Citation
Agrarian reform cooperatives in Honduras

Material Information

Title:
Agrarian reform cooperatives in Honduras
Creator:
Martin, Michael, 1954-
Publication Date:
Language:
English
Physical Description:
xii, 211 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Agrarian reform ( jstor )
Agricultural land ( jstor )
Agriculture ( jstor )
Allocative efficiency ( jstor )
Corn ( jstor )
Economics ( jstor )
Efficiency metrics ( jstor )
Farmers ( jstor )
Human capital ( jstor )
Production efficiency ( jstor )
Agriculture -- Honduras ( lcsh )
Dissertations, Academic -- Food and Resource Economics -- UF ( lcsh )
Food and Resource Economics thesis, Ph. D ( lcsh )
Land reform -- Honduras ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1996.
Bibliography:
Includes bibliographical references (leaves 199-210).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Michael J. Martin.

Record Information

Source Institution:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
026432142 ( ALEPH )
36750970 ( OCLC )

Downloads

This item has the following downloads:


Full Text










AGRARIAN REFORM COOPERATIVES
IN HONDURAS

















By

MICHAEL J. MARTIN


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1996

































Copyright 1996 by

Michael J. Martin
































This dissertation is dedicated to my Honduran friends who were patient and kind enough to share their rich lives and culture with me. It is also dedicated to my parents, whose unquestionable love and concern turned a Central American sojourn into a larger home.













ACKNOWLEDGMENTS


This dissertation is the product of the collaboration and assistance among many friends in Honduras and the University of Florida. I am especially grateful to my Chair, Professor Timothy Taylor, who generously contributed his guidance and support throughout the entire research process. The door to his abundant professional faculties was always open. I am also very grateful to my Cochair, Professor Uma Lele, whose amiably shared insights in the realm of international development have been invaluable in broadening my perspectives. Professor Chris Andrew, in whose class I presented the embryonic stages of this dissertation and who helped in the conceptualization and identification of the research problem, deserves much gratitude, as does Professor James Seale for sharing his rich technical skills and philosophizing on techniques and context. I extend a special appreciation to University of Florida President John V. Lombardi, who took time out of his busy schedule to serve on the committee.

My sincere thanks go to Professor Keith Andrews, Director of the Pan-American School of Agriculture in Honduras, for providing fieldwork support and guidance. All the extension agents and staff at Zamorano have my gratitude for their assistance and friendship. I am indebted to the enumerators for turning tedium into fun and for relating so successfully with farmers. The farmers who participated in the survey merit particular appreciation for their time and trust.

I am thankful for my brothers, Jim, Dan and Steve who provided invaluable encouragement that reminded me of the benefits of cooperation. I am also indebted to friends here in Gainesville who provided moral support when it was crucially needed.


iv















TABLE OF CONTENTS


page

iv


ACKNOWLEDGMENTS ....


LIST OF TABLES ............................................ viii

LIST OF FIGURES ................................................. x

A BSTR A C T . . .. . . . .. . .. .. . .. . . .. . . . . . . . . .. . .. . .. . .. .. . . . . . . xi

CHAPTERS


1
4
5


1 INTRODUCTION ....................
Objectives .......... ...........
The Arrangement of the Dissertation .....


2 EVOLUTION OF FORCES THAT INSTITUTED HONDURAN AGRARIAN


REFORM COOPERATIVES


. . . . . . . . . . . . . . . . . . . . . . . . . 7


The Pre-Colonial Period .................................... 8
Agriculture ............................................ 8
Land Tenure and Labor .................................... 9
The Colonial Period ........................................... 10
A griculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Land Tenure and Labor ............................... 11
Independence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Honduran Independence ................................... 15
Honduran Land Tenure 1821 - 1898 ....................... 17
Honduran Land Tenure: 1898 - 1940 ...................... . 19
Campesino Unions and the Design of HARCs during the Cold War ......... .20
Cuba and Bananas ........................................ 20
The Land Reform Law of 1962 ......... ........ . ... .. 22
Decretos Leyes (Legal Decrees) Nos. 8 & 170 ................. 25
Agrarian Reform Amid Political Tumult: 1975 - 1988 ............ 28
The Land Titling Program .............................. 33
Collectivization .............................................. 34
Incentives in Collectives .............. 34
Structural Contradictions of Collectives in Agrarian Reforms .. ... . 36 HARC Collectivization: Underlying Forces ........ .. . 38


v


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









3 INITIATING EFFICIENCY GAINS: TECHNOLOGY ADOPTION ............ .40

Introduction.... ... .................................... 40
Determinants of Technology Adoption ........................... 41
Methodology ........ ..................... ................. 43
Data and Model Specification ............................ 45
Results ..... ......................... ................ 47
Summary ................................................. 50

4 TECHNICAL AND ALLOCATIVE EFFICIENCY: COLLECTIVE VS
INDIVIDUAL ..... ............ .............................. 51

Productive Efficiency ........... ... ........................... 54
Technical Efficiency .............. .................... 56
Allocative Efficiency ...................... .............. 59
Nonparametric Frontiers ......... ......................... 62
Parametric Frontiers ....... .... ..... ................... 65
Technical Efficiency vis A vis Technology Adoption .............. 69
The HARC Stochastic Frontier ................................ 71
Results ..................................................... 76
Technical and Allocative Efficiencies ............................ 80
Summary .... ................................. ........... 87

5 TRADITIONAL AND ADVANCED TECHNOLOGY: A COMPARISON OF
BEANS AND MAIZE ......................................... 89

Beans: A Traditional Crop .................................. 90
General Efficiency Comparisons: Maize and Beans .................... 98
Sum m ary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

6 THE INFLUENCE OF HUMAN AND SOCIAL CAPITAL ON TECHNICAL AND
ALLOCATIVE EFFICIENCY ................................ 103

Introduction .................................... ........... 103
Theoretical Underpinnings of Human Capital .................. 104
Empirical Applications of Human Capital Components .............. 105
Education ........................... ........ 106
Training and experience ............................. 108
Health ........................................ 109
Religion ... ........................ ........... 110
Social Capital ......................................... 111
Empirical M odel and Data ................................... 113
M aize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
B eans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Training and Education: Complements or Substitutes? .............. 131
Summ ary ................................. ............ 143


vi










7 THE NATURE OF HONDURAN AGRARIAN REFORM COOPERATIVES AS
PRIVATE ENTERPRISES ................................

Introduction ...............................................
The problem ........................................
Cooperative Enterprises and the New Institutional Economics ...........
W hy Firm s Exist .................................
Cooperatives in the NIE ................................
Contracts ..........................................
Contracts Related to HARC Basic Grain Production ............
The Standard Administrative Chart ........................
Transaction Association by Activity ......................
Internal Contracts: The Black Box of the Firm ................
Labor-labor contracts .............................
Administration-labor contracts ....................


External Contracts: Association with Support
Campesino unionizing . . . . . . . . .


Land tenure .........
Technical assistance ....
Credit .............
Marketing ..........
Vigilance Committees ........ The Design of Efficient Institutions ....
Credible Commitments .......
Non-credible Commitments .
An Empirical Observation: The Guanchfas, Summ ary ....................


Agencies


. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
. . . . . . . . . .
Cooperative . . . . . . . . . . . .


145

146 147 148 149 150
151 152 153 156 158 159 159 160 161 163 163
164 165 166 167 168 168 169 171


8 SUMMARY AND RECOMMENDATIONS .............................. 175

Design of the Analysis ............................. ........... 175
Summary of the Results ........................................ 176
Recommendations ................................ ........... 178
Future Research ........................................ . 179
The Potential Role Private Capital . . . . . . . . . . . . . . . . . . . . . . . . . 179
The Potential Role of NGOs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

APPENDICES

I DATA AND STUDY AREA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181


II FIELD SURVEY ..................

III REFERENCE LIST . . . . . . . . . . . . . . . .

IV BIOGRAPHICAL SKETCH ...........


