in fertilizer from a
low level to a high level is preferred by moderately risk averse farmers, independent of farmers’
discount rate and type of variety (Table 47). This does not happen in the same way for farmers

154

in other villages. For farmers in other villages, the change in input level use is also preferred by
most moderately risk averse farmers; however, it is preferred only for the traditional variety. This
different result between varieties is due to the mix of a constant increase in yield for both
varieties by the additional fertilizer applied and to the higher risk in modern varieties. For
moderately risk averse farmers, the CERs range from 0.981 to 1.046. CERs for farmers in La
Cienega are at least as high for traditional varieties as for modern varieties. In addition,
moderately risk averse farmers have higher CERs than high risk averse farmers due to the risk
aversion effect by farmers reacting to a higher risk involved in higher fertilizer use. In other
villages, most high risk averse farmers with modern varieties have CERs less than one. This
means that farmers would prefer not to increase the quantity of fertilizer used within the modern
variety, without the use of insurance. In general, this is because of the thicker lower tails of yield
probability distributions due to the higher downside risk of higher input use and modern variety
use (i.e., the more risky choice in variety use). The differences in villages are due to bean price
differences among villages. This situation shows how insurance is important in some situations.
The trend of higher CERs for higher input levels is practically eliminated in all other
villages (Table 47, 48 and 49), except for a few cases. The increase in CE in all other villages
due to the input change causes a change in CER of at most 1.018 and many less than 1. The
difference in CERs being higher in La Cienega is due to the higher price of beans in La Cienega
that makes the marginal revenue higher, compared to the marginal cost of fertilizer. This higher
marginal revenue in La Cienega results in a higher gross income (i.e., assuming a linear
production function) most of the time, compared to other villages.

155

Table 47. CE ratios of high input use over low input use for each land preparation system and type of farmer (i.e., risk
aversion and discount rate) in La Cienega, and Talanga and Guaimaca.

Land
Preparation
System

Variety and
Input Package

La Cienega

Talanga and Guaimaca

Discount Rate

Discount Rate

7%
CRRAC
2.0
4.0

14%
CRRAC
2.0
4.0

7%
CRRAC
2.0
4.0

14%
CRRAC
2.0
4.0

Traditional
Variety

1.046

1.003

1.037

1.002

0.999

0.968

1.002

0.974

Modern
Variety

1.017

0.989

1.014

0.989

0.990

0.968

0.993

0.971

Non-Mechanized

156

Table 48. CE ratios of high input use over low input use for each land preparation system and type of farmer (i.e., risk
aversion and discount rate) in Alauca and Arauli and Chirinos.

Land
Preparation
System

Variety and
Input Package

Alauca

Arauli and Chirinos

Discount Rate

Discount Rate

7%
CRRAC
2.0
4.0

14%
CRRAC
2.0
4.0

7%
CRRAC
2.0
4.0

14%
CRRAC
2.0
4.0

Traditional
Variety

1.007

0.973

1.009

0.980

1.000

0.972

1.002

0.977

Modern
Variety

0.981

0.961

0.989

0.967

0.979

0.961

0.985

0.965

Non-Mechanized

157

Table 49. CE ratios of high input use over low input use for each type of farmer (i.e., risk aversion and discount rate) in Villa
de San Francisco and Sabaneta de Valle de Angeles.

Discount Rate
Land Preparation System

Variety and Input
Package
7%

14%

CRRAC

CRRAC

2.0

4.0

2.0

4.0

Traditional Variety

1.018

0.982

1.015

0.985

Modern Variety

0.991

0.970

0.995

0.973

Non-Mechanized

158

5.2.4. Type of variety
Type of variety is analyzed by observing the ratios of CEs of modern varieties over
traditional varieties within the same type of farmer (i.e., same discount rate and CRRAC), and
same use of input level without the use of insurance or credit (i.e., Table 50, 51, and 52). The
change to a modern variety is preferred by all moderately risk averse farmers throughout all
villages. However, the benefits are only modest within the high input use. For several cases
with high risk averse farmers the change is not perceived to be beneficial.
The change in variety is perceived to be most beneficial within the low input use. This is
because the increase in yield due to the change in variety is the same for both input levels and the
high input use is riskier and also has a higher base value, therefore, a lower relative change. The
time preference rate also affects the change in variety. In La Cienega, the 7% discount rate has
higher CERs than the 14% discount rate, while in Talanga and Guaimaca is reversed. In other
villages, this result is mixed depending on the CRRAC. The reason for this is explained in the
time preference section below. The trend for the change in variety continues in the same manner
as from the change in input use, where the change is preferred by moderately risk averse farmers
than to high risk averse farmers and preferred within the low input use (i.e., the less risky choice
in input use). Again, this shows how insurance might benefit some farmers depending on their
risk aversion, the riskiness of the change in variety and the insurance contract.

159

Table 50. CE ratios of modern variety use over traditional variety use for each type of farmer (i.e., risk aversion and discount
rate) in La Cienega, and Talanga and Guaimaca.

Land
Preparation
System

Input Level

La Cienega

Talanga and Guaimaca

Discount Rate

Discount Rate

7%
CRRAC
2.0
4.0

14%
CRRAC
2.0
4.0

7%
CRRAC
2.0
4.0

14%
CRRAC
2.0
4.0

Low input use

1.059

1.007

1.048

1.006

1.013

0.977

1.015

0.983

High input use

1.029

0.993

1.025

0.993

1.004

0.977

1.006

0.980

Non-Mechanized

160

Table 51. CE ratios of modern variety use over traditional variety use for each type of farmer (i.e., risk aversion and discount
rate) in Villa de San Francisco and Sabaneta de Valle de Angeles, and Alauca
Villa de San Francisco and Sabaneta de
Alauca
Valle de Angeles
Land
Preparation
System

Discount Rate

Input Level

7%
CRRAC
2.0
4.0

Discount Rate

14%
CRRAC
2.0
4.0

7%
CRRAC
2.0
4.0

14%
CRRAC
2.0
4.0

Low input use

1.049

1.002

1.040

1.003

1.052

1.000

1.043

1.004

High input use

1.020

0.989

1.019

0.991

1.025

0.987

1.023

0.991

Non-Mechanized

161

Table 52. CE ratios of modern variety use over traditional variety use for each type of farmer (i.e., risk aversion and discount
rate) in Arauli and Chirinos.
Arauli

Land Preparation System

Discount Rate

Input Level
7%

14%

CRRAC

CRRAC

2.0

4.0

2.0

4.0

Low input use

1.050

1.005

1.041

1.007

High input use

1.028

0.994

1.024

0.994

Non-Mechanized

162

5.2.5. Farmers’ risk aversion and rate of time preferences
The risk aversion level and time preference rate affect the CERs for variety package
change. CERs for variety package changes, with and without insurance, are higher for medium
risk averse farmers than for high risk averse farmers. This is because high risk averse farmers
would be willing to pay more than medium risk averse farmers for a certain pay (i.e., lower CE)
within the low risk technology (i.e., traditional variety package), and would pay even more for a
certain pay (i.e., lower CE) within a high risk technology (i.e., modern variety package). Thus,
given that modern varieties have a higher wealth expected value, CERs are lower for high risk
averse farmers.
The time preference or discount rate has different effects on CERs depending on the signs
(i.e., positive or negative) of the streams of gross income. When most of the outcomes are
positive, increases in the rate of time preference, or discount rate, reduces the CERs for a variety
package change because the streams of gross income in the future get discounted more.
However, if many of the outcomes in the future are negative, the negative impact is reduced by
the higher discount rate. These two effects are observed in the higher CERs in La Cienega for
low discount rates, compared to high discount rates, but nearly about the same values or lower in
CERs in other villages. Most of the positive outcomes in La Cienega are due to La Cienega’s
higher bean price.
5.2.6. Use of credit and insurance
CERs have shown that the variety package change is warranted in all villages for
moderately risk averse farmers; however, the use of credit and insurance is only warrated for a
moderately risk averse farmer in La Cienega for the case of a subsidized insurance contract. In
other ways, farmers in Honduras are, for the most part, highly urged by banks to obtain crop

163

insurance to access credit. Thus, for farmers who need credit to change the variety package from
a traditional variety package to a modern variety package would need to purchase insurance.
However, results in these simulations show that many farmers who need credit would not opt for
the change in variety package even with insurance. This can be observed in CERs lower than
one for the variety package change for a farmer that needs credit and uses insurance. The
interest paid on the credit reduces the CERs. Farmers in La Cienega would perceive the change
in variety package beneficial when using credit and insurance, while farmers in other villages
would not. The price difference among villages is the reason for the discrepancy.
As it is expected, an insurance contract with a load of 1.4 does have lower CERs than the
actuarially fair premium, and this effect is more evident for farmers that need credit. However,
one of the hypotheses of this study is that a credit-insurance contract would increase farmer
adoption of a modern variety package, and the results observed within farmers that need credit
and with a reasonably priced insurance contract are not in accordance to this hypothesis. All risk
averse farmers are worst off with the modern variety package together with the insurance
contract with 75% coverage and load of 1.4, than the better of the traditional variety package
with no insurance and modern variety package without insurance. This means that it is better for
moderately risk averse farmers who need credit to not access credit and insurance to change
variety package because the mix of interests paid and high premium is not worth compared to the
risk transferred and transaction cost involved.
Contrary to many cases discussed above, high risk averse farmers in La Cienega who do
not need credit with a subsidized insurance would make the change in variety package. The
hypothesis is only partially true for the villages involved, type of insurance contract and type of
farmer specified above.

164

5.2.7. Price of beans
The price of beans has an impact on CERs. This impact has been observed in the
differences in CERs among villages. Higher gross income make the change in modern variety
package less sensitive to changes in cost from variety package change, changes in farmer
discount rate, and cost of credit. La Cienega is the village with the highest price and Arauli and
Chirinos are the villages with the lowest price. The Cienega is the village with highest CERs,
however, Arauli and Chirinos does not have the lowest CERs; Talanga and Guaimaca has the
lowest CERs. This is because the relative increase in gross income is higher in Arauli and
Chirinos due to a lower base yield of the traditional variety package, while the increase in yield
due to the variety package change is the same, and prices are about same for those villages. Thus,
the ratios of CERS would be higher for Arauli and Chirinos. La Cienega have higher bean prices
than Talanga and Guaimaca, which results in higher gross income. The higher gross income in
La Cienega make the change in modern variety package less sensitive to changes in cost from
variety package change, changes in farmer discount rate, and cost of credit.
5.3. Percentage Chance of Going Insolvent
This section will compare the percentage chance of going insolvent (PCGI) for each of
the choices for all villages. The percentages of chance of going insolvent for all villages are
presented from Table 53 to Table 57.
The change in variety package, from a traditional to a modern variety package, increases
the risk, PCGI, because of a thicker lower tail of the modern variety package. This is seen in all
villages in Tables 53 through Table 57, where the modern variety package shows at least 0.74
percentage points of PCGI above the traditional variety package, and up to 5.54 percentage
points.