. . . . . . . . . . . . . . . . . . . . . . . . 18 8


199


. . . . . . . . . . . . . . . . . . . . . . . . . 2 1 1


vii















LIST OF TABLES


Table Dage

2.1 Distribution of Farmland by Farm Size in Honduras, 1974 ................ 29

3.1 IPM Adoption: Pre-infestation maximum likelihood estimates .............. 49

4.1 Average function: Maize ordinary least squares regression . . . . . . . . . . . . . . . . 77

4.2 Frontier function: Maize maximum likelihood estimates half normal distribution . . . 78 4.3 Frontier function: Maize maximum likelihood estimates exponential distribution . .. 79 4.4 Technical and allocative efficiencies for maize half normal truncated distribution 81 4.5 Technical and allocative efficiencies for maize exponential truncated distribution . 83 4.6 Differences in efficiency averages between collective and individual parcels . . . . . 86 5.1 Average function: Beans ordinary least squares regression . . . . . . . . . . . . . . . . 93

5.2 Frontier function: Beans maximum likelihood estimates half normal distribution . . . 94 5.3 Frontier function: Beans maximum likelihood estimates exponential distribution . . . 95

5.4 Technical and allocative efficiencies: Beans half normal and exponential truncated
distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

5.5 Differences in allocative efficiency between an imputed standard wage and zero
wage for unpaid labor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

5.6 Average Technical and Allocative Efficiencies for Maize and Beans . . . . . . . . . . 99

5.7 Coefficients of variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

6.1 Personal and household characteristics - maize . . . . . . . . . . . . . . . . . . . . . . . . 117

6.2 Physical capital - m aize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

6.3 Social capital - m aize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120


viii








6.4 Extension methods - maize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

6.5 Experience - maize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

6.6 Personal and household characteristics - beans . . . . . . . . . . . . . . . . . . . . . . . . 125

6.7 Physical capital - beans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

6.8 Social capital - beans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

6.9 Extension - beans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

6.10 Experience methods - beans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

6.11 Lecture only - maize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

6.12 Publication only - maize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

6.13 Lecture and publication - maize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135

6.14 Lecture and visual aids - maize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

6.15 Control group - maize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137

6.16 Lecture only - beans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

6.17 Publication only - beans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

6.18 Lecture and publication - beans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

6.19 Lecture and visual aids - beans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

6.20 Control group - beans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

AL. 1. Asentamientos by name, membership and land access . . . . . . . . . . . . . . . . . . . 184

AI.2 Comparison of sample averages with national averages . . . . . . . . . . . . . . . . . . 185


ix

















LIST OF FIGURES


Figure page

2.1 Adjudicated land: 1962 - 85 ....................................... 26

4.1 Input Requirement Set ............................................ 57

4.2 Farrell Technical and Allocative Efficiency .......................... 60

7.1 Adm inistration chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

7.2 A ctivity chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157


x














Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy AGRARIAN REFORM COOPERATIVES
IN HONDURAS

By

Michael J. Martin

December, 1996


Chairperson: Timothy G. Taylor
Major Department: Food and Resource Economics

Honduran agrarian reform cooperatives (HARCs) perform poorly as economic enterprises. HARCs' inefficiency is generally attributed to their collective mode of production and to internal mismanagement. In this study, a stochastic frontier production function is estimated with cross-sectional maize production data on over 400 individual producers and 28 HARCs. "Debreu-Farrell" efficiency estimates indicate that collective production is more technically and allocatively efficient than individual production.

Frontier estimates on individual bean production data show much higher allocative efficiencies than those of maize. Input requirements for maize are more varied and advanced than those of traditional bean production and require adjustments in input mixes. That farmers are not optimally adjusting input mixes is attributed to ineffective input distribution systems.

The most prominent single factor influencing HARC efficiency is commercialization. Group mean differences show that farmers who sell more of their output demonstrate higher technical and allocative efficiencies for both maize and beans. Commercialization is also a


xi








prominent factor in inducing farmers to adopt IPM technologies, as demonstrated by a bivariate logit regression.

Technical and allocative efficiencies are compared for a variety of human and social capital group means. Extension is shown to be a substitute for education at low levels of education in improving technical and allocative efficiency. However, at higher levels of education, different types of extension methods, such as publication circulation, appear to complement general education. Education, literacy, health and experience show positive influences on efficiency.

The greater efficiency of collective parcels suggests that the internal dynamics of the institutional "black box" of the firm are sound. However, transactions with support agencies are undependable as evidenced by ineffective distribution and marketing systems. Failure of implicit transaction contracts results from disparate incentives between HARC members and agency personnel.

Collaboration with the private sector, which shares HARCs' economic aspirations, offers a means to capitalize on the economies inherent in the cooperative structure and efficiencies spurred by commercialization. The few cooperatives which contracted private capital show that HARCs can evolve into cooperatives that produce high quality products and generate investment returns that significantly increase member income and achieve social goals.


xii















CHAPTER 1
INTRODUCTION


The Honduran economy deteriorated significantly during the "lost decade" of the 1980s (see Hayes 1989) in spite of large infusions of U.S. financial aid. Per capita GDP, which grew annually at 2.2 percent between 1961 and 1980, fell at an average annual rate of 0.5 percent over the next ten years. Gross domestic investment and the real minimum wage also fell over the same period. The share of external public debt to GNP was 113.8 percent in 1991, twice the 1980 level (World Development Report, 1993). Most of those funds were invested in state and parastatal agencies that are in the process of being privatized under pressure from the IMF and the World Bank.

In 1991 Honduras' per capita GNP stood at US$ 580 (World Development Report 1993), exceeding only that of Haiti in the western hemisphere. Over half of the country's 5.3 million inhabitants live in rural areas where absolute poverty, the inability to afford enough food to meet minimal nutritional requirements, is more pronounced. About 20 percent of the population is unemployed (Europa Yearbook, 1991), but casual observation suggests a high degree of underemployment as well. Infant mortality stands at 49 per 1,000 births and stunting afflicts 34 percent of all children between the ages of two and five (World Development Report, 1993). At least 12,000 Honduran children die each year due to preventable illnesses, 25 percent of Honduran families suffer from protein deficiency and 62 percent show inadequate caloric intake (Barry and Norsworthy, 1990). Seventy-three percent of the population is literate (World


1









2

Development Report, 1993), but functional literacy - the capacity to read sufficiently well to cope with modern instructions - is probably lower.

Unlike its Central American neighbors, no powerful oligarchy emerged in Honduras' history as the dominant political and economic force to beget extremes of wealth and poverty. Still, incomes and land ownership are skewed. The richest 20 percent earn 63.5 percent of total national income while the poorest 20 percent earn 2.7 percent (World Development Report, 1993). Four percent of Honduran farms comprise 56 percent of total farmland, while 64 percent of the farms comprise nine percent of the farmland (Barry and Norsworthy, 1990). Such distribution may be acceptable in developed countries, but the distress of resource maldistribution in an agrarian based economy is inversely related to the size of a nation's resource base and directly related to the size of its population. It is interesting to note that only one in five Honduran farms is worked by its owner. The majority of farmers work under sharecropping or tenancy arrangements, or on municipal lands termed ejidos (Kurian, 1987).

In an attempt to redress the problem of land maldistribution and rural poverty, Honduran governments have made sporadic attempts at land reform. Most farmers who wanted to obtain land under reforms were obligated to join an asentamiento (precooperative) and work collectively with other land reform participants in order to qualify for credit. By 1984, about 294,422 hectares of land had been allotted to land reform groups. Total membership in the groups, however, declined from 61,176, counted at the time of land apportionment, to 48,129 in 1984 (Instituto Nacional Agrario, 1985).

Cooperatives are ostensibly preferred because they permit independent farmers to jointly make investments in production, storage and marketing activities that they would be unable to make individually, and because they provide a sense of security to members who individually can be devastated by the uncertainties of agriculture. Cooperatives also serve as educational










3

institutions and provide an efficient structure for development agencies to perceive production constraints and rural needs so that responses can be formulated in the form of credit and technical assistance. Cooperatives are thus intended to achieve both economic and social goals.

However, Honduran agrarian reform cooperatives (HARCs) often fall into bankruptcy or never achieve a position of financial independence even though ample financial and material support have been provided. In 1982, 57 percent of the loans to 'he reform sector were delinquent (Stringer, 1989). Blame for cooperative failure is commonly placed on members, and accusations are heard of corruption, non-cooperation and apathy. But such problems, or at least the potential for them, are present to some degree in all firms. Given the extreme poverty in Honduras, and the physical and mental frailty that inevitably accompanies it, cooperative failure may result from an underinvestment in human capital and institutional flaws which fail to account for human capital deficiencies.

HARCs have a notorious record of inefficiency, owed principally to the high rate of loan defaults. The precise nature of the inefficiency, however, has never been rigorously analyzed. An abundance of conjectural notions attribute inefficiency to ignorance, corruption or culture. Even development planners and social scientists are quick to discount cooperatives as infeasible forms of association, in spite of their theoretical advantages.

It is possible that cooperatives do capture some forms of efficiency better than, or at least as well as, other forms of organization, but due to persistent bankruptcy they are rejected in whole. Do cooperatives fail because production is inadequate? Are cooperatives allocatively inefficient, investing in inputs that do not generate a justifiable value with respect to their marginal addition to production? Or are they both technically and allocatively efficient, but experience financial collapse because they are poorly managed? Distinguishing services that coops can perform successfully vis A vis those which they cannot will suggest specific aspects of









4

policy measures that hold a higher probability of solidifying sustainable cooperative enterprises.




Objectives


This dissertation endeavors to examine how HARCs can become more "efficient." The concept of efficiency is multi-faceted, related to physical input-output relationships, human capital investments. internal firm organization, market operations and relations with public and private organizations. The lines between each facet are not always conveniently distinct for the purpose of analysis. Efficiency concerns of HARCs are intuitively characterized in three areas: technology adoption, a requisite for improving technical efficiency; production efficiency in terms of both physical input-output, or "technical," efficiency and price or "allocative" efficiency; and institutional efficiency, or the rules that govern HARC operations. All are examined vis A vis human capital investments upon which HARCs both depend and are mandated to augment.

Although HARCs have had a notably poor record in achieving financial viability, little research has been done concerning the possible means of rectifying charaLteristic problems. Thus the primary objective of this dissertation is to identify efficiency weaknesses in HARC operations. Additionally, this dissertation contributes to the general knowledge of agricultural development concerning aspects of cooperative organizations that depend on human capital investments and broadens the scope of empirical cooperative and human capital studies.

The following objectives are designed to identify aspects of institutional behavior and human capital investments characteristic of Honduran cooperatives so that policies may be drawn to improve economic performance:

1. Review the evolution of forces that institutionalized HARCs.










5

2. Examine how human capital, extension methods and demographic characteristics
influence the adoption of integrated pest management (IPM) techniques.

3. Compare the technical and allocative efficiencies of individual production systems vis
A vis collective production systems.

4. Examine how human capital, extension methods and demographic characteristics
influence technical and allocative efficiency of individual production.

5. Examine the complementarity of education and varying extension methods in
improving technical and allocative efficiency.

6. Synthesize the results in a new institutional economics (NIE) rubric in a manner which
yields viable policy alternatives.



The Arrangement of the Dissertation


The chapters below probe three principle underlying facets of HARC operations. Chapter 2 reviews the evolution of forces that instituted HARCs. How humai capital influences the adoption of advanced technology, an intermediate step in improving and modernizing agriculture, is examined in Chapter 3.

Chapter 4 presents a stochastic frontier model and empirical results of collective and individual production systems. Collectivization has been indirectly mandated within the reform sector as it has traditionally been a requirement for HARCs to receive credit and support services from government agencies and because ownership of individual parcels has been precluded by land reform legislation. Technical and allocative efficiencies are compared for collective and individual parcels. A stochastic frontier estimation of bean production is presented in Chapter 5 and compared with maize estimations presented in Chapter 4. Beans are grown using a more traditional technology than maize.

Human and social capital deficiencies are often cited as sources of inefficiency within HARCs. Human and social capital are examined directly in Chapter 6 as they pertain to










6

allocative and technical efficiency. Chapter 6 further examines the complementarity of education and extension techniques in improving technical and allocative efficiency.

Finally, empirical results are synthesized in Chapter 7 and interpreted within the broader discrete institutional framework in which HARCs operate. The gathering of data involved extensive fieldwork, dozens of cooperative meetings and contact with government support agencies. Relating empirical results to general observations of the operating environment is a further means of capitalizing on direct fieldwork. The intention is to identify short-run modifications in HARC governance to account for human capital shortages that generally require long-run investments.
















CHAPTER 2
EVOLUTION OF FORCES THAT INSTITUTED HONDURAN AGRARIAN REFORM COOPERATIVES (HARCS)


Agrarian structures in Central America have evolved over the last 500 years in response to political and economic shocks. Unlike North America, where agricultural settlements completely supplanted Indian communities, Hispanic colonization incorporated indigenous peoples to extract scarce labor and foodstuffs.

Aspects of current Central American agricultural organization can be traced back to colonial and even pre-colonial eras. Placing the current system within the context of political history reveals institutional transactions (Bromley, 1989) that served to minimize transaction costs (Coase, 1937 and Williamson, 1985) for interests that influenced Honduran agrarian reform cooperatives (HARCS). Institutions that ostensibly serve HARCs warrant reevaluation in the post-Cold War world to identify structures that misaligned incentives with enforceable responsibilities. Transaction costs may be reduced by reorganization within the constraints of available technology and capital investments, major focuses of this study.

The purpose of this chapter is to review the forces that both motivated and constrained Honduran agrarian reform throughout history. Those forces explain peculiarities in the design of HARCs, relative to the organization of classical economic firms, that provide a context within which to interpret empirical observations.


7










8


The Pre-Colonial Period


The ancient ethnic conflicts persist in Guatemala and southern Mexico are relatively absent in Honduras. However, a review of pre-colonial institutions offers insight into the Spanish colonial framework that was established to control indigenous populations and the post-colonial tenure systems that emerged.


Agriculture


A variety of Indian cultures inhabited Honduras when the Spanish initiated their conquest in the sixteenth century. Historical evidence is available for the Mayan ;ociety, which stretched from western Honduras north to the Yucatan. Information is scarce regarding Indian societies that resided in the central and eastern regions.1 However, a few reliable observations about the pre-colonial Honduran agricultural economy may be obtained from previous studies.

Maize and beans, which continue to form the basic staples of Central American diets. were cultivated in all areas of Honduras along with manioc and sweet potato. Except for a few small irrigated systems in the Mayan region, farming in Honduras was rainfed. Remnants of gardens and orchards have been found, especially along the banks of rivers where flood waters brought natural fertilization. Only the mute dog and the turkey were domesticated for subsistence. Tribes in eastern Honduras were more reliant on hunting and gathering and were generally smaller and more dispersed than those in western and central Honduras.

Human carriers and canoes formed the only modes of transport. Central American Indians used no animal traction and, in spite of impressive Mayan intellectual achievements in




'A discovery of a burial tomb of "glowing skulls" was recently made in the area, but research results are still being processed.









9

mathematics, calendrics, architecture and astronomy, never developed the wheel. Trade was thus limited primarily to clothing goods such as fabrics and feathers and light agricultural products.

Swidden agriculture, a constrained form of which is still practiced in Central America, was the predominant farming method. This is an extensive system of cultivation which generally requires at least two years of fallow for soil rejuvenation. Land was cleared by hand with stone axes and cultivated with hoes and digging sticks (the barreta, which is .;till in use). Typically, farmers abandoned one site after cultivation and slashed and burned the debris from another, the fertility of which had been restored by time and natural vegetation. The length of the rejuvenation cycle varied across regions, but at least twice as much fallow land was required to sustain output.

Institutional land constraints have deprived most contemporary Central American farmers of the rejuvenation portion of the cycle, although debris from the previous season's crops are still burned. Rising populations have further tightened those constraints, contributing to acute social and political conflicts.


Land Tenure and Labor


The Mayan Empire was never ruled by one authority, but their :enters were larger and more advanced in terms of abstract knowledge, technology and economic diversification than other contemporary indigenous populations in Honduras. The degree (f centralization and the bases of economic support and activity varied across settlements and time. The most fundamental organization of ancient Mayan society was, as it continues to be, the nuclear family.

Morley et al. (1983) argue that as class distinctions widened, the Mayan elites coerced peasant classes into a feudal relationship whereby the peasants were apportioned land to cultivate, the product of which was shared with the nobles. In exchange for their igricultural production,









10

as well as military service and non-agricultural labor, the peasants may 'iave received protection from outside raiders.

The chiefdoms in western and central Honduras were primarily dependent on subsistence agriculture. They were not as diversified as the Maya, but nobles employed commoners or slaves for the goods and services they consumed (Torquemada, 1723). The smaller tribes in eastern Honduras were more egalitarian; the only division of labor was based on sex and age. Men undertook the clearing of lands and hunting and fishing, while agriculture and child rearing was primarily entrusted to women.


The Colonial Period


Although Honduras was discovered by Columbus by 1502, colonization was not effectively imposed for decades. Other conquests in the new world offered greater reward at less expense, owed principally to Honduras' rough terrain, relative lack of processed precious minerals and lack of broad social structure. The riches of the Aztec and Inca made the opportunity costs of conquering the scattered tribes of Honduras prohibitive to the conquistadores. Nonetheless, colonization eventually came and obliterated aboriginal chiefdoms and tribes in Honduras. Newson (1986) speculates that by the end of colonial rule, over 90 percent of the Indian population had been annihilated by slavery, forced labor and disease. Miscegenation further eroded Indian communities, giving rise to the predominance of Honduras' mestizo population.


Agriculture


The Spaniards introduced new crops and agricultural methods, but the technology of production and the crops grown on Indian lands changed little. Indigenous foods of maize and









11

beans became a fundamental part of the colonialists' diet. The Spanish either produced their own conventional grains of wheat and rice, or extracted them as tribute from Indians who had to be supplied with implements and draft animals. Converse to current production patterns, wheat production expanded in colonial Honduras to meet domestic market demands, though rice did not. Rice now constitutes a staple and is produced throughout Central America. Virtually no wheat is grown in Central America, though wheat products are becoming increasingly popular, owed principally to PL-480 shipments. Colonists also successfully introduced bananas and citrus products, commodities that now form the backbone of the Honduran export economy.

Agriculture in colonial Honduras was dominated by livestock production. Cattle, horses and mules found a compatible new environment in the disease and predator-free high grasslands of the Honduran savanna. The mule and horse industry flourished to satisfy demands in agricultural production, mining and, most importantly, in transport. Indigo, hides and tabasco were produced for export to Spain, but were minuscule relative to mining operations (Newson, 1986).