165

In general, insurance does reduce the chance of going insolvent, as long as it has a large
enough coverage and is not priced to high. PCGI is lower for farmers that use actuarially fair
insurance contracts and high coverage, independent of the need for credit in all villages, except
in Arauli and Chirinos. Arauli and Chirinos are the villages with the lower bean price, which
would explain why even with insurance these would have higher values of PCGI.
As expected, PCGI values in all villages are lower for high coverages compared to low
coverages for actuarially fair insurance contracts. However, these values are higher for
moderately priced insurance contracts. This result is due to the decapitalization of the farmer
through the years because of the payment of the loaded premium.
One unexpected result is that an insurance contract with an actuarially fair premium and a
low coverage has a higher percentage chance of going insolvent than with no insurance. This is a
counterintuitive result because the actuarially fair insurance contract should protect more farmers
from going insolvent than with no insurance. For one year shot, the actuarially fair insurance
contract does protect more farmers; however, with a longer year sequence the result changes.
This result is because the premium reduces capital from the farmer while only in a few years the
farmer gets indemnized. .
Results of PCGI are not always lower for no credit, which initially one would expect
because of the interest payments. There are several cases where the corresponding comparisons
of credit are having lower PCGI. This is because of a mix of coverage level, bean price and the
switch to traditional variety when the farmer goes below the threshold of being credit-worthy.
The higher coverage level decreases the PCGI. Also, the lower bean price villages (i.e., Arauli
and Chirinos, and Talanga and Guaimaca) have the farmers that need credit with higher PCGI
than those that do not need credit, independent of the coverage or load. However, the village

166

with higher bean price (i.e., La Cienega) has PCGI values lower for farmers that do not need
credit compared to those who do. Villa de San Francisco and Sabaneta de Valle de Angeles, the
villages with a bean price somewhere in the middle, switch depending on the coverage level and
the need for credit.
As expected, subsidized insurance reduces PCGI more than a moderately priced
insurance contract because of a lower premium.
The PCGI throughout all villages are in a range from 31.00% to 40.04%. These levels
are not within a normal range of other studies (Leatham, et.al., 1987).
It should be noted that all farmers who find the change in variety packages appealing, in
terms of CERs, have already taken into consideration the higher risk and higher cost of a modern
variety package when coupled with insurance, and still find the change valuable. This would
apply for the case of subsidized insurance in La Cienega.

167

Table 53. Percentage chance of going insolvent for each land preparation system and credit-insurance package in La Cienega.
No Credit
With Credit
Load
Land
Variety
Preparation and Input
1.0
1.4
1.0
1.4
No
System
Package
Insurance
Coverage
Coverage
Coverage
Coverage
45%
75%
45%
75%
45%
75%
45%
75%
Traditional
Variety
Low Input

32.64

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

Modern
Variety
High Input

33.38

33.90

32.58

34.56

35.56

33.78

32.14

34.52

35.22

NonMechanized

168

Table 54. Percentage chance of going insolvent for each land preparation system and credit-insurance package in Talanga
and Guaimaca.
No Credit
With Credit
Load
Land
Variety
Preparation and Input
1.0
1.4
1.0
1.4
No
System
Package
Insurance
Coverage
Coverage
Coverage
Coverage
45%
75%
45%
75%
45%
75%
45%
75%
Traditional
Variety
Low Input

27.26

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

Modern
Variety
High Input

32.36

32.88

31.46

33.62

35.56

33.94

31.72

34.64

36.24

NonMechanized

169

Table 55. Percentage chance of going insolvent for each land preparation system and credit-insurance package in Villa de San
Francisco y Sabaneta de Valle de Angeles.
No Credit
With Credit
Load
Land
Variety
Preparation and Input
1.0
1.4
1.0
1.4
No
System
Package
Insurance
Coverage
Coverage
Coverage
Coverage
45%
75%
45%
75%
45%
75%
45%
75%
Traditional
Variety
Low Input

31.00

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

Modern
Variety
High Input

34.32

34.64

33.48

35.42

37.24

34.70

32.64

35.26

37.04

NonMechanized

170

Table 56. Percentage chance of going insolvent for each land preparation system and credit-insurance package in Alauca.
No Credit
With Credit
Load
Land
Variety
Preparation and Input
1.0
1.4
1.0
1.4
No
System
Package
Insurance
Coverage
Coverage
Coverage
Coverage
45%
75%
45%
75%
45%
75%
45%
75%
Traditional
Variety
Low Input

26.28

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

Modern
Variety
High Input

31.82

32.10

29.44

33.22

34.10

33.30

30.92

34.14

35.42

NonMechanized

171

Table 57. Percentage chance of going insolvent for each land preparation system and credit-insurance package in Arauli and
Chirinos.
No Credit
With Credit
Load
Land
Variety
Preparation and Input
1.0
1.4
1.0
1.4
No
System
Package
Insurance
Coverage
Coverage
Coverage
Coverage
45%
75%
45%
75%
45%
75%
45%
75%
Traditional
Variety
Low Input

30.40

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

N.A.

Modern
Variety
High Input

35.06

35.82

35.08

36.78

39.50

36.50

35.42

37.64

40.04

NonMechanized

172

5.4. Results for Cases of Interest
The discussion on the use of a credit-insurance package for adopting a modern variety
package will be presented for the two cases of interest: 1) for a risk averse and not capital
constrained farmer, and 2) risk averse and capital constrained farmer.
5.4.1. Results for a risk averse and not capital constrained farmer
High risk averse farmers that do not need credit would be the only farmers that would
benefit from insurance use for adoption of the modern variety package because moderately risk
averse farmers already consider the change in variety package beneficial. Farmers that do not
need credit in all villages would not purchase insurance to adopt the modern variety package,
except for high risk averse farmers in La Cienega. The risk protection by the insurance is only
modest or just not enough to compensate for the risk transferred for high risk averse farmers
compared to its base practice of using a traditional variety package. CERs for highly risk averse
farmers that do not need credit and use insurance are at most one or two percentage points above
39

the CER of the change in variety package alone or less than 1.03 , except for a high risk averse
farmer in La Cienega. The exception applies only to the subsidized insurance premium (i.e.,
load=1) case. This can be observed with the CERs of Table 42 which are lower, or modestly
higher, than the CER of the variety change only; also observe CE of choices 1 and 5 from Table
4, in Tables 75-114 in Appendices.
The difference in incentive to change variety packages for farmers in La Cienega,
compared to farmers in other villages, is the higher price of beans in La Cienega.

39

The two conditions were established for the change in variety package to be considered
beneficial.
173

PCGI is decreased by higher insurance coverage and increased by changing variety
package from a traditional variety package to a modern variety package.
In summary, the results show that for all cases of reasonably priced insurance contract
and even for most cases of the subsidized insurance contract, the insurance package would not
help increase the adoption of modern varieties for all farmers that do not need credit in all
villages, except for La Cienega.
5.4.2. Results for a risk averse and capital constrained farmer
Contrary to a not capital constrained farmer, both the high or moderate risk averse type of
farmer could be benefited by the use of insurance because the change in variety package is only
achieved through credit. This means that the comparison of the CER is with only the threshold
of 1.03 and not with both conditions established before.
Farmers in all villages that need credit to aquire the modern variety package would not
acquire the credit-insurance package to adopt the modern variety package, except for moderately
risk averse farmers in La Cienega. Moderately risk averse farmers in La Cienega with capital
constraints would benefit from the credit-insurance package to change the variety package, as
long as the insurance is subsidized with a coverage of 75%. Farmers in all other villages would
not be benefited even when subsidized. This result is observed because the interest paid for the
credit, above any modest improvements of the use of insurance, is just too high for farmers to
make the change. A moderately risk averse farmer in La Cienega who needs credit to purchase
the modern variety and inputs would benefit from the credit by being able to purchase the inputs
and from the insurance by reducing the downside risk. The credit-insurance package benefit is
shown in Table 42, where there are CER values of 1.03 for the subsided insurance contract.
Moderately priced insurance would not benefit farmers in any of the villages for farmers that are

174

capital constrained. The change in variety package with capital constrained farmers might be
possible if the bean price increases or other parameters change so as to increase the gross income.
As in the previous case of interest, PCGI is decreased by higher insurance coverage and
increased by changing variety package from a traditional variety package to a modern variety
package.
Again, the difference in incentive to change variety packages for farmers in La Cienega,
compared to farmers in other villages, is the higher price of beans in La Cienega.
5.5. Chapter Summary
The metrics certainty equivalent (CE), certainty equivalent ratios (CER) and insolvency
percentages were used to compare changes in technology and type of insurance contract for each
type of farmer and village. The changes compared were the variety package change as a whole
as well as the individual elements of the variety package, levels of different credit-insurance
contract characteristics, and differences among villages. The previous is done to answer if the
credit-insurance package increases the adoption of a modern variety package for two cases of
interest: 1) a risk averse farmer who is not capital constrained, and 2) a risk averse farmer that is
capital constrained. The use of a modern variety package alone, without a credit-insurance
package, improves the moderate risk averse farmers’ wellbeing except for farmers in Talanga
and Guaimaca. CER values for moderately risk averse farmers are equal or higher than 1.03, the
threshold established to change variety package. The variety package change includes the
change in input (i.e., fertilizer) plus the change in variety. Both changes carry more risk for the
farmer, thus the perception of increased wellbeing is higher for the moderately risk averse than to
the high risk averse.

175

Farmers in all villages that do not need credit to acquire the modern variety package
would not purchase insurance to adopt the modern variety package, except for high risk averse
farmers in La Cienega. Farmers in all villages that need credit to acquire the modern variety
package would not acquire the credit-insurance package to adopt the modern variety package,
except for moderately risk averse farmers in La Cienega. This latter moderately risk averse
farmers would benefit from the credit-insurance package if it is actuarially fair with a 75%
coverage. The use of insurance, as expected, reduced the percentage chance of going insolvent
(PCGI) when the premium is actuarially fair and has high coverage. When it is a moderately
priced insurance contract, the PCGI is higher for a high coverage due to decapitalization of the
farmer.

176

6. CHAPTER VI: Conclusions and Recommendations
This section provides the conclusions based on survey results, budgeting, and stochastic
simulation runs and recommendations for future research.
Simulations in this study are modeled as a steady state since price and yield have the
same mean through time. It also does not facilitate investment so that growth is incorporated.
Finally, the study is strongly dependent on the shape of the density functions of price and yield.
Moderate changes in this density functions would change the results of whom, if anyone,
benefits from the use of the credit-insurance package.
6.1. Conclusions


The change from a traditional variety to a modern variety is profitable. The

partial budget analysis shows that the increase in marginal revenue per hectare due to the change
in variety is higher than the marginal costs. Moreover, the change from a traditional variety
package (i.e., traditional variety and low levels of inputs) to a modern variety package (i.e.,
modern variety and high levels of inputs) is even more profitable because the marginal revenue
due to the use of fertilizer is greater than the fertilizer marginal cost. Profitability holds
throughout all the villages although with varying magnitudes, mainly because of bean price
differences among villages.


Bean farmers face a moderate bean and corn price risk with price distributions

with a coefficient of variation (CV) of 20%. Time series data at the market level in the capital
city of Tegucigalpa revealed that beans have approximately a 19% CV, and corn has a 16% CV,
which were adjusted due to differences of CVs from the market level and the farm level, and
because of experiences seen at other studies.