Land Tenure and Labor


The design of colonial institutions was endogenous to the accessibility of indigenous populations and the nature of social structures, particularly with respect to labor hierarchy and land tenure, that confronted the Spanish. Tribal divisions and warfare, which had been accelerated by the Quiche overextending its dominance to other Central American tribes, also provided incentives for tribal leaders to collaborate with the militarily s iperior Spaniards.

According to Newson (1986), the Spanish extracted slaves from easily accessible lowland areas in Central America to work in colonial mining operations. In the relatively remote highlands, the existence of a jurisdictional structure provided the conquitadores with a low cost









12

means to secure a ready labor supply. The Crown granted encomiendas to colonists that in essence conferred property rights over the Indians while ostensibly providing the Indians with protection and instruction in the Catholic faith. In exchange, colonists were allowed to levy tributes in the form of goods or money and, initially, labor, on the Indians.

Agriculture initially played a subsidiary role to mining in the colonial economy of Honduras. Colonists turned increasingly to agriculture as urban living expenses spiraled and opportunities for wealth creation declined in urban areas (Macleod, 1983). High food prices also served as an inducement to establish agricultural enterprises and supply urban dwellers (Frank, 1979). Tributes, exacted through encomenderos that often exceeded legal bounds, constituted the most important source of Crown revenue.

Rampant violations of colonial duty and the abdication of protection obligations forced the Crown to transfer tribute exaction to local administrators who established labor quotas (repartamientos) in indigenous villages. By that time, disease and forced slavery had taken a serious toll on Indian populations, and thus on the available labor supply. Colonists then divided up Indian labor for not just public and religious institutions, which had been the original intention of the repartamiento, but also for private individuals. The work was so arduous and the pay so minimal, laborers suffered malnourishment and were unable to work their own land adequately. Poor diets for the Indian workers and their families further weakened their resistance to prevailing epidemics.

Encomiendas and repartamientos served as viable tools of enfo:cement in areas where Indian states and chiefdoms already existed because most Indians wero accustomed to paying



2"Encomienda" comes from the Spanish word encomendar, which means "to entrust."

3"Repartamiento" is derived from the Spanish word repartir, whi-h means "to divide" or "to distribute."









13

tributes to authorities and because the hierarchical structure provided a low cost means of exercising control over large populations. In areas absent of large social structures where nomadic tribes subsisted on swidden farming, hunting and gathering, the incorporation of Indians into the colonial economy was cost prohibitive. The Crown charged Catholic Missions with the task of training Indians how to function in the culture that had over-powered them. However, small disparate tribes were only brought under control when their attacks caused costly interference in Spanish commerce. In such instances Indians were routinely exterminated or captured for slavery.

Pre-colombian Indian land holdings were technically honored by royal decree. In reality, however, Spaniards either ignored the decree or circumvented it by various means. The very act of proclaiming property rights to indigenous peoples was a defacto expropriation of the land they inhabited. Ultimate ownership rested with the Crown and was exercised on its behalf through judicial appointees for mutual economic benefit.

Land was offered to Spaniards as an incentive for colonization and the conquest of the Americas and came to be a valuable source of government revenues necessary to protect royal property. Martinez Peldez (1975) contends that the Crown extended property rights to the Indians as a means of ensuring their presence to pay the agricultural tr-butes needed to supply mining activities and urban lifestyles.

In 1591, the Crown instituted two portentous cedulas (decrees) affecting land tenure. The first, which ostensibly would have benefitted the Indians, ordered all lands illegally usurped to be returned to the Crown. The second cedula, however, granted that illegally owned lands could be rightfully purchased. Colonists, by virtue of their much greater wealth and knowledge of Spanish law relative to the Indians, were able to circumvent the first cedula to benefit from the second. Redistribution of power or influence was confined to Spanish appointees or









14

entrepreneurial creoles who accumulated enough capital to invest in lucrative official positions (Brading, 1986). This skewed distribution of land prevails in Honduras.

The encomienda system lasted longer in Central America than in any other area of Spanish America (Newson, 1986). The current demographic complexicn of Honduras began to take shape as the Indian population was surpassed by the mestizos through slavery, disease and miscegenation. Royal Decrees offered no property rights to mestizos as it had to Indians. Latifundios expanded through the exploitations of mestizos, who were forced to devote portions of their production in exchange for usufruct privileges (Martinez Peldez, 1975).


Independence


Latin American revolutions in the 19th century were inspired by their North American counterparts and facilitated by Napoleon's incursions into Spain. However, while much of the independence rhetoric was similar, the Spanish aristocracy continue. to rule. Apart from forswearing formal titles that eliminated colonial taxes, little else changed for the majority of the people.

Central America was polarized between two forces that would continue to squabble, at great expense to economic development, into the twentieth century. The conservatives arose from Spanish aristocracy and sought to maintain colonial institutions and sccial classes without the burden of royal tribute. The liberals, comprised mostly of elite bureaucratic professionals that formerly served the aristocracy, saw economic benefit in reforms that expanded economic opportunity.

A major source of contention between liberals and conservative! concerned the Church as a formal component of government. The Church conferred moral justification on the privileges enjoyed by elite conservatives and was instrumental in quelling the masses. Perhaps









15

most destructive to long-run stability and development was the disp:ite on education. The conservatives wanted to maintain an elitist system of education under the auspices of the Church, while the liberals promoted secular, and ultimately mass, education. Like the Crown, conservatives claimed and exercised the right to make ecclesiastical appointments (Barnadas, 1986).

Liberals sought to reduce trade barriers and eliminate monopoly rights that conservatives held by virtue of their royal commission. Liberals also wanted to ban the nepotism of conservatives in governmental, commercial and ecclesiastical positions.

Both parties were comprised of elite, urban, intellectual interests and encompassed only a small portion of the total population. Notions of broad-based utilitarianism, promoted most notably by Jeremy Bentham, provided a philosophical rallying force to revolt against Imperial Spain. However, the exact terms of "utilitarian" economic design did not result from such popular consensus.


Honduran Independence


Honduras experienced three independences in the frenzied rebellions that swept the Spanish Americas in the nineteenth century. It was first liberated from Spain as part of what was to be the Kingdom of Mexico. Anarchy in Mexico afforded the Aud.ncia de Guatemala an opportunity to institute its own autonomous government, Provincias Unidades del Centro de America. With the exception of Chiapas, all the states of the Audencia revolted against Mexico and joined the Provincias.

Central American unity, however, was hampered by the surge of hemispheric independence that permeated all levels of society. Motivations for local participation and autonomy were born of nascent ideals and economic opportunity. Powerful regional caudillos









16

(political/military strongmen) in each state provided conservatives the fo:ces needed to defeat the liberals, led by Francisco Morazan, but they ultimately dissolved Central America into five independent states. Independence for Honduras was thus a by-product of war strategy rather than a popular consensus based on cultural heritage or economic principles.

Elections notwithstanding, caudillos ruled Honduras from 1839 until the mid twentieth century. The private acquisition of public and Church lands, a hallmjark of liberal economic programs intended to stimulate exports and growth, aroused a symbiotic alliance between colonial elites and the Church on the one hand, and Ladino and Indian communities on the other. Conservative forces respected Church property rights and supported preservation of ejido lands for subsistence farmers, who constituted the majority of the population. Economic growth slowed under conservative rule, the country returned to traditions that reflected traditional HispanicCatholic principles, and subsistence farming gained in prominence vis vis export agriculture.

Apprehensions about capitalism also arose from the lack of government capacity to administer a broad-based development agenda. Liberals attempted to opmn alarming proportions of property to foreign investment. A debilitating shadow over liberal economics was cast by William Walker, a U.S. citizen who, in the name of democracy and economic freedom, proclaimed himself president of Nicaragua. His group of predominantly Mississippi Valley mercenaries recruited after the Mexican-American War expected to receive land in return for their services and were extremely unpopular in Honduras, which had proclaimed war on the United States to defend Mexico. They represented the detrimental excess of economic liberalization: foreign usurpation of domestic resources and political control. Walker, who was executed in Honduras after a failed filibuster, severely discredited Liberal policies throughout Central America and intensified anti-American sentiments that prevail tc this day.









17


Honduran Land Tenure 1821 - 1898


The land tenure patterns that evolved in post-independence Latin America differed substantially from those in North America. The United States and Canada opened their frontiers to settlers, allowing for a more equitable distribution of wealth and ensuring property rights to people within the political and economic system (ethical considerations regarding the usurpation of Indian lands aside). Property rights were much more tenuous in Latin America, owed to the continuation of ejido lands and absentee landlordship of large holdings. Agricultural technology lagged in Latin America, restricting labor productivity and the consequent labor and food surpluses exigent to industrial growth.

The industrial revolution in the mid 19th century that transformed the global economy influenced Honduras' Central American neighbors much more than Honcuras. Recorded history and data on the 19th century Central American economy are sparse, but some policies and especially the reactions to them reveal distinguishing features of Honduras. Coffee plantations arose in Guatemala, El Salvador and Costa Rica in response to rising world demand, but did not take hold in Honduras until the twentieth century.