177

Farmers face a perceived yield risk at a range of approximately 32-34% CV for
beans, depending on the bean technology, and 32% for traditional corn varieties. Aggregated
data at the regional level in Honduras presented yield distributions with CVs of 14% for beans
and 9% for corn. However, aggregated data and farmer level data are very different. Based on
farmers elicited yield distributions and experiences seen at other studies the former
aforementioned CVs for the farmer level yield risk are appropriate.
Farmer yield distribution elicitation was not successful in this study. Many
improper probability distributions were obtained from the elicitation as seen from the shape of
the distributions. The elicitation method, and enumerator training and bias were part of the
problem of obtaining the improper distributions. The visual impact method without some safety
guards is not appropriate for farmers with low schooling and enumerators with few or no
experiences in yield distribution elicitation. Enumerators need intensive training on the
elicitation procedure. More pre-testing of the elicitation method with more funding would have
resulted in better estimation of yield density functions.
In any given year, farmers using a traditional variety package face an 11.0%
chance of not being able to meet their normal consumption expenditures without having to resort
to a mechanism to cope with shocks. Farmers’ cumulative density function at the level of half of
their consumption expenditures revealed this chance, given that farmers rely on gross income
from risky bean and corn production and assuming a stable income from other sources. Farmers
throughout the villages on average reduce modestly that chance to 10.5% in a given year by
changing to the modern variety package with a reasonably priced insurance contract for farmers
that are not capital constrained.

178



Sixty percent of farmers that do not have enough capital to balance their

consumption expenditures when perils occur rely mainly on working in another job besides their
crop enterprise (50%), selling their assets (24%), and asking for a loan from friends, family or
other person (14%).


Adoption of an improved variety package can occur when the change in package

is profitable, even without the use of insurance. However, price and yield risk can deter adoption
of the modern package even when it is profitable to do so. Some risk averse farmers would not
adopt a riskier variety package because of the effect bad outcomes has on their livelihoods.
Insurance is a risk management tool that allows risk averse farmers adopt technologies when
faced with the previous situation.


There is no literature on evaluations of existing insurance policies in Honduras

regarding their coverage and loading. Crop insurance companies in Honduras offer catastrophic
crop insurance policies (i.e., coverages of 45%) that do not provide high levels of risk transfer,
although the contract provides the means to grow another crop in case of a devastating peril
striking. However, the policies offered have unusually high insurance premium loads that range
from 3 to 5. A new crop insurance policy is proposed with 75% coverage and a load of 1.0 to
incentivize certain type of farmers to adopt modern bean varieties and increase their gross
income. This would need government intervention since no insurance company would sell
contracts with a load of one. However, this would not be unusual since some funds have been
allocated already in the National Development Bank in Honduras (BANADESA for its acronym
in Spanish) for purchasing insurance, when this institution lends to farmers.
A subsidized credit-insurance package would help increase the adoption of modern varieties for
those that do and do not need credit in La Cienega, however not for other villages.

179



Without the use of insurance, all high risk averse farmers throughout the villages

would not adopt the modern variety package because of the higher risk involved in the modern
package. However, most moderately risk averse farmers would adopt the modern package
without the use of insurance.
Farmers in all villages that do not need credit to change to a modern variety
package would not benefit from the insurance, except for high risk averse farmers in La Cienega.
The risk protection by the insurance is not enough to compensate for the risk transfer for high
risk averse farmers.
Farmers in all villages that do need credit to change to a modern variety package
would not adopt the modern variety package by using the credit-insurance package, except for
low risk averse farmers in La Cienega. The risk protection by the insurance is only modest and
not enough to compensate for the transaction costs involved in its acquisition for farmers in other
villages. The conditions for these farmers to adopt the modern variety package would be with
credit and a subsidized insurance contract with high coverage.
The bean price differential among villages is a determinant factor for adopting a
modern variety package in a risky setting with or without the use of insurance and/or credit.
Farmers under general conditions would be benefited by a subsidized creditinsurance package (i.e., load of 1.0) that enables them to change from a traditional variety
package (i.e., low fertilizer and traditional seed) to a modern variety package (i.e., high fertilizer
level and modern seed).
The specific conditions in which insurance would benefit is when farmers are
highly risk averse and do not need credit, living in villages with relatively high prices, or,
moderately risk averse farmers that need credit in villages with relatively high prices. The high

180

price has an effect on insurance demand or value to the farmer in those villages with high prices
because of the forgone expected gross income of the modern variety is high compared to those
farmers in villages with low prices.
Insolvency has an impact on the benefits of insurance. The modern variety
package was not preferred in some cases of low coverage, even when covered by an actuarially
fair insurance. Insolvency makes indemnity payments to not be balanced with the premium paid.
In an actuarially fair insurance, the premium is equal to the expected loss, so indemnity
payments received by the farmer through the years would be equal to the premiums paid by
him/her. However, farmers that go insolvent do not receive the indemnities from the time they
go insolvent and onward, while in the previous years the farmer has already paid the premium.
This mismatch in the quantity of indemnity payments made by the insurance company to the
farmer and premiums paid by the farmer would act as a load greater than one on the premium
and make the change in variety package not attractive.
PCGI is reduced by insurance with high coverage and increased by change in
variety package. Both, PCGI and CERs determine the demand for the modern variety package.
However, some unexpected results arise sometimes due to the mix of factors and their levels.
6.2. Limitations and Recommendations for Future Research
It is advisable to revisit the approach of estimating the yield probability distributions due
to the problems encountered; the general approach has been applied across a broad range of
studies. To avoid obtaining improper density functions, it is recommended to include one of two
practices when using the visual impact method for yield elicitation as well as intensive
enumerator training. The specific visual impact method used in this study provided tokens to
farmers for them to assign to a fixed number of intervals created from farmers expressed

181

minimum and maximum yields. The mutually exclusive practices added to that method are: 1)
anchor one point by asking the most likely yield; after this value is established, the frequency of
the most likely relative to the number of counters must be established; then, counters must be
placed on the next to most likely yield to the right, and next to most likely yield to the left, and
so on, by asking what is the relative frequency of the interval being filled compared to the most
likely interval; finally, farmers must be given a chance to rearrange the counters; and 2) after
doing several pre-tests, fix the endpoints for all farmers and then let farmers fill out the yield
distribution. These processes will increase the understanding of the concept of relative frequency.
Careful preparation of enumerators is needed. Five day training session should be
devoted to enumerators on visual impact method, including three days of field pre-testing. Focus
groups with farmers would be preferred to use for each of the technology categories, when
farmers are of low schooling.
To determine why farmers who do not need credit would not switch from a traditional
bean variety package to a modern bean variety, it is essential to obtain their knowledge and
perception of modern varieties regarding yields for both modern and traditional varieties. Also,
the cost structure of these farmers in the case of switching to a modern variety is helpful to
understand their motives of why are they not using modern varieties. To corroborate or adjust
the yield probability distributions, it would be helpful to organize focus groups to elicit their
yield probability distributions, cost structure for traditional and modern varieties, creating
scenarios of disease pressure and drought, and obtaining the frequency of the scenarios. Thus,
researchers should quantify the perception of losses due to disease vulnerability together with
rainfall vulnerability to create yield distributions for these common cases of crop loss and better
understand the choices farmers make.

182

To encourage more farmers to obtain crop insurance, coverage levels need to increase
with a commensurable premium (i.e., a premium that will also increase suitably) for farmers to
perceive appropriate risk transfer and also to reduce the farmers that go insolvent. More efforts
are needed on the design of contracts to reduce the effect of moral hazard and adverse selection,
which affect crop insurance premiums, to reduce the insurance premiums. For example,
collection of yield data at the municipality level to develop area yield insurance contracts of on
most productive areas of Honduras. Also, making these data accessible to public, as it is done
for rainfall data, which enables to develop weather-index based insurance contracts. The latter
type of contracts may need placement of rainfall stations in the most productive areas, which has
been one of the constraints of establishing weather-index crop insurance. In addition, doing more
research on weather-index insurance, for example, the use of satellite data do develop contracts,
would help reduce insurance premiums.

183

APPENDICES

184

Appendix 1. Determining yield probability distributions

Table 58. Probability Distribution Analysis of Variance for Traditional Varieties,
Honduras 2007.
Source
Type III
Degrees
Mean
F-value
p-Value
Sum of
of
Square
Squares
Freedom
Corrected Model
3.562(a)
319
.011
4.323
.000
Intercept
14.998
1
14.998 5807.187
.000
Village
.000
9
.000
.000
1.000
Interval
.611
4
.153
59.112
.000
Enumerator
.000
11
.000
.000
1.000
Village*Interval
.151
36
.004
1.619
.013
Village*Enumerator
.000
43
.000
.000
1.000
Interval*Enumerator
.510
44
.012
4.487
.000
Village*Interval*Enumerator
.807
172
.005
1.816
.000
Error
2.208
855
.003
Total
52.770
1175
Corrected Total
5.770
1174
(a) R Squared = .617 (Adjusted R Squared = .475)
Source: Pejuan Bean Farmer Survey, 2007.

185

Table 59. Probability Distribution Analysis of Variance for Modern Varieties, Honduras
2007.
Source
Type III
Degrees
Mean
F-value
p-Value
Sum of
of
Square
Squares
Freedom
Corrected Model
1.650(a)
244
.007
2.684
.000
Intercept
8.876
1
8.876 3522.206
.000
Village
.000
9
.000
.000
1.000
Interval
.199
4
.050
19.739
.000
Enumerator
.000
10
.000
.000
1.000
Village*Interval
.204
36
.006
2.247
.000
Village*Enumerator
.000
29
.000
.000
1.000
Interval*Enumerator
.327
40
.008
3.247
.000
Village*Interval*Enumerator
.647
116
.006
2.214
.000
Error
.693
275
.003
Total
23.143
520
Corrected Total
2.343
519
(a) R Squared = .704 (Adjusted R Squared = .442)
Source: Pejuan Bean Farmer Survey, 2007.

186

Table 60. Probability Distribution Analysis of Variance for Traditional Varieties,
Honduras 2007.
Source
Type III Degrees of Mean
F-value
p-value
Sum of
Freedom
Square
Squares
Corrected Model
3.372(a)
416
.008
3.511
.000
Intercept
3.385
1
3.385 1466.234
.000
Labor Cost per ha
.000
1
.000
.000
1.000
Input Cost per ha
.000
1
.000
.000
1.000
Village (Vill)
.000
9
.000
.000
1.000
Interval (Int)
.364
4
.091
39.461
.000
Enumerator (Enum)
.000
11
.000
.000
1.000
Land Preparation System
.000
2
.000
.000
1.000
(LPS)
Vill*Int
.145
36
.004
1.739
.006
Vill*Enum
.000
35
.000
.000
1.000
Int*Enum
.367
44
.008
3.615
.000
Vill*Int*Enum
.455
140
.003
1.409
.004
Vill*LPS
.000
6
.000
.000
1.000
Int*LPS
.036
8
.005
1.956
.050
Vill*Int*LPS
.062
24
.003
1.124
.312
Enum*LPS
.000
9
.000
.000
1.000
Vill*Enum*LPS
.000
1
.000
.000
1.000
Int*Enum*LPS
.103
36
.003
1.244
.160
Vill*Int*Enum*LPS
.017
4
.004
1.818
.124
Error
1.150
498
.002
Total
41.122
915
Corrected Total
4.522
914
a R Squared = .746 (Adjusted R Squared = .533)
Source: Pejuan Bean Farmer Survey, 2007.

187

Table 61. Probability Distribution Analysis of Variance for Modern Varieties, Honduras
2007.
Source
Type III
Degrees
Mean
F-value
p-value
Sum of
of
Square
Squares
Freedom
Corrected Model
.314(a)
76
.004
.872
.680
Intercept
.018
1
.018
3.897
.061
Labor Cost per ha
.000
1
.000
.000
1.000
Input Cost per ha
.000
1
.000
.000
1.000
Village (Vill)
.000
2
.000
.000
1.000
Interval (Int)
.024
4
.006
1.272
.310
Enumerator (Enum)
.000
7
.000
.000
1.000
Land Preparation System
.000
1
.000
.000
1.000
(LPS)
Vill *Int
.031
8
.004
.831
.584
Vill *Enum
.000
0
.
.
.
Int*Enum
.206
28
.007
1.552
.143
Vill*Int*Enum
.000
0
.
.
.
Vill*LPS
.000
0
.
.
.
Int*LPS
.005
4
.001
.254
.904
Vill*Int*LPS
.000
0
.
.
.
Enum*LPS
.000
0
.
.
.
Vill*Enum*LPS
.000
0
.
.
.
Int*Enum*LPS
.000
0
.
.
.
Vill*Int*Enum*LPS
.000
0
.
.
.
Error
.109
23
.005
Total
4.422
100
Corrected Total
.422
99
a R Squared = .742 (Adjusted R Squared = -.109)
Source: Pejuan Bean Farmer Survey, 2007.

188

Table 62. Probability Distribution Analysis of Variance for Traditional Varieties
(enumerator five removed), Honduras 2007.
Source
Type III Degrees
Mean
F-value
p-value
Sum of
of
Square
Squares
Freedom
Corrected Model
3.176(a)
366
.009
3.808
.000
Intercept
3.160
1
3.160 1386.515
.000
Labor Cost per ha
.000
1
.000
.000
1.000
Input Cost per ha
.000
1
.000
.000
1.000
Village (Vill)
.000
8
.000
.000
1.000
Interval (Int)
.354
4
.089
38.872
.000
Enumerator (Enum)
.000
10
.000
.000
1.000
Land Preparation System
.000
2
.000
.000
1.000
(LPS)
Vill *Int
.167
32
.005
2.288
.000
Vill*Enum
.000
29
.000
.000
1.000
Int*Enum
.328
40
.008
3.596
.000
Vill*Int*Enum
.368
116
.003
1.394
.009
Vill*LPS
.000
6
.000
.000
1.000
Int*LPS
.048
8
.006
2.613
.008
Vill*Int*LPS
.062
24
.003
1.139
.296
Enum*LPS
.000
7
.000
.000
1.000
Vill*Enum*LPS
.000
1
.000
.000
1.000
Int*Enum*LPS
.079
28
.003
1.238
.190
Vill*Int*Enum*LPS
.017
4
.004
1.842
.120
Error
1.010
443
.002
Total
36.586
810
Corrected Total
4.186
809
(a) R Squared = .759 (Adjusted R Squared = .560)
Source: Pejuan Bean Farmer Survey, 2007.

189

Table 63. Probability Distribution Analysis of Variance for Modern Varieties (enumerator
five removed), Honduras 2007.
Source
Type III
Degrees of Mean
F-value
p-value
Sum of
Freedom
Square
Squares
Corrected Model
.254(a)
61
.004
.745
.805
Intercept
.012
1
.012
2.146
.160
Labor Cost per ha
.000
1
.000
.000
1.000
Input Cost per ha
.000
1
.000
.000
1.000
Village (Vill)
.000
1
.000
.000
1.000
Interval (Int)
.013
4
.003
.601
.666
Enumerator (Enum)
.000
5
.000
.000
1.000
Land Preparation System
.000
1
.000
.000
1.000
(LPS)
Vill*Int
.010
4
.002
.429
.786
Vill *Enum
.000
0
.
.
.
Int*Enum
.195
20
.010
1.740
.121
Vill*Int*Enum
.000
0
.
.
.
Vill*LPS
.000
0
.
.
.
Int*LPS
.005
4
.001
.214
.927
Vill*Int*LPS
.000
0
.
.
.
Enum*LPS
.000
0
.
.
.
Vill*Enum*LPS
.000
0
.
.
.
Int*Enum*LPS
.000
0
.
.
.
Vill*Int*Enum*LPS
.000
0
.
.
.
Error
.101
18
.006
Total
3.555
80
Corrected Total
.355
79
a R Squared = .716 (Adjusted R Squared = -.245)
Source: Pejuan Bean Farmer Survey, 2007.

190

Table 64. T-tests for differences between reported mean and different points of the elicited
distribution.
Midpoint Value
Standard
Type of
Error of
variety and
Maximum
Average
mean
pth
th
4
Interval
5
Interval
Distribution
Reported Reported N difference value
Traditional
N.A.
1393
N.A.
1375
83
44.7
0.685
variety Half
distribution
Traditional
N.A.
N.A.
1550
1375
83
46.6
0.001
variety Half
distribution
Modern
N.A.
1295
N.A.
1453
51
72.5
0.034
variety Half
distribution
Modern
N.A.
N.A.
1439
1453
51
75.6
0.854
variety Half
distribution
Traditional
variety More
1006
N.A.
N.A.
1091
54
43.4
0.058
than Half
distribution
Traditional
variety More
N.A.
1276
N.A.
1091
54
47.5
0.001
than Half
distribution
Modern
variety More
952
N.A.
N.A.
1248
31
64.2
0.001
than Half
distribution
Modern
variety More
N.A.
1211
N.A.
1248
31
66.3
0.579
than Half
distribution
N.A.= Not Available/Applicable
Source: Pejuan Bean Farmer Survey, 2007.

191

Table 65. Traditional Variety Probability Distribution by Village and Interval (Half
Distribution).
Interval
Village
1
2
3
4
5
Total
SAN PEDRO
0.147
0.107
0.160
0.213
0.373
0.200
ALAUCA
0.151
0.061
0.040
0.061
0.205
0.141
3
3
3
3
3
15
0.099
0.142
0.201
0.247
0.312
0.200
CHIRINOS
0.031
0.035
0.015
0.033
0.048
0.083
13
13
13
13
13
65
0.100
0.132
0.204
0.252
0.312
0.200
ARAULI
0.034
0.033
0.040
0.033
0.056
0.087
10
10
10
10
10
50
0.103
0.147
0.206
0.240
0.305
0.200
LA CIENEGA
0.028
0.023
0.024
0.020
0.044
0.076
19
19
19
19
19
95
VILLA DE SAN
0.093
0.147
0.213
0.267
0.280
0.200
FRANCISCO
0.023
0.046
0.023
0.023
0.040
0.079
3
3
3
3
3
15
0.103
0.143
0.194
0.234
0.326
0.200
TALANGA
0.021
0.039
0.028
0.036
0.054
0.086
7
7
7
7
7
35
0.119
0.151
0.211
0.219
0.300
0.200
GUAIMACA
0.057
0.033
0.055
0.041
0.107
0.088
10
10
10
10
10
50
0.120
0.160
0.220
0.240
0.260
0.200
ZABANETAS
0.033
0.000
0.023
0.000
0.023
0.057
4
4
4
4
4
20
0.106
0.143
0.203
0.239
0.308
0.200
Total
0.043
0.032
0.033
0.033
0.069
0.084
69
69
69
69
69
345
Source: Pejuan Bean Farmer Survey, 2007.

192

Table 66. Traditional Variety Probability Distribution by Village and Interval (More than
Half Distribution).
Interval
Village
1
2
3
4
5
Total
SAN PEDRO
0.147
0.187
0.240
0.213
0.213
0.200
ALAUCA
0.023
0.023
0.000
0.023
0.046
0.040
3
3
3
3
3
15
0.103
0.183
0.229
0.269
0.217
0.200
CHIRINOS
0.031
0.045
0.020
0.050
0.056
0.069
7
7
7
7
7
35
0.140
0.220
0.200
0.260
0.180
0.200
ARAULI
0.028
0.028
0.000
0.028
0.028
0.046
2
2
2
2
2
10
0.116
0.200
0.229
0.244
0.211
0.200
LA CIENEGA
0.022
0.031
0.044
0.042
0.051
0.059
11
11
11
11
11
55
VILLA DE SAN
0.133
0.204
0.244
0.253
0.164
0.200
FRANCISCO
0.045
0.042
0.024
0.028
0.068
0.063
9
9
9
9
9
45
0.093
0.227
0.253
0.240
0.187
0.200
TALANGA
0.046
0.023
0.023
0.000
0.046
0.066
3
3
3
3
3
15
0.131
0.234
0.240
0.206
0.189
0.200
GUAIMACA
0.038
0.071
0.052
0.054
0.103
0.075
7
7
7
7
7
35
0.12
0.2
0.16
0.24
0.28
0.2
ZABANETAS
0
0
0
0
0
0.063246
1
1
1
1
1
5
0.122
0.206
0.233
0.242
0.197
0.200
Total
0.035
0.044
0.036
0.043
0.065
0.062
43
43
43
43
43
215
Source: Pejuan Bean Farmer Survey, 2007.

193

Appendix 2. Completing the distributions due to problems encountered

Several ways were tried to obtain a complete shape of the yield distribution. The elicited
distributions classified as half distributions were completed using assuming a normal, right
skewed, and left skewed distributions to be able to decide which distribution was sensible to
assume, once the distributions were completed. The more than half distributions were
completed using a triangular distribution.
To complete each type of the half distributions (i.e., normal, left skewed, and right
skewed distributions), half distributions were first pooled across farmers within the same village,
technology category, and distribution interval from the elicited yield distribution. With the
resulting pooled distributions, interval midpoints were calculated for the second half of the
distribution.
To complete the half distribution and make it a normal distribution, the interval
midpoints followed a mirror image in length of the first half of the distribution. In the same way,
probabilities for each interval in the second part of the distribution reflected probabilities (i.e.,
relative frequency of counters from the elicitation process) in the first part of the distribution
resulting in a bell-shaped distribution. For left skewed distribution, the midpoints were
calculated in the following way:
Midpoint Interval i = Reported Maximum + (Reported Maximum – Interval Midpoint j) 0.8
i = 6,7,8,9,10 & j = 5,4,3,2,1.
The factor of 0.8 results in a conservative left skewness. The counters for the second part of the
distribution were calculated in the following way:
Counters Interval 6 = 1.1* Counters Interval 5

194

Counters Interval 7 = 1.2* Counters Interval 5
Counters Interval 8 = 1.2* Counters Interval 5
Counters Interval 9 = 0.8* Counters Interval 5
Counters Interval 10 = 0.5* Counters Interval 5
The probabilities are the result of ratio of number of counters in each interval to the total number
of counters.
To complete the half distribution and assume a right skewed distribution, the midpoints were
calculated in the following way:
Midpoint Interval i = Reported Maximum + (Reported Maximum – Interval Midpoint j) 1.2
i = 6,7,8,9,10 & j = 5,4,3,2,1.
The factor of 1.2 results in a conservative right skewness. The counters for the second
part of the distribution were calculated in the following way:
Counters Interval 6 = 0.8* Counters Interval 5
Counters Interval 7 = 0.6* Counters Interval 5
Counters Interval 8 = 0.4* Counters Interval 5
Counters Interval 9 = 0.2* Counters Interval 5
Counters Interval 10 = 0.1* Counters Interval 5
Finally, completed half distributions were not used because these did not provide a sense
of skewness from the beginning.
On the other hand, the more than half distributions were tried to be completed by two
methods. The first method was eliminated based on the results it provided. The second method
provided the empirical distribution parameters.