Liberal reforms regained prominence among Honduran political philosophers who also contended that, unlike the laissez faire reforms promoted by Francisco Morazan, government had a pivotal role to play in economic development. Lacking a bourgeois class, Honduran economic strategists considered government necessary to stimulate growth and secure integration with the world economy (IHDER, 1980). A new constitution in 1880 mandated the state to establish credit for Honduran enterprises and provide the infrastructure necessary to attract foreign investment.

Liberal reform began in 1876 when President Jos6 Marfa Medina, under economic liberalization pressures from Guatemala, yielded the chief post to Marco Soto (Yankelvich, 1988).









18

Soto considered ejido structures inefficient and in 1877 instituted an agrarian law that slated coffee, cacao and Indian rubber as favored exports (Guevara, 1983). The intent of the reform was to propel Honduran agriculture into modern specialized production. Central to the agrarian law was the dictate that municipalities sell any national lands to bidding entrepreneurs at a "just price." The government took an active role in designing and monitoring the agricultural system. Landowners and laborers of "modern farming units" were exempted from military and civil obligations and the tariffs on imported agricultural inputs. The Government established strict guidelines regarding such things as fencing, crops and laborers that constituted modern farming units. Participating farmers were required to make written reports to demonstrate their compliance every six months.

Interventions to stimulate the production of export crops failed. Reasons for the failure are not clear. Coffee, which initially was considered the crop of greatest potential, began a longterm decline in world demand soon after the program was mounted (Gue-ara, 1983). Also, labor was not as inexpensive relative to other Central American countries, partially because some ejido lands remained inalienable (IHDER, 1980; and Quifiones and Argueta. 1978). Rugged terrain and poor transportation and communication systems also hindered commerce.

Government support for transportation infrastructure was hampered by the enormous debt it incurred from trying to build a transcontinental railroad earlier in the century. In 1892 the total debt was over 73 times annual export volumes. Consequently, foreign investors were reluctant to invest in Honduras. The failure to attract upstanding foreign investors rendered Honduras vulnerable to riskier international investors who promised great returns but did nothing more than further deplete the treasury. The bulk of Honduran agriculture thus remained a decentralized system of traditional subsistence farming into the 20th century (Guevara, 1983).









19


Honduran Land Tenure 1898 - 1940


Banana exports began growing along the north coast of Honduras around 1880. Most of the production came from small, independent producers who sold their produce to intermediaries close to the coast or with river transport vessels. Banana production grew in popularity among small commercial farmers. Unlike coffee, the favored export of the government, bananas could be grown with relatively little capital investment and along rivers that enabled transport. Most important, bananas presaged short-run gains.

The liberal reforms' hapless attempts to establish a bourgeois class in the 19th century were dealt a decisive blow by the emergence of international banana companies. Twentieth century commodity production, storage and transportation required capital investments beyond the reach of most Hondurans and certainly beyond that of small-scale farmers engaged in banana production. Ultimately, foreigners established large banana plantations that sold directly to exporters, thus undercutting both small Honduran producers and intermediary transporters (Guevera, 1983).

Honduran land reforms during 1898-1961 reserved public lands for family parcels. The Honduran Institute of Rural Development (IHDER, 1980) identified two characteristics of land reforms during that period. First, the reforms established inalienable property rights, the intent of which was to impede land concentration and proletarianization of rural labor. The 1924 decree stipulated that family parcels be 20 hectares (Stokes, 1947). Second, because reform parcels were small and used primarily for subsistence crops requiring only about .00 labor days annually, land reforms provided an abundant, low wage seasonal labor force. The motivations for the various decrees during this period have been attributed to the government's desire to spur economic growth and generate tax revenue (Villanueva, 1968) and to ameliorate lay-offs in the banana industry (IHDER, 1980).








20

Most of the pre-Cold War land reform debate was rooted in disputes between campesino unions and large American fruit companies. Land reform initiated in the basic grain sector is distinct in many ways from land reform in the banana sector. Integral to both sectors were the campesino unions that fused political strength among the landless to pressure for land redistribution.


Campesino Unions and the Design of HARCs during the Cold War


The first campesino unions appeared around the late 1920s. Financed and influenced by the communist party (Posas, 1981), they were characterized by the inflammatory Marxist rhetoric that began fueling the reactionary backlashes to a broad range of grassroots social organizations characteristic of, and crucial to. open democratic societies. Campesino unions were influenced by national and international forces which, particularly throughout the Cold War, had enormous leverage over the design of HARCs. Those forces arose from political exigencies that, unlike the forces that forge autonomous cooperatives, did not necessarily hold economic efficiency - or even viability - as a guiding principle.


Cuba and Bananas


Two events in the 1950s had dynamic impacts on Honduran agrarian reform, the Cuban revolution and the lay-off of over 18,000 laborers on banana plantations. Labor organizing was already afoot in Honduras prior to the Cuban revolution. Labor and farm organizations were inspired significantly by Cuba, as were academic institutions that could provide scarce technical and logistic support for mobilization activities. Massive floods and an unprecedented strike in 1954 in which 25,000 workers participated gave United Brands incentive to restructure by closing operations on marginal lands and by making labor-saving capital investments. The company








21

reduced its labor force by 69 percent between 1954 and 1963 (Posas, 1987). Those workers were forced to return to subsistence production, but they returned with a new form of human and social capital embodied in the capacity to organize for worker rights.

Campesino unions modeled their organization on urban and banana company union mentors, whose objectives and optimal strategies were different from landless farmers. Urban and plantation workers tend to be more concentrated than campesinos, allowing for relatively low costs of coordinating activities. It was easier to define and achieve consensus on the goals among urban and plantation union members than among campesinos. Association with and support of internationally polarized "labor" organizations also drew campesinos into conflicts about which they had little understanding and to which, by virtue of their geographical dispersion and illiteracy, their interests were more vulnerable.

Two principle campesino unions emerged in Honduras at the inception of the Cold War. The first, the National Federation of Honduran Campesinos (FENACH, was an offshoot of the unionizing apparatus of the 1954 strike. Although FENACH was spawned by autonomous local organizations largely devoid of revolutionary ideology, two of its three principle leaders were militant communists (Posas, 1987). Within two months of the founding of FENACH in 1962, another, less radical, campesino organization appeared, the National Association of Honduran Campesinos (ANACH). ANACH could claim some lineage to the grassroots 1954 organizations, but it also received support from the American Institute for Free Labor Development (AIFLD), an AFL-CIO affiliate that received funding from the U.S. Agency for International Development (USAID) and from its parent organization, the Inter-American Regional Organization of Workers (ORIT). ANACH was clearly a union established by foreign funding as : moderating alternative to FENACH.








22

Both ANACH and FENACH applied for personerfajuridica, which technically confers legal recognition on individuals or organizations. In practice it is used as an institutional mechanism to support - or leave unmolestedthose in the government's favor. ANACH was granted personerfa juridica two months after solicitation. FENACH, whose principle disputes were with the Tela Railroad Company (a subsidiary of United Bran's), was never granted personerta juridica (Posas, 1981).

One would be hard-pressed to argue, given the collapse of communism, that Marxist intellectuals' abstract notions of an ideal society, and particularly its attainment, were in sync with the aspirations of uneducated laborers and farmers. On the other hand, opponents of communism, however defined, reacted in Honduras and throughout L-tin America on several well documented occasions by means contrary to the democratic ideals which they espoused. Cold-War reactions were also evidenced more subtly in development assistance programs. Honduran agrarian reform cooperatives became tools of partisan international concerns which relegated the long-term benefit of the HARCs subordinate to the attainment of political goals.


The Land Reform Law of 1962


The Cuban revolution prompted anti-communist forces within Latin America and the United States to respond to landless farm workers as a means of quelling violent uprisings in the western hemisphere. The Alliance for Progress, accorded at the Inter itmerican Conference at Punta del Este in 1961, stipulated that aid recipient countries must mount land reform programs in order to qualify for assistance.

Though motivated by political exigencies, land reform also had economic underpinnings. The redistribution of idle lands was consistant with the Kaldor-Hicks compensation principle in terms of wealth in that resources began generating substantive economic gains; a potential transfer








23

could compensate absentee landowners for the negligible utility derived from owning but not utilizing the land. Redistribution also constituted a stricter Pareto improvement in terms of income in that the former owners derived no income from idle lands.

The Alliance induced democratically elected President Ram6n Villeda Morales to establish the National Agrarian Institute (INA) in 1961. INA was mandated to respond to land petitions and coordinate land distribution. It was also charged with providing support services, principally the organization and support of cooperatives.

The Villeda government passed the first land reform law in 1962 that slated not only government lands for distribution, but idle private lands as well, thus requiring all lands to serve social purposes. Unused lands were ostensibly subjected to progressive taxation or expropriation, but ambiguously written laws and ineffectual enforcement capacity mitigated the threat to vigilant landowners or land holding companies (IHDER, 1980). Property rights have never been clearly defined in Honduras. To the extent that a real-estate market exists in Honduras, INA and the judiciary serve as its clearing houses.

Latifundistas could also legally activate idle lands by establishing cattle ranching operations, a convenient activity for absentee landlords that resulted in a threefold increase in beef production while domestic consumption of beef declined (Monthly Review, 1985). Transferring pastures to crops would apparently be a much more efficient use of the land in terms of income, employment and foreign exchange earnings or savings (Garcfa et al. 1988).

Growing popularity of structural reforms, legitimized by the Alliance for Progress, pressured Villeda on one side while the landed status quo and banana companies pressured him from the other. The opposing forces destabilized a balance that was vulnerable to interference by a virtually autonomous military. Villeda's attempts to weaker the military through institutional reform enraged commanders who were encouraged by dictators in Nicaragua and the








24

Dominican Republic to overthrow the government (Morris, 1984). Subsequently, aspirations for land reform were cut short by the coup d'etat in October, 1963, led by Colonel Oswaldo LUpez Arrellano.

Campesino unions' expanding influence had apparently posed a greater threat than the 1962 agrarian reform. The military did not reverse the agrarian reform law, but weakened INA and destroyed FENACH by ransacking its offices and assassinating its leaders. Consequently, the slow rate at which land was being distributed was halved (Bueso, 1987).

LUpez's tenure as dictator and later as a not-so-freely elected president was marred by traditional bureaucratic corruption. Forces arose to effectively challenge the traditional caudillo politics. Labor unions and business organizations recognized mutual benefit in extending government participation to all sectors of the Honduran public. A severe hurricane and the infamous "football war" with El Salvador mobilized Honduran society t.o work collectively for the first time, propagating the wide-spread recognition of governmental rights. The agrarian reform of Peru had also inspired younger military personnel, many of whom came from campesino families and empathized with agrarian concerns.

In 1971 Colonel L6pez ostensibly relinquished control of the political forum to democratic elections. The campaign of 1971 brought Ram6n Ernesto Ciuz to the presidency for a brief interlude from military rule. Cruz represented the conservative National Party, which owed its strength to its alliance with, and the campaign management of, the military.

The Cruz administration prohibited land "invasions", a dangerous and final recourse campesinos had to demonstrate the acuteness of their poverty and thus their resolve. Six campesinos who had invaded land in the department of Olancho were assassinated and several others wounded and incarcerated in early 1972. Campesino unions contended land seizures were "recuperations" rather than invasions because such land was actually government or ejidal land








25

illegally expropriated by latifundistas through forged documents or outright bribery. Sensing their alienation from the real political processes., ANACH planned a massive hunger march and other actions that threatened social stability. In December the military restored order by peacefully relieving Cruz of the presidency and reinstalling General Oswaldo Ldpez Arellano as chief of state.

Ldpez returned to power cognizant of the need to broaden his political base. At a huge rally shortly before the coup he emphasized that "The Armed Forces a-e composed of workers and campesinos....., the Armed Forces are not enemies of the workers and the campesinos." L6pez, who arose from campesino stock, proceeded to carry out the most extensive agrarian reform in Honduran history.


Decretos Leyes (Legal Decrees) Nos. 8 & 170


Within a month after assuming power, L6pez promulgated terrporary agrarian reform legislation, Law Decree Number 8 (popularly referred to as Decreto No. 8). Decreto No. 8 was billed as an emergency two-year measure necessary to respond to campesino needs until longterm legislation could be designed. Unlike previous agrarian laws that were rendered ineffective through myriad loopholes and ambiguities, Decreto No. 8 bestowed clear, decisive power on INA to grant access to lands that were not fulfilling a social purpose. Consequently, the pace of land adjudication accelerated dramatically (Figure 2.1). The purpose of Decreto No. 8 was to incorporate campesinos into the economy and give willing workers the opportunity to improve their livelihoods.









































































G




t'861- z9e6 oesL. eL6L 9L6k


9- Z961 :puel p unpnfpV I~ -Z aaiC86 - '8dOHI e:D-AnoS tL6L ZL6L OLEL. 996L~ 996L t9E Z96L


.-GIq2jV p.JQ 12t0j


9z


... ... .. ... ... .. ... .\ .. .. . . .. . . . . . .. . . . . . . . .t- - ------ - * ---- --- -------------- - -- -----. ......... - - - - ---- -- - - - ---


000'01000 0



0'0E 000 'OE. 000 '0t








27

The declared "temporary" nature of Decreto No. 8 and the swiftness with which it was implemented reduced the potential for opponents to organize and contributed greatly to its impact. More than twice as much total and arable land was adjudicated under the scheduled two year enforcement of Decreto No. 8 than the previous ten years under the agrarian reform of 1962. Campesino mobilization is considered the pivotal factor that precipitated the reform. 92 percent of the groups had to pressure INA in extraordinary ways to obtain land. Over 23,000, or 15 percent of all rural landless families, (about 140,000 people) were granted land-use rights to 108,496 manzanas (IHDER, 1980).

Upon expiration of Decreto No. 8 a new agrarian decree, Decreto No. 170 was enacted. Decreto No. 170 consisted of 180 articles that in practice marked a return to legal sidestepping and thus a deceleration in reform. Decreto No. 170 granted permanenc. to the temporary landuse permits granted under Decreto No. 8 and maintained the authority to expropriate idle lands. However, Decreto No. 170 also denied legal recourse to campesinos that obtained land by invasion, a principle means of procuring land under lethargic government agencies.

Concomitant with the surge of land adjudications was the growth in public credit to serve the increasing number of asentamientos4 in the reform sector. The low rate of loan repayment to public lending institutions has come to be one of the major criticisms of HARCs. However, large farmers have exhibited equivalently high default rates (Lele 1974. IBRD 1975). In 1980, the total amount of non-payment in the nonreform sector was more than three times higher than in the reform sector (IHDER, 1980), though the non-reform sector repaid 63 cents to the Lempira compared to 56 cents for the reform sector.


4"Asentamientos are considered "pre-cooperatives" in that they Eave not been awarded personerla juridica, although they invariably belong to a campesino union that is legally recognized. Technically, for a group to qualify for status as a cooperaive, it must be granted a personera juridica. Legal recognition is reserved for those organizations that have demonstrated their good citizenship to the government.








28

Administrative corruption is not uncommon in Honduras. General L6pez was ousted from power after the Wall Street Journal alleged he had accepted a millien-dollar bribe that saved United Brands over 75 million dollars in taxes. Loan defaults become more understandable, if not excusable, under such circumstances when looked at from the perspective of a politically and economically weak campesino.


Agrarian Reform Amid Political Tumult: 1975 - 1988


Ldpez was succeeded by General Juan Alberto Melgar Castro in a peaceful coup d'etat in April 1975. Melgar's power abided in anti-reform forces intent on curbing the pace of land distribution. Over one-half of rural agricultural families were still landless after the reforms achieved under the L6pez regime, 56 percent of the land was still claimed by less than five percent of the families (Table 2. 1).








29


Table 2.1 Distribution of Farmland by Farm Size in Honduras, 1974




Farm Size (has.) No. of Farms % No. of has. %

Under 1 33,771 17.3 21,542 0.8
1 - 2 38,650 19.8 53,648 2.0

2 - 5 52,360 26.8 163,803 6.2
5 - 10 28,264 14.5 201,274 7.7

10 - 50 34,390 17.6 729,361 27.7

50 - 100 4,433 2.3 301,228 11.5

100 - 1,000 3,304 1.7 763,673 29.0

1,000+ 169 0.1 395,330 15.0

Total 195,341 100 2,62),859 100
Source: Stringer, 1984.








30

Land invasions continued with little effective response from the government. Landed interests belonging to the National Federation of Agriculturalists and Stock Raisers of Honduras (FENAGH) began an intimidation campaign against campesino unions and land reform proponents. Tensions were manifested in the summer of 1975 when, again in the Department of Olancho, ten leaders of the National Union of Campesinos (UNC), two priests and two women were murdered at the hands of ranchers and military officers. UNC has endured more violent intimidation than any other campesino union in Honduras.

Frustrated with the Melgar government's unwillingness to provide basic protection and its reluctance to enforce provisions pertaining to both idle and ejidal lands, the three national campesino unions forged an alliance, The United Campesino Front (FUNC). Melgar responded to FUNC with his plan "Operativo Reldmpago Juan A. Melgar Castro." The plan resulted in the distribution of 8,722 hectares to 3,160 families. Melgar's reluctance tc either implement legal reforms or decisively reverse them angered campesinos and the "Superior Council of the Armed Forces" (CONSUFFAA) that installed him. Melgar eventually dismissed heads of INA and the Labor Ministry who had been appointed by Ldpez Arellano and were proponents of reform, but he was too late to demonstrate control. CONSUFFAA announced Melgir's resignation in 1978 and installed General Policarpo Paz Garcfa. Paz ruled until 1982, during which time few land requests were processed.