195

The first method used to try to complete the more than half distributions, assumed a
triangular distribution by using a formula that takes into account the midpoint and probability of
the highest interval, and assumed 0.5 probability was left to fill. More specifically, the first
method used the probability of the highest interval together with the assumed 0.5 probability left
to fill the second part of the distribution. With these two pieces of information (i.e., probability
of highest interval (i.e., fifth interval), and 0.5 probability) and using Pythagoras Theorem, a new
maximum value was estimated. Also, probabilities in between this new maximum and midpoint
of the fifth interval were estimated. Finally, when using this method, the maximum values were
out of the possible physical range of bean yields and therefore it was not used.
The second method applied to complete the distributions used the midpoints of the
th

interval with the highest probability (X1), the midpoint of the 5 interval (X2), and the ratio of
th

the density of the interval with the highest probability to the 5 interval (λ). Pythagoras’s
theorem was used to obtain the maximum. The maximum yield followed the following equation:
Maximum yield = λ (2X1 – X2) + X1
This equation starts with the assumption that the maximum is the sum of the mode and an
adjusted value of the difference between double the mode and the reported maximum. The
th

adjustment is made by multiplying the ratio of the density of the modal interval to the 5 interval
density (λ). A higher ratio increases the maximum. Contrarily, the larger the difference between
the maximum reported and the mode’s midpoint, the lower the maximum expected. This is due
to the assumption that a larger section of the distribution was provided by the farmer.
With the new maximum yield estimated, the midpoints of the second part of the
distribution were estimated in the following way:

196

Midpoint Interval i = X2 + (i – 5)(Maximum Yield – X2)/5 ; i = {6,7,8,9,10}
Probabilities were calculated by assuming a constant tangent parameter. This tangent ratio is the
th

number of counters in the 5 interval divided by the distance from the estimated maximum yield
th

to the midpoint of the 5 interval. Then, the counters for each interval were calculated in the
following way:
Counters in interval i = Tangent ratio* Distance from New Maximum to Midpoint i ; i
={6,7,8,9,10}.
Next, the probabilities for each interval of the completed elicited probability distributions
are the result of ratio of number of counters in each interval to the total number of counters. This
completion of the elicited distributions was done for all farmers with the more than half
distribution.

197

Appendix 3. Village Information

Table 67. Bean and corn yield (kg/ha) response to location, land preparation system, type
of variety, and input levels.
Pooled Villages
Bean
Corn
-392**
-793**
Alauca
[11.94]
[3.01]
-432**
368
Arauli & Chirinos
[14.17]
[1.42]
-298**
-573+
Guaimaca & Talanga
[8.80]
[1.83]
-153**
-496+
Sabaneta & Villa de San Francisco
[4.99]
[1.87]
65**
Land Preparation System
152
[2.70]
[0.67]
5
84
Type of Variety
[0.15]
[0.49]
172**
N.A.
Input Level
[8.94]
N.A.
623**
1,544**
Constant
[27.48]
[8.40]
196
142
Observations
0.76
0.18
R-squared
Absolute value of t statistics in brackets
+ significant at 10%; * significant at 5%; ** significant at 1%
Source: Pejuan Farmer Survey, 2007.

198

Table 68. Regression of expert opinion on bean and corn yield (kg/ha) responses to rainfall,
elevation, and fertilizer.
Bean
Corn
18.860**
2.741**
Rainfall
[14.56]
[10.88]
-0.019**
N.A.
Rain square
[15.33]
N.A.
-0.392**
-0.759**
Elevation
[7.05]
[6.15]
0.746**
3.447**
Fertilizer kg/ha
[3.45]
[7.18]
-3,483.709**
733.026**
Constant
[10.56]
[4.65]
20
20
Observations
0.96
0.93
R-squared
Absolute value of t statistics in brackets
+ significant at 10%; * significant at 5%; ** significant at 1%
Source: Pejuan Farmer Survey, 2007.

Table 69. Partial budget analysis: change in income for substituting one hectare of a
modern bean variety for a traditional bean variety in La Cienega.
Increase
Increased income:
1056 kg/ha * L. 11.69/kg
L. 12,345
Reduced cost:
0
L.
0
Total increase
L. 12,345
Decrease
Reduced income:
856 kg/ha * L. 12.99/kg
L. 11,119
Increased cost:
Seed cost (((33+11.69+11.69)/3 (L/kg)) – L. 12.99/kg) * 45.45
L. 264
kg/ha
Harvesting and (1,056-856) kg/ha * L. 1.1/kg
L. 220
threshing
Total decrease
L. 11,603
Net change

L 741

199

Table 70. Partial budget analysis: change in income for substituting one hectare of a
modern bean variety for a traditional bean variety in Talanga and Guaimaca.
Increase
Increased income:
1,246 kg/ha * L. 9.65/kg
L. 12,024
Reduced cost:
0
L.
0
Total increase
L. 12,024
Decrease
Reduced income:
1,046 kg/ha * L. 10.72/kg
L. 11,213
Increased cost:
Seed cost (((33+9.65+9.65)/3 (L/kg)) – L. 10.72/kg) * 45.45
L. 305
kg/ha
Harvesting and (1,246-1,046) kg/ha * L. 1.1/kg
L. 220
threshing
Total decrease
L. 11,738
Net change

L 286

Table 71. Partial budget analysis: change in income for substituting one hectare of a
modern bean variety for a traditional bean variety in Arauli and Chirinos.
Increase
Increased income:
1,051 kg/ha * L. 9.22/kg
L. 9,690
Reduced cost:
0
L.
0
Total increase
L. 9,690
Decrease
Reduced income:
851 kg/ha * L. 10.24/kg
L. 8,714
Increased cost:
Seed cost (((33+9.22+9.22)/3 (L/kg)) – L. 10.24/kg) * 45.45
L. 314
kg/ha
Harvesting and (1,051-851) kg/ha * L. 1.1/kg
L. 220
threshing
Total decrease
L. 9,248
Net change

L 442

200

Table 72. Partial budget analysis: change in income for substituting one hectare of a
modern bean variety for a traditional bean variety in Villa de San Francisco and Sabaneta
de Valle de Angeles.
Increase
Increased income:
1,091 kg/ha * L. 10.05/kg
L. 10,965
Reduced cost:
0
L.
0
Total increase
L. 10,965
Decrease
Reduced income:
891 kg/ha * L. 11.17/kg
L. 9,952
Increased cost:
Seed cost (((33+10.05+10.05)/3 (L/kg)) – L. 11.17/kg) * 45.45
L. 297
kg/ha
Harvesting and (1,091-891) kg/ha * L. 1.1/kg
L. 220
threshing
Total decrease
L. 10,469
Net change

L 495

Table 73. Partial budget analysis: change in income for substituting one hectare of a
modern bean variety for a traditional bean variety in Alauca.
Increase
Increased income:
1,039 kg/ha * L. 9.90/kg
L. 10,286
Reduced cost:
0
L.
0
Total increase
L. 10,286
Decrease
Reduced income:
839 kg/ha * L. 11.00/kg
L. 9,229
Increased cost:
Seed cost (((33+9.90+9.90)/3 (L/kg)) – L. 11.00/kg) * 45.45
L. 300
kg/ha
Harvesting and (1,039-839) kg/ha * L. 1.1/kg
L. 220
threshing
Total decrease
L. 9,749
Net change

L 537

201

Table 74. Percentage of farmers quoting their main way of coping with perils.
Way of coping with perils
Percentage
Look for other sources of income
Other job
50
Other business
4
Reduce consumption
1
Ask for a loan from a friend or family
6
Ask for a loan in other place
14
Sell assets
24
Other
1
Source: Pejuan Farmer Survey, 2007. N=72. 42 Non responses.

202

Table 75. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in La
Cienega; with a RTP*** of 7%, insurance coverage of 75%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
112,480
93,353
32.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance

117,679

93,597

32.64%

Non-mechanized+modern variety+low input use+no credit+no insurance

119,120

93,956

31.94%

Non-mechanized+modern variety+high input use+no credit+no insurance

121,091

92,947

33.38%

Non-mechanized+modern variety+high input use+no credit+with insurance

121,373

95,928

32.58%

Non-mechanized+modern variety+high input use+with credit+with insurance
(switch*****)

116,101

93,416

32.14%

Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

129,825
133,667
132,542
135,293
136,659

104,015
102,469
97,913
103,718
105,729

21.58%
24.60%
28.80%
25.58%
21.40%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

203

Table 76. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in La
Cienega; with a RTP*** of 7%, insurance coverage of 75%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
112,480
93,353
32.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
117,679
93,597
32.64%
Non-mechanized+modern variety+low input use+no credit+no insurance
119,120
93,956
31.94%
Non-mechanized+modern variety+high input use+no credit+no insurance
121,091
92,947
33.38%
Non-mechanized+modern variety+high input use+no credit+with insurance
116,807
92,969
35.56%
Non-mechanized+modern variety+high input use+with credit+with insurance
112,136
90,864
35.22%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

129,825
133,667
132,542
129,165
130,615

104,015
102,469
97,913
99,616
101,623

21.58%
24.60%
28.80%
25.58%
21.40%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

204

Table 77. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in La
Cienega; with a RTP*** of 7%, insurance coverage of 45%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
112,480
93,353
32.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
117,679
93,597
32.64%
Non-mechanized+modern variety+low input use+no credit+no insurance
119,120
93,956
31.94%
Non-mechanized+modern variety+high input use+no credit+no insurance
121,091
92,947
33.38%
Non-mechanized+modern variety+high input use+no credit+with insurance
120,507
93,324
33.90%
Non-mechanized+modern variety+high input use+with credit+with insurance
115,819
91,493
33.78%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

129,825
133,667
132,542
132,377
133,770

104,015
102,469
97,913
98,900
100,844

21.58%
24.60%
28.80%
28.86%
26.68%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

205

Table 78. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in La
Cienega; with a RTP*** of 7%, insurance coverage of 45%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
112,480
93,353
32.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
117,679
93,597
32.64%
Non-mechanized+modern variety+low input use+no credit+no insurance
119,120
93,956
31.94%
Non-mechanized+modern variety+high input use+no credit+no insurance
121,091
92,947
33.38%
Non-mechanized+modern variety+high input use+no credit+with insurance
119,441
92,710
34.56%
Non-mechanized+modern variety+high input use+with credit+with insurance
114,906
90,980
34.52%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

129,825
133,667
132,542
136,936
132,357

104,015
102,469
97,913
98,064
100,020

21.58%
24.60%
28.80%
29.64%
27.46%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