Characterization of campesino unions was often based on association, not unlike the McCarthy years in the United States. UNC sprung from rural development programs sponsored by the Catholic Church and assistance from Christian Democratic Party. After the Second Vatican Council in 1964 the Catholic Church placed new emphases on ministering to the poor, which unsettled traditional power structures throughout Latin America. UNC was a part of the General Central of Workers (CGT), an activist organization backed by Christian socialists in








31

Latin America. Although founded by 1972, UNC had to wait until 1984 to obtain its personerta juridica (Posas, 1987). In another instance, Efrafn Dfaz Galeas, a former president of ANACH, riled financial sponsors by soliciting assistance from UNAH faculty, whom the sponsors considered communists. The Executive Committee of ANACH yielded to threats of funding cuts by dismissing Dfaz. Dfaz later reappeared as the president of the Federa'ion of Agrarian Reform Cooperatives (FECORAH). a campesino union that was also delayed foir years in obtaining its personeria jurfdica.

ANACH, by contrast, long the largest campesino union in Honduras, had been relatively unperturbed by the military, owed principally to its alliance with AIFLD. Generous operating budgets financed by the AFL-CIO, USAID, and a number of private corporations ranging from IT&T and Mobile Oil to Sterling Drug and Bacardi Co., (SITRAUNAH, 1979) gave ANACH the capacity to bid organizers away from grassroots unions or hire talented organizers otherwise uninterested in union mobilization. ANACH organizers were thus more financially beholden to foreign interests than to the campesinos whom they represented. Unions that arise from autonomous forces are by nature more democratic in that they must respond to the economic needs of their members. Incentives existed in ANACH, on the other hand, to subordinate the economic welfare of the HARCs to political priorities.

This is not to deny that the AFL-CIO had sincere ideological interests in preventing unions from being overcome by communist control. However, the principle source of funding for AIFLD was the US government (92 percent in 1969). Several members of the original AIFLD Administrative Board headed companies that had large financial interests in Latin America.

The strings attached to ANACH were unmasked in 1978 when Antonio Julfn M6ndez accused the sitting president of ANACH, Reyes Rodriguez Ardvalo, of fraudulent reelections.








32

M6ndez demonstrated the veracity of his charges by congregating a majority of assembly representatives in opposition to Rodriguez. A year later at an ANACH national convention in a Lion's Club hall in Siguatepeque, M6ndez again outnumbered Rodrfguez supporters 173 to 109. Upon realizing he was at a political disadvantage, Rodriguez abandoned the convention to reconvene in a military battalion. Paz Garcfa had shown his support for Rodriguez earlier by deporting a missionary priest who had denounced Rodrfguez's corruption (Posas, 1987). The military did not recognize the majority of ANACH delegates, led by M6ndez and Camilio Padilla, but ratified the Ardvalo faction. However, the M6ndez and Padilla coalition received a valuable endorsement from the Confederation of Honduran Workers. M6ndez's authority would again be unsuccessfully challenged in 1984 by an operative of General Gustavo Alvarez Martfnez, the army commandeer responsible for the civillian crackdown and several disappearances in the 1980s.

The failures and tribulations of campesino unions cannot be solely attributed to external forces. UNC and FECORAH arose from other labor organizations and, like ANACH, were always in the process of splintering themselves. The fragmenting of unions was easier to accomplish in rural-based unions than in urban or plantation unions because rural systems are necessarily less centralized. Regional and local campesino union leaders hold more member allegiance than national leaders.

Under Melgar, INA embarked on a policy that divided its resources between export crops and basic grains. The Paz regime slated eight percent of INA's total budget to the basic grain sector while designating almost 65 percent to commercial export crops. Bananas account for about 22 percent of all national value-added and occupies less than five percent of area planted to crops. Basic grains on the other hand, contribute only about 14 percent to value added even though it uses about 65 percent of cropland (Stringer, 1987). Critics charged on social grounds








33

that land reform had become a colonization project through coerced collectivization and the transferring of campesinos from their homes to areas where land was less constraining. On the other hand, allocating labor resources to a higher level of employment has distinctive economic benefits.

One persistent aspect of the post-WWII reforms was the attempt to propel the agricultural population into modern agricultural production (IHDER, 1980). Neighboring Central American countries had more advanced production and higher growth rates, but the maldistribution of wealth and income contributed to social upheaval (Bulmer-Thomas, 1989). Honduran reforms endeavored to escape that fate.


The Land Titlin2 Pro2ram


In 1981 Roberto Suazo C6rdova became Honduras' first democratically elected president in 20 years. Suazo campaigned strongly on an agrarian reform platform. The pace of land redistribution increased 75 percent over the previous four years. Nonetheless, at the end of Suazo's tenure over 125,000 rural families still had no secure access to land (Ruhl, 1989). Rural population growth alone fed the landless population faster than the land reform could reduce it. Suazo had not satisfied his political mandate with respect to land reforrr.

In 1981, at the behest of USAID, the focus of land tenure adjustments switched from outright redistribution of public or private lands to land titling. In order to qualify, a farmer must demonstrate he has worked a specific parcel of land, but that his tenure was not secure or legally binding. The primary intentions of the titling program were to provide workers of the land collateral for credit and long term incentives to make land improvements. The quality of soil and the extent to which crops are commercialized influences participation in the land titling








34

program (Seligsan and Nesman, 1989). Those factors have also been siown to be important in the adoption of technology (Martin and Taylor, 1995).

Structural adjustment measures have encouraged privatization throughout all sectors of the Honduran economy, including the agrarian reform sector. The land reform law of 1992 opened vast tracts of land for sale by cooperatives in the agrarian reform sector.


Collectivization


Beginning in 1962 land reforms in Honduras reversed the liberal tradition of distributing specific parcels to individual families, opting for a collective form ot enterprise. In theory, collectivization improves efficiency by permitting large "lumpy" investments in production, storage, marketing and input purchases that are beyond the capacity of small, capital-poor farmers. Cooperatives also theoretically reduce monitoring costs associat-d with large enterprises and lower transaction costs of providing extension and support services to large numbers of farmers. In practice however, collectives have had a poor economic record. Attrition rates have always been high in HARCs, averaging over 30 percent between 1962 and 1985 (Bueso, 1987). Loan default rates were also disproportionately higher for HARCs than for individual farmers. More than half of the National Bank of Agricultural Development's (BANADESA) loans to the reform sector were delinquent as of 1982 (Stringer, 1989). Incentives in Collectives


The conditions under which voluntary collectivization might occur are the focus of agricultural production cooperative (APC) theory (Putterman 1986, 1989; Bonin, 1987; Carter, 1987) which contrasts the behavior of collective production based on va-ying work and revenue sharing rules. APC theory customarily attributes cooperative failure to "shirking," the








35

relinquishment of individual responsibility to other members. The problem is that widespread shirking leaves many responsibilities unfulfilled and negatively impacts cooperative performance. The important question, however, is not if shirking occurs, but whether it overrides gains in efficiency arising from worker collectivization.

APC theory does not take into account the role of human capital. Two of the collective production systems often alluded to in discussions of APC theory, the Kibbutz in Israel and the Hutterite communities in the U.S., place a high value on human capital investments. Nor does APC theory - or the cooperative literature in general - distinguish the types of monitoring costs that different cooperatives may be more successful in reducing. Illiterate members cannot carryout simple bookkeeping procedures and are obviously ill-equipped to oversee complicated contractual obligations. On the other hand, landless farm laborers are capable of monitoring the unspecialized agricultural labor tasks to which they are accustomed. The shirking alleged by Alchian and Demsetz (1972), would be relatively more difficult to accomplish under such circumstances than the financial rent-seeking cited by Jensen and Mecklirg (1979). To the extent that dignity is derived from labor, shirking may be unlikely to occur. Fieldwork is one of the few endeavors available to HARC members, and one in which pride is evidenced in traditional folklore, dance and song.

The exact nature of HARC breakdown cannot be easily attributed to shirking - at least at the level of manual labor. In fact, a common charge leveled against campesinos is that they are simple people who "love to work with their machetes in their milpas (cornfields)," but that they are disinterested about the more cerebral matters of managing technologically advanced enterprises. HARCs' dismal financial history may abide in elements other than the technical inefficiency implied by shirking. Even firms that are one hundred percent technically and








36

allocatively efficient could be ruined by financial mismanagement - especially if institutions do not safeguard interest holders from "opportunism with guile" (Williamsn, 1985).


Structural Contradictions of Collectives in Agrarian Reforms


The justification for land reform sits on tenuous ground when cooperatives are promoted as part of the reform to capture economies of size. Many reform advocates contend that skewed land distribution results in inefficiency because small farmers are considered to be more efficient than large farmers (See Dorner and Kanel 1971; Berry and Cline 1979; Cornia, 1985). On the other hand, development planners impose cooperative structures in an attempt to capture efficiencies that are supposedly absent in large operations. The economic argument for agrarian reform cooperatives are thus fallacious unless large collective enterprises could attain greater efficiency than those that are individually operated.