206

Table 79. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in La
Cienega; with a RTP*** of 14%, insurance coverage of 75%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
114,132
101,211
32.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
118,408
101,450
32.64%
Non-mechanized+modern variety+low input use+no credit+no insurance
119,660
101,848
31.94%
Non-mechanized+modern variety+high input use+no credit+no insurance
121,385
100,766
33.38%
Non-mechanized+modern variety+high input use+no credit+with insurance
122,312
104,421
32.58%
Non-mechanized+modern variety+high input use+with credit+with insurance
118,117
101,762
32.14%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

126,997
130,420
130,578
133,227
134,119

110,490
109,672
105,728
111,884
113,356

21.58%
24.60%
28.80%
25.58%
21.40%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

207

Table 80. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in La
Cienega; with a RTP*** of 14%, insurance coverage of 75%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
114,132
101,211
32.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
118,408
101,450
32.64%
Non-mechanized+modern variety+low input use+no credit+no insurance
119,660
101,848
31.94%
Non-mechanized+modern variety+high input use+no credit+no insurance
121,385
100,766
33.38%
Non-mechanized+modern variety+high input use+no credit+with insurance
118,690
101,530
35.76%
Non-mechanized+modern variety+high input use+with credit+with insurance
114,919
99,271
35.22%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

126,997
130,420
130,578
128,630
129,678

110,790
109,672
105,728
108,082
109,699

21.58%
24.60%
28.80%
29.56%
26.14%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

208

Table 81. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in La
Cienega; with a RTP*** of 14%, insurance coverage of 45%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
114,132
101,211
32.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
118,132
101,450
32.64%
Non-mechanized+modern variety+low input use+no credit+no insurance
119,660
101,848
31.94%
Non-mechanized+modern variety+high input use+no credit+no insurance
121,385
100,766
33.38%
Non-mechanized+modern variety+high input use+no credit+with insurance
121,174
101,443
33.90%
Non-mechanized+modern variety+high input use+with credit+with insurance
117,593
99,643
33.78%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

126,997
130,420
130,578
130,738
131,945

110,790
109,672
105,728
107,059
108,923

21.58%
24.60%
28.80%
28.86%
26.68%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

209

Table 82. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in La
Cienega; with a RTP*** of 14%, insurance coverage of 45%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
114,132
101,211
32.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
118,408
101,450
32.64%
Non-mechanized+modern variety+low input use+no credit+no insurance
119,660
101,848
31.94%
Non-mechanized+modern variety+high input use+no credit+no insurance
121,385
100,766
33.38%
Non-mechanized+modern variety+high input use+no credit+with insurance
120,316
100,820
34.56%
Non-mechanized+modern variety+high input use+with credit+with insurance
116,839
99,109
34.52%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

126,997
130,420
130,578
129,621
130,870

110,790
109,672
105,728
106,233
108,134

21.58%
24.60%
28.80%
29.64%
27.46%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

210

Table 83. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Talanga and Guaimaca; with a RTP*** of 7%, insurance coverage of 75%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
120,486
98,689
27.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
120,342
95,503
30.42%
Non-mechanized+modern variety+low input use+no credit+no insurance
122,082
96,392
29.62%
Non-mechanized+modern variety+high input use+no credit+no insurance
120,813
93,341
32.36%
Non-mechanized+modern variety+high input use+no credit+with insurance
121,203
97,083
31.46%
Non-mechanized+modern variety+high input use+with credit+with insurance
115,656
94,174
31.72%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

137,493
134,735
130,898
134,795
136,130

109,122
103,513
98,133
105,825
107,827

18.44%
23.22%
28.12%
22.78%
19.36%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

211

Table 84. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Talanga and Guaimaca; with a RTP*** of 7%, insurance coverage of 75%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
120,486
98,689
27.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
120,342
95,503
30.42%
Non-mechanized+modern variety+low input use+no credit+no insurance
122,082
96,392
29.62%
Non-mechanized+modern variety+high input use+no credit+no insurance
120,813
93,341
32.36%
Non-mechanized+modern variety+high input use+no credit+with insurance
115,635
93,301
35.56%
Non-mechanized+modern variety+high input use+with credit+with insurance
110,821
90,920
36.24%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

137,493
134,735
130,898
127,392
128,845

109,122
103,513
98,133
100,479
102,487

18.44%
23.22%
28.12%
27.90%
24.64%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

212

Table 85. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Talanga and Guaimaca; with a RTP*** of 7%, insurance coverage of 45%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
120,486
98,689
27.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
120,342
95,503
30.42%
Non-mechanized+modern variety+low input use+no credit+no insurance
120,082
96,392
29.62%
Non-mechanized+modern variety+high input use+no credit+no insurance
120,813
93,341
32.36%
Non-mechanized+modern variety+high input use+no credit+with insurance
119,978
93,682
32.88%
Non-mechanized+modern variety+high input use+with credit+with insurance
115,343
91,865
33.94%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

137,493
134,735
130,898
130,956
132,504

109,122
103,513
98,133
99,447
101,562

18.44%
23.22%
28.12%
27.82%
25.82%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

213

Table 86. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Talanga and Guaimaca; with a RTP*** of 7%, insurance coverage of 45%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
120,486
98,689
27.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
120,342
95,503
30.42%
Non-mechanized+modern variety+low input use+no credit+no insurance
122,082
96,392
29.62%
Non-mechanized+modern variety+high input use+no credit+no insurance
120,813
93,341
32.36%
Non-mechanized+modern variety+high input use+no credit+with insurance
118,747
92,965
33.62%
Non-mechanized+modern variety+high input use+with credit+with insurance
114,201
91,176
34.64%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

137,493
134,735
130,898
129,415
130,979

109,122
103,513
98,133
98,507
100,507

18.44%
23.22%
28.12%
28.68%
26.58%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

214

Table 87. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Talanga and Guaimaca; with a RTP*** of 14%, insurance coverage of 75%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
120,082
105,986
27.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
120,320
103,195
30.42%
Non-mechanized+modern variety+low input use+no credit+no insurance
121,928
104,153
29.62%
Non-mechanized+modern variety+high input use+no credit+no insurance
121,056
101,139
32.36%
Non-mechanized+modern variety+high input use+no credit+with insurance
122,215
105,587
31.46%
Non-mechanized+modern variety+high input use+with credit+with insurance
117,781
102,561
31.72%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

132,478
131,121
129,155
132,549
133,294

115,195
110,660
105,955
113,523
114,804

18.44%
23.22%
28.12%
22.78%
19.36%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

215

Table 88. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Talanga and Guaimaca; with a RTP*** of 14%, insurance coverage of 75%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
120,082
105,986
27.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
120,320
103,195
30.42%
Non-mechanized+modern variety+low input use+no credit+no insurance
121,928
104,153
29.62%
Non-mechanized+modern variety+high input use+no credit+no insurance
121,056
101,139
32.36%
Non-mechanized+modern variety+high input use+no credit+with insurance
117,855
101,972
35.56%
Non-mechanized+modern variety+high input use+with credit+with insurance
113,966
99,487
36.24%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

132,478
131,121
129,155
127,135
128,109

115,195
110,660
105,955
108,812
110,325

18.44%
23.22%
28.12%
27.90%
24.64%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

216

Table 89. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Talanga and Guaimaca; with a RTP*** of 14%, insurance coverage of 45%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
120,082
105,986
27.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
120,320
103,195
30.42%
Non-mechanized+modern variety+low input use+no credit+no insurance
121,928
104,153
29.62%
Non-mechanized+modern variety+high input use+no credit+no insurance
121,056
101,139
32.36%
Non-mechanized+modern variety+high input use+no credit+with insurance
120,736
101,860
32.88%
Non-mechanized+modern variety+high input use+with credit+with insurance
117,191
100,073
33.94%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

132,478
131,121
129,159
129,511
130,788

115,195
110,660
105,955
107,608
109,575

18.44%
23.22%
28.12%
27.82%
25.82%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

217

Table 90. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Talanga and Guaimaca; with a RTP*** of 14%, insurance coverage of 45%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
120,082
105,986
27.26%
Non-mechanized+traditional variety+high input use+no credit+no insurance
120,320
103,195
30.42%
Non-mechanized+modern variety+low input use+no credit+no insurance
121,928
104,153
29.62%
Non-mechanized+modern variety+high input use+no credit+no insurance
121,056
101,139
32.36%
Non-mechanized+modern variety+high input use+no credit+with insurance
119,747
101,132
33.62%
Non-mechanized+modern variety+high input use+with credit+with insurance
116,281
99,402
34.64%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

132,478
131,121
129,155
128,325
129,637

115,195
110,660
105,955
106,697
108,687

18.44%
23.22%
28.12%
28.68%
26.58%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

218

Table 91. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in Villa
de San Francisco and Sabaneta de Valle de Angeles; with a RTP*** of 7%, insurance coverage of 75%, load=1.0, and two
CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
112,034
94,714
31.00%
Non-mechanized+traditional variety+high input use+no credit+no insurance
114,079
93,012
32.92%
Non-mechanized+modern variety+low input use+no credit+no insurance
117,502
94,910
30.86%
Non-mechanized+modern variety+high input use+no credit+no insurance
116,386
92,028
34.32%
Non-mechanized+modern variety+high input use+no credit+with insurance
116,643
95,198
33.48%
Non-mechanized+modern variety+high input use+with credit+with insurance
111,332
92,377
32.64%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

129,221
128,313
127,462
130,342
131,697

105,850
101,456
97,207
103,679
105,650

19.78%
24.38%
29.04%
24.92%
20.90%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

219

Table 92. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in Villa
de San Francisco and Sabaneta de Valle de Angeles; with a RTP*** of 7%, insurance coverage of 75%, load=1.4, and two
CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
112,034
94,719
31.00%
Non-mechanized+traditional variety+high input use+no credit+no insurance
114,079
93,012
32.92%
Non-mechanized+modern variety+low input use+no credit+no insurance
117,502
94,910
30.86%
Non-mechanized+modern variety+high input use+no credit+no insurance
116,386
92,028
34.32%
Non-mechanized+modern variety+high input use+no credit+with insurance
111,977
91,972
37.24%
Non-mechanized+modern variety+high input use+with credit+with insurance
107,320
89,628
37.04%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

129,221
128,313
127,462
123,880
125,282

105,850
101,456
97,207
98,994
100,926

19.78%
24.38%
29.04%
29.70%
25.54%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

220

Table 93. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in Villa
de San Francisco and Sabaneta de Valle de Angeles; with a RTP*** of 7%, insurance coverage of 45%, load=1.0, and two
CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
112,034
94,719
31.00%
Non-mechanized+traditional variety+high input use+no credit+no insurance
114,079
93,012
32.92%
Non-mechanized+modern variety+low input use+no credit+no insurance
117,502
94,910
30.86%
Non-mechanized+modern variety+high input use+no credit+no insurance
116,386
92,028
34.32%
Non-mechanized+modern variety+high input use+no credit+with insurance
115,836
92,432
34.64%
Non-mechanized+modern variety+high input use+with credit+with insurance
111,305
90,552
34.70%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

129,221
128,313
127,462
127,207
128,709

105,850
101,456
97,207
98,226
100,254

19.78%
24.38%
29.04%
28.88%
25.94%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

221

Table 94. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in Villa
de San Francisco and Sabaneta de Valle de Angeles; with a RTP*** of 7%, insurance coverage of 45%, load=1.4, and two
CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
112,034
94,719
31.00%
Non-mechanized+traditional variety+high input use+no credit+no insurance
114,079
93,012
32.92%
Non-mechanized+modern variety+low input use+no credit+no insurance
117,502
94,910
30.86%
Non-mechanized+modern variety+high input use+no credit+no insurance
116,386
92,028
34.32%
Non-mechanized+modern variety+high input use+no credit+with insurance
114,778
91,791
35.42%
Non-mechanized+modern variety+high input use+with credit+with insurance
110,328
89,946
35.26%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

1292,221
128,313
127,462
125,724
127,125

105,850
101,456
97,207
97,313
99,234

19.78%
24.38%
29.04%
29.62%
26.88%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

222

Table 95. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in Villa
de San Francisco and Sabaneta de Valle de Angeles; with a RTP*** of 14%, insurance coverage of 75%, load=1.0, and two
CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
113,672
102,395
31.00%
Non-mechanized+traditional variety+high input use+no credit+no insurance
115,381
100,833
32.92%
Non-mechanized+modern variety+low input use+no credit+no insurance
118,175
102,701
30.86%
Non-mechanized+modern variety+high input use+no credit+no insurance
117,530
99,921
34.32%
Non-mechanized+modern variety+high input use+no credit+with insurance
118,219
103,742
33.48%
Non-mechanized+modern variety+high input use+with credit+with insurance
114,219
100,744
32.64%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

126,208
126,182
126,365
129,158
129,894

112,132
108,647
105,036
111,615
112,831

19.78%
24.38%
29.04%
24.92%
20.90%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

223

Table 96. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in Villa
de San Francisco and Sabaneta de Valle de Angeles; with a RTP*** of 14%, insurance coverage of 75%, load=1.4, and two
CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
113,676
102,395
31.00%
Non-mechanized+traditional variety+high input use+no credit+no insurance
115,381
100,833
32.92%
Non-mechanized+modern variety+low input use+no credit+no insurance
118,175
102,701
30.86%
Non-mechanized+modern variety+high input use+no credit+no insurance
117,530
99,921
34.32%
Non-mechanized+modern variety+high input use+no credit+with insurance
114,841
100,648
37.24%
Non-mechanized+modern variety+high input use+with credit+with insurance
111,008
98,136
37.04%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

126,208
126,182
126,365
124,352
125,270

112,132
108,647
105,036
107,424
108,828

19.78%
24.38%
29.04%
29.70%
25.54%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

224

Table 97. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in Villa
de San Francisco and Sabaneta de Valle de Angeles; with a RTP*** of 14%, insurance coverage of 45%, load=1.0, and two
CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
113,672
102,395
31.00%
Non-mechanized+traditional variety+high input use+no credit+no insurance
115,381
100,833
32.92%
Non-mechanized+modern variety+low input use+no credit+no insurance
118,175
102,701
30.86%
Non-mechanized+modern variety+high input use+no credit+no insurance
117,530
99,921
34.32%
Non-mechanized+modern variety+high input use+no credit+with insurance
117,363
100,625
34.64%
Non-mechanized+modern variety+high input use+with credit+with insurance
113,814
98,709
34.70%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

126,208
126,182
126,365
126,488
127,656

112,132
108,647
105,036
106,398
108,179

19.78%
24.38%
29.04%
28.88%
25.94%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

225

Table 98. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in Villa
de San Francisco and Sabaneta de Valle de Angeles; with a RTP*** of 14%, insurance coverage of 45%, load=1.4, and two
CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
113,672
102,395
31.00%
Non-mechanized+traditional variety+high input use+no credit+no insurance
115,381
100,833
32.92%
Non-mechanized+modern variety+low input use+no credit+no insurance
118,175
102,701
30.86%
Non-mechanized+modern variety+high input use+no credit+no insurance
117,530
99,921
34.32%
Non-mechanized+modern variety+high input use+no credit+with insurance
116,513
99,982
35.42%
Non-mechanized+modern variety+high input use+with credit+with insurance
113,027
98,112
35.26%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

126,208
126,182
126,365
125,331
126,486

112,132
108,647
105,036
105,500
107,254

19.78%
24.38%
29.04%
29.62%
26.88%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

226

Table 99. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Alauca; with a RTP*** of 7%, insurance coverage of 75%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
116,122
98,369
26.28%
Non-mechanized+traditional variety+high input use+no credit+no insurance
116,886
95,709
29.76%
Non-mechanized+modern variety+low input use+no credit+no insurance
122,133
98,366
27.48%
Non-mechanized+modern variety+high input use+no credit+no insurance
119,834
94,493
31.82%
Non-mechanized+modern variety+high input use+no credit+with insurance
120,970
98,794
29.44%
Non-mechanized+modern variety+high input use+with credit+with insurance
114,563
94,933
30.92%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

135,252
132,654
132,227
137,367
138,363

111,520
104,884
100,378
109,647
111,337

15.24%
21.66%
26.22%
18.68%
15.64%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

227

Table 100. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Alauca; with a RTP*** of 7%, insurance coverage of 75%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
116,122
98,369
26.28%
Non-mechanized+traditional variety+high input use+no credit+no insurance
116,886
95,709
29.76%
Non-mechanized+modern variety+low input use+no credit+no insurance
122,133
98,366
27.48%
Non-mechanized+modern variety+high input use+no credit+no insurance
119,834
94,493
31.82%
Non-mechanized+modern variety+high input use+no credit+with insurance
115,753
95,023
34.10%
Non-mechanized+modern variety+high input use+with credit+with insurance
110,118
91,791
35.42%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

135,252
132,654
132,227
130,090
131,137

111,520
104,884
100,378
104,036
105,709

15.24%
21.66%
26.22%
24.18%
21.36%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

228

Table 101. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Alauca; with a RTP*** of 7%, insurance coverage of 45%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
116,122
98,369
26.28%
Non-mechanized+traditional variety+high input use+no credit+no insurance
116,886
95,709
29.76%
Non-mechanized+modern variety+low input use+no credit+no insurance
122,133
98,366
27.48%
Non-mechanized+modern variety+high input use+no credit+no insurance
119,834
94,493
31.82%
Non-mechanized+modern variety+high input use+no credit+with insurance
119,365
95,096
32.10%
Non-mechanized+modern variety+high input use+with credit+with insurance
114,166
92,653
33.30%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

135,252
132,654
132,227
132,493
133,876

111,520
104,884
100,378
102,047
104,065

15.24%
21.66%
26.22%
25.72%
23.64%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

229

Table 102. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Alauca; with a RTP*** of 7%, insurance coverage of 45%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
116,122
98,369
26.28%
Non-mechanized+traditional variety+high input use+no credit+no insurance
116,886
95,709
29.76%
Non-mechanized+modern variety+low input use+no credit+no insurance
122,132
98,366
27.48%
Non-mechanized+modern variety+high input use+no credit+no insurance
119,834
94,493
31.82%
Non-mechanized+modern variety+high input use+no credit+with insurance
118,171
94,329
33.22%
Non-mechanized+modern variety+high input use+with credit+with insurance
113,123
91,987
34.14%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

135,252
132,654
132,227
130,875
132,256

111,520
104,884
100,378
101,007
103,002

15.24%
21.66%
26.22%
26.64%
24.48%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

230

Table 103. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Alauca; with a RTP*** of 14%, insurance coverage of 75%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
116,511
105,377
26.28%
Non-mechanized+traditional variety+high input use+no credit+no insurance
117,535
103,304
29.76%
Non-mechanized+modern variety+low input use+no credit+no insurance
121,554
105,813
27.48%
Non-mechanized+modern variety+high input use+no credit+no insurance
120,235
102,353
31.82%
Non-mechanized+modern variety+high input use+no credit+with insurance
121,789
106,967
29.44%
Non-mechanized+modern variety+high input use+with credit+with insurance
116,763
103,166
30.92%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

130,036
129,163
129,783
133,698
134,153

116,415
111,603
108,010
116,256
117,172

15.24%
21.66%
26.22%
18.68%
15.64%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

231

Table 104. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Alauca; with a RTP*** of 14%, insurance coverage of 75%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
116,511
105,377
26.28%
Non-mechanized+traditional variety+high input use+no credit+no insurance
117,535
103,304
29.76%
Non-mechanized+modern variety+low input use+no credit+no insurance
121,554
105,813
27.48%
Non-mechanized+modern variety+high input use+no credit+no insurance
129,235
102,353
31.82%
Non-mechanized+modern variety+high input use+no credit+with insurance
117,806
103,529
34.10%
Non-mechanized+modern variety+high input use+with credit+with insurance
113,301
100,283
35.42%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

130,036
129,173
129,783
128,612
129,251

116,415
111,603
118,010
111,647
112,775

15.24%
21.66%
26.22%
24.18%
21.36%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

232

Table 105. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Alauca; with a RTP*** of 14%, insurance coverage of 45%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
116,511
105,377
26.28%
Non-mechanized+traditional variety+high input use+no credit+no insurance
117,535
103,304
29.76%
Non-mechanized+modern variety+low input use+no credit+no insurance
121,554
105,813
27.48%
Non-mechanized+modern variety+high input use+no credit+no insurance
129,235
102,353
31.82%
Non-mechanized+modern variety+high input use+no credit+with insurance
120,183
103,277
32.10%
Non-mechanized+modern variety+high input use+with credit+with insurance
116,135
100,839
33.30%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

130,083
129,173
129,783
130,275
131,317

116,415
111,603
108,010
109,925
111,646

15.24%
21.66%
26.22%
25.72%
23.64%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

233

Table 106. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Alauca; with a RTP*** of 14%, insurance coverage of 45%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
116,511
105,377
26.28%
Non-mechanized+traditional variety+high input use+no credit+no insurance
117,535
103,304
29.76%
Non-mechanized+modern variety+low input use+no credit+no insurance
121,554
105,813
27.48%
Non-mechanized+modern variety+high input use+no credit+no insurance
129,235
102,353
31.82%
Non-mechanized+modern variety+high input use+no credit+with insurance
119,257
102,547
33.22%
Non-mechanized+modern variety+high input use+with credit+with insurance
115,308
100,198
34.14%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

130,036
129,173
129,783
129,062
130,131

116,415
111,603
108,010
108,951
110,691

15.24%
21.66%
26.22%
26.64%
24.48%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

234

Table 107. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Arauli and Chirinos; with a RTP*** of 7%, insurance coverage of 75%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
109,837
94,599
30.40%
Non-mechanized+traditional variety+high input use+no credit+no insurance
109,877
91,953
34.60%
Non-mechanized+modern variety+low input use+no credit+no insurance
115,305
95,078
30.82%
Non-mechanized+modern variety+high input use+no credit+no insurance
112,937
91,355
35.06%
Non-mechanized+modern variety+high input use+no credit+with insurance
112,516
93,985
35.08%
Non-mechanized+modern variety+high input use+with credit+with insurance
107,038
90,718
35.42%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

126,951
124,201
124,277
127,040
128,314

106,400
100,425
97,027
103,721
105,611

18.78%
24.96%
28.78%
24.20%
19.44%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

235

Table 108. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Arauli and Chirinos; with a RTP*** of 7%, insurance coverage of 75%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
109,837
94,599
30.40%
Non-mechanized+traditional variety+high input use+no credit+no insurance
109,877
91,953
34.60%
Non-mechanized+modern variety+low input use+no credit+no insurance
115,305
95,078
30.82%
Non-mechanized+modern variety+high input use+no credit+no insurance
112,937
91,355
35.06%
Non-mechanized+modern variety+high input use+no credit+with insurance
107,769
90,541
39.50%
Non-mechanized+modern variety+high input use+with credit+with insurance
103,119
87,958
40.04%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

126,951
124,201
124,277
120,518
121,916

106,400
100,425
97,027
98,691
100,619

18.78%
24.96%
28.78%
29.90%
25.46%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

236

Table 109. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Arauli and Chirinos; with a RTP*** of 7%, insurance coverage of 45%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
109,837
94,599
30.40%
Non-mechanized+traditional variety+high input use+no credit+no insurance
109,877
91,953
34.60%
Non-mechanized+modern variety+low input use+no credit+no insurance
115,305
95,078
30.82%
Non-mechanized+modern variety+high input use+no credit+no insurance
112,937
91,355
35.06%
Non-mechanized+modern variety+high input use+no credit+with insurance
112,208
91,629
35.82%
Non-mechanized+modern variety+high input use+with credit+with insurance
107,490
89,379
36.50%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

126,951
124,201
124,277
123,901
125,303

106,400
100,425
97,027
97,996
99,959

18.78%
24.96%
28.78%
28.98%
26.18%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

237

Table 110. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Arauli and Chirinos; with a RTP*** of 7%, insurance coverage of 45%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
109,837
94,599
30.40%
Non-mechanized+traditional variety+high input use+no credit+no insurance
109,877
91,953
34.60%
Non-mechanized+modern variety+low input use+no credit+no insurance
115,305
95,078
30.82%
Non-mechanized+modern variety+high input use+no credit+no insurance
112,937
91,355
35.06%
Non-mechanized+modern variety+high input use+no credit+with insurance
111,081
90,895
36.78%
Non-mechanized+modern variety+high input use+with credit+with insurance
106,524
88,758
37.64%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

126,951
124,201
124,277
122,406
123,837

106,400
100,425
97,027
97,016
98,960

18.78%
24.96%
28.78%
30.04%
27.14%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

238

Table 111. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Arauli and Chirinos; with a RTP*** of 14%, insurance coverage of 75%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
111,868
102,195
30.40%
Non-mechanized+traditional variety+high input use+no credit+no insurance
112,051
99,877
34.60%
Non-mechanized+modern variety+low input use+no credit+no insurance
116,399
102,892
30.82%
Non-mechanized+modern variety+high input use+no credit+no insurance
114,684
99,275
35.06%
Non-mechanized+modern variety+high input use+no credit+with insurance
115,245
102,583
35.08%
Non-mechanized+modern variety+high input use+with credit+with insurance
110,741
99,199
35.42%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

124,328
123,010
123,592
126,324
126,905

112,323
107,700
104,804
111,382
112,379

18.78%
24.96%
28.78%
24.20%
19.44%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

239

Table 112. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Arauli and Chirinos; with a RTP*** of 14%, insurance coverage of 75%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
111,868
102,195
30.40%
Non-mechanized+traditional variety+high input use+no credit+no insurance
112,051
99,877
34.60%
Non-mechanized+modern variety+low input use+no credit+no insurance
116,399
102,892
30.82%
Non-mechanized+modern variety+high input use+no credit+no insurance
114,684
99,275
35.06%
Non-mechanized+modern variety+high input use+no credit+with insurance
111,483
99,344
39.50%
Non-mechanized+modern variety+high input use+with credit+with insurance
107,560
96,562
40.04%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

124,328
123,010
123,592
121,534
122,348

112,323
107,700
104,804
107,049
108,313

18.78%
24.96%
28.78%
29.90%
25.46%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

240

Table 113. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Arauli and Chirinos; with a RTP*** of 14%, insurance coverage of 45%, load=1.0, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
111,868
102,195
30.40%
Non-mechanized+traditional variety+high input use+no credit+no insurance
112,051
99,877
34.60%
Non-mechanized+modern variety+low input use+no credit+no insurance
116,399
102,892
30.82%
Non-mechanized+modern variety+high input use+no credit+no insurance
114,684
99,275
35.06%
Non-mechanized+modern variety+high input use+no credit+with insurance
114,430
99,874
35.82%
Non-mechanized+modern variety+high input use+with credit+with insurance
110,685
97,607
36.50%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

124,328
123,010
123,592
123,696
124,738

112,323
107,700
104,804
106,156
107,785

18.78%
24.96%
28.78%
28.98%
26.18%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

241

Table 114. CE* values (L.)* and IP** for different technology categories, and credit and crop insurance use for farmers in
Arauli and Chirinos; with a RTP*** of 14%, insurance coverage of 45%, load=1.4, and two CRRAC****.
Certainty Equivalent
Percent
Technology Category/credit and insurance
Insolvent
CRRAC=2
CRRAC=4
Non-mechanized+traditional variety+low input use+no credit+no insurance
111,868
102,195
30.40%
Non-mechanized+traditional variety+high input use+no credit+no insurance
112,051
99,877
34.60%
Non-mechanized+modern variety+low input use+no credit+no insurance
116,399
102,892
30.82%
Non-mechanized+modern variety+high input use+no credit+no insurance
114,684
99,275
35.06%
Non-mechanized+modern variety+high input use+no credit+with insurance
113,535
99,183
36.78%
Non-mechanized+modern variety+high input use+with credit+with insurance
109,917
97,018
37.64%
(switch*****)
Mechanized+traditional variety+low input use+no credit+no insurance
Mechanized+traditional variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+no insurance
Mechanized+modern variety+high input use+no credit+with insurance
Mechanized+modern variety+high input use+with credit+with insurance (switch)

124,328
123,010
123,592
122,541
123,636

102,323
107,700
104,804
105,218
106,900

18.78%
24.96%
28.78%
30.04%
27.14%

* CE=Certainty equivalent; L.=Lempira, Honduran currency.US $1=19.03 in 2007; **IP=insolvency percentage out of 5,000;
Insolvent is a producer with an ending period wealth less than or equal to 1.25 times the house value. ; ***RTP=Rate of time
preference; ****CRRAC=Constant relative risk aversion coefficient.
*****Switch refers to the farmer switching back from the actual technology category (e.g., non-mechanized land preparation,
modern variety, high input) to base technology category (i.e., non-mechanized land preparation, traditional variety, low input). Base
category for non-mechanized technologies is non-mechanized traditional low input, and for mechanized technologies is mechanized
traditional low input. The switch is done when the farmer is no longer eligible for credit by the bank´s standard (i.e., below 1.75
times the farmer´s house value).

242

BIBLIOGRAPHY

243

BIBLIOGRAPHY

Armendariz, B. and Morduch, J. The Economics of Microfinance; The MIT Press; Cambridge ,
Massachusetts 2005.
Bellman, R., & Dreyfus, S. E. (1962).Applied dynamic programming. Princeton, N.J.: Princeton
University Press.
Carroll, C. D.; “Buffer-Stock Saving and the Life Cycle/Permanent Income Hypothesis”, The
Quarterly Journal of Economics, Vol.112, No. 1. (Feb., 1997), pp. 1-55.
Castillo, O.; personal communication. 2006.
Deaton, A. “Saving and Liquidity Constraints”, Econometrica, Vol. 59, No. 5 (Sep., 1999); pp.
1221-1248.
Erazo, D.; Personal communication. 2007.
Feder, G.; Just, R.; and Zilberman, D.; Adoption of Agricultural Innovations in Developing
Countries: A Survey; Economic Development and Cultural Change, Vol. 33, No. 2 (Jan., 1985),
255-298.
Francisco, E.; Anderson, J.; “Chance and Choice west of the Darling”; Australian Journal of
Agricultural Economics, 16(1972) 82-93.
Grisley, W.; Kellogg, E.; Farmers subjective probabilities in Northern Thailand; American
Journal of Agricultural Economics, Vol 65, No. 1 (Feb., 1983) 74-82.
Guzman, C.; Costos de Administracion y Transaccion de Tecnologias Crediticias no
Tradicionales: Foro de Microfinanzas; 2006.
Hess, U. and Syroka, J.; Weather-based Insurance in Southern Africa: The case of Malawi. The
World Bank, 2005.
IICA (Instituto Interamericano de Cooperación para la Agricultura) (2006), “Mapeo del mercado
de semillas de maíz blanco y frijol en Centroamérica, Proyecto Red SICTA, COSUDE. 2009.
Just, R. E., L. Calvin, and J. Quiggin. "Adverse Selection in crop insurance: Actuarial and
asymmetric information incentives." American Journal of Agricultural Economics 81(1999):
834-849.
Kurosaki, T.; “Production Risk and Advantages of Mixed Farming in Pakistan Punjab”; The
Developing Economies Mar (1997): 28-47.

244

Mather, D.; Essays on the Economic impact of Disease-Resistant Bean Research in Honduras;
Department of Agricultural Economics; Dissertation, 2003, Michigan State University.
Miranda, M. "Area-yield crop insurance reconsidered." American Journal of Agricultural
EconomicsMay(1991): 233-242.
Moneda (May 12-16 2003) Honduras: Oposicion a Seguro Agricola Obligatorio, vol. 2004,
http://www.terra.com.hn/moneda/noticias/mnd12313.htm.
Moneda (October 14 to November 10, 2003) Honduras: Poco Interes en Seguro Agricola, vol.
2004, http://www.terra.com.hn/moneda/noticias/mnd12313.htm.
Morduch, J.Income Smoothing and Consumption Smoothing; Journal of Economic Perspective,
vol. 9, No. 3, (1995):103-114.
th

Nicholson, W. Microeconomic Theory, 7 Edition; Fort Worth; The Dryden Press; 1998.
Nyambane, G.; The Dynamics of Agricultural Insurance and Consumption Smoothing;
Department of Agricultural Economics; Dissertation, 2005, Michigan State University.
Osborne, T.; “Credit and risk in rural developing economies”, Journal of Economic Dynamics &
Control 30 (2006), pp. 541-568
Pease, J.;, Prediction of Subjective Yield Expectations Using Farm Records, Staff Paper 89-29
1988 Virginia Polythechnic Institute and University.
Rothschild, Michael and Joseph Stiglitz [1976], “Equilibrium in Competitive
Insurance Markets: An Essay on the Economics of Imperfect Information,”
Quarterly Journal of Economics, vol. 90, 630-649.
Skees, J.; Buzby, J.; Pease, J.; The Usefulness of Subjective Data in the Analysis of Yield Risk;
Department of Agricultural Economics; Staff paper 285, Jan 1991, University of Kentucky.
Smeltekop, H; Constraints to Greater Bean Yields in Honduras; Department of Crop and Soil
Sciences; Dissertation, 2005, Michigan State University.
Williams, J. R., G.L. Carriker, G.A. Barnaby, and J.K. Harper. "Crop insurance and disaster
assistance designs for wheat and grain sorghum." American Journal of Agricultural Economics
75(1993): 435-447.
The World Bank; En Breve: Learning by Doing: Applying a Country-led PSIA Approach to
Improve Policy Making in the Agro Sector in Honduras Report Number 153. 2005.

245

The World Bank; Managing Agricultural Production Risk: Innovations in Developing Countries.
Report Number 32727-GLB. 2005.

246