This may be a tenable proposition for land reform cooperatives in Honduras and elsewhere in Central America. Honduran agrarian structures are still emerging from precapitalist modes of production and are often still characterized by absentee ownership and sharecropping practices that maintain traditional production methods. By the same t-ken, cooperatives have lacked the managerial skills in both production technology and administration needed to exploit the potential for economies of size. Scale economies that require specialization exact transaction costs for monitoring because production processes become less personal (Bardhan, 1989), partially, at least, explaining why sharecropping practices persist (Alchiam and Demsetz, 1972). But the cooperative naturally reduces monitoring costs because members are owners and bear


' Economics traditionally treats efficiencies obtained from the simultaneous expansion of all inputs as economies of scale. Economies of size, on the other hand, cccurs when long term, often "lumpy" and not necessarily simultaneous, investments lower the average total cost of production. Economies of size is used here because it is more comprehensive and thus more precisely represents efficiencies sought by the process of collectivization.







37

substantial peer pressure on one another. The desire to collaborate in a productive manner was quite noticeable among the majority of HARC members.

Size economies in HARCs are likely to become increasingly attainable as technological advances are adopted because long-run capital investments such as farm machinery, input inventory, transportation and storage equipment reduce per unit costs of output. Such investments provide new opportunities for coop members to exploit comparative labor advantages, underscoring the importance of the division of labor to carry out specialized tasks and a managerial workforce which can dedicate its time to identifying and implementing profitable activities. Many of those opportunities depend on labor talents that can emerge only through concomitant investments in human capital. The combination of monitoring cost reduction gained through collectivity, and the division of labor gained through human capital may thus allow for the capturing of size economies beyond the grasp of large individual farms or poorly skilled cooperatives.

Beyond efficiencies in production, collectives reduce costs incurred by service delivery agencies. New technologies inevitably require a learning process and are aggressively promoted by input suppliers who exploit the glamour appeal those technologies hold on traditional farmers. Cooperatives are efficient structures for technical and support agencies to monitor operations and provide agricultural training to large numbers of farmers. Budgets in developing countries are inadequate to employ enough support personnel and extension workers to train and monitor individual farming operations. In cooperatives training is conducted through seminars from professional instructors and, more importantly, transferred through other members (Martin and Taylor, 1995).








38


HARC Collectivization: Underlying Forces


It is difficult to discern precisely why the collective was adopted as the predominant form of operation and production in the Honduran agrarian reform sector. Stringer (1984) attributes collectivization to Roberto Sondoval Corea who was appointed Director of INA in 1968. Cardona (1979) ascribed the origins of collectivization to Jorge St. Siegens, a Romanian economist who was contracted at the recommendation of UNESCO to serve as the Technical Director and Professor of the Autonomous University of Honduras (UNAH). Posas (1987) claims Virgilio Carias, Head of the Economic Research Institute at UNAH. initially promoted the Israeli mode of production in the successful and broadly renowned Guanchfas Cooperative on the North Coast.

Although the origin of collective production is unclear, one thing is certain: HARCs have lacked proper incentive structures and enforcement mechanisms of comnracts, particularly with government support agencies, have been notoriously weak. Property assurances are also fragile. HARC members hold no individual title to specific parcels. In fact, very few HARCs technically own the land which they work collectively.

The motives for imposing collective production. through denial of credit and technical assistance to individual farmers, are seen in almost all the official organizations that support the HARCs. The government appreciates collectivization as HARCs can absorb additional farmers who are more reluctant to risk their lives invading land than to arrange an agreement with HARC members for formal admittance (IHDER, 1980). Collectivization also permits large financial transactions to be made between loan agents and a few campesino union officials, outside the








39

scrutiny of individual HARC members. "Soft-state" governments such as Honduras are vulnerable to nepotism and rent-seeking agency staff that could detrimentally exploit HARCs.

The notion of collectivization was also supported by nongovernmental entities in Honduras for political, not economic, reasons. The campesino unions maintain more political control over collectives than they could over individual farmers who, once owning a land parcel, would be less dependent on the union and more difficult to mobilize. Cold War development planners also found collectives (within a free enterprise rubric) to be a means of mollifying communist supporters in allied countries.














CHAPTER 3
INITIATING EFFICIENCY GAINS:
TECHNOLOGY ADOPTION


Introduction


Technology adoption lies at the foundation of improving production efficiency in the agrarian reform sector of Honduras. Technological change has rapidly become available to HARC members whose production systems, rooted in ancient Mayan traditions, are still practiced throughout Central America. Slash and burn methods of field preparation developed reflexively over generations as farmers sought to maximize efficiency within the physical and institutional parameters imposed on them. Abundant land resources of Mayan civilizations allowed farmers to abandon fields long enough to become naturally rejuvenated. Fields allowed to lie fallow for several years benefit from the natural growth of vegetation and composting which restore essential nutrients to the soil. Unlike their ancestors, however, modern-day Central American farmers are not afforded the luxury of moving on to another parcel of naturally rejuvenated land. Colonial institutions and population growth have tightened land constraints and forced farmers to cultivate the same land every year. Perennial cultivation tires the soil and gradually lowers yields from traditional production methods.

Worsening soil fertility has provided impetus for farmers to seek and experiment with various technologically advanced production methods. Expanded markets and complementary technologies have provided further impetus. The capacity of farmers to respond to such pressures often corresponds to their ability to purchase complementary human capital, either in themselves


40








41

or through hired labor. Small farmers usually lack the funds to purchase necessary assistance for technology adoption.

Environmental strain has also motivated development institutions to promote environmentally friendly, or "sustainable," technologies. High-input strategies of the "green revolution" depend on non-renewable natural resources. International agricultural development institutions are now mandated to simultaneously pursue environmentally friendly and productivity enhancing technologies (GREAN, 1995). Integrated pest management (IPM) technologies are considered crucial to environmentally friendly components of agricultural systems.

Integrated pest management strategies require the identification of, and the complex interactions among, organisms that occur in the crop system. IPM attempts to introduce biological control agents that are compatible with management schemes and the ecological dynamics of the agricultural system. The ultimate goal is to develop minimally disruptive methods of managing pests that do not reduce the value of agricultural output. Those technologies are being developed, but perhaps more challenging is the effective promotion of technology adoption among farmers in developing countries. This chapter addresses that task.


Determinants of Technology Adoption


Education and extension have long been considered indispensable to technology diffusion and the correction of attendant economic disequilibria (Schultz, 197.i). Lack of skill and knowledge lie at the root of production inefficiencies which have restriLted production systems in developing countries from reaching their technical frontiers.

The level of human capital in Central American farmers in general and Honduran farmers in particular is low. Functional literacy levels in rural Honduras stands at about fifty percent (World Fact Book 1988). Institutional factors also influence the accessibility and appropriateness








42

of technology. Land tenure arrangements of small farms in Central America are tenuous, increasing the risk associated with new technologies. Credit is available on a sporadic basis, reducing producers' capacity to sustain newly adopted technologies. Farm and household size vary in the region as well, playing important roles in the selection of technologies.

Technology transfer in rural Honduras is confronted with some formidable obstacles. Assistance programs to introduce new technologies are often criticized fcr not taking into account a common farmer's opportunity set, production constraints and, perhaps most important of all, his mental preparedness to respond to assistance. If outside technology is introduced to an area that lacks the skills to interpret and screen relevant information, then the "spill in" of the technology will not occur.

In developing countries, much of the social science literature regarding technology lies outside economics, focusing on normative notions related to political and social aspects of production (Roy and Clark 1994). Birkhaeuser, Evenson and Feder (1991) review several studies related to the determinants technology adoption and efficiency. Jamison and Lau (1982) indicated that education positively influences productivity and technclogy adoption of Thai farmers. Khaldi (1975) and Pudasaini (1983) indicate that education improves allocative efficiency progressively more as the rate of technology adoption increases.

Martin and Taylor (1995) showed that producers of commercial crops and members of cooperatives are more inclined to adopt new technologies than subsistence and independent farmers. Their results also demonstrated that T&V training methods iave a multiplier effect through personal contact of experts and friends, and are very effective in motivating technology adoption in contrast to impersonal multi media techniques.

This chapter examines the adoption of integrated pest management (IPM) techniques developed by the Integrated Pest Management Program (Spanish acroinym MIPH) of the Pan








43

American Agricultural School. MIPH promoted IPM procedures among HARCS in the study area (See Appendix I for data and study area).

IPM technologies merit inquiry on three counts. First, they are specific technologies which can be investigated in depth with a few questions (See Appendix II for the survey instrument). The second refers to the word "integrated" in the term "integrated pest management." IPM technologies integrate several other production technologies ranging from soil preparation to the point of harvest. Other technologies such as fertilizer application or row spacing may be components of IPM, but they do not depend on other technologies to the same degree as IPM. Thus IPM technologies serve as a proxy for the adoption of comprehensive systems. Finally, IPM is pertinent to cooperative organizations because IPM techniques often work more effectively at the level of common property (Rook and Carlson 1985; Meister 1980).

MIPH randomly assigned four different training types to HAR-s in the sample. One group was set aside with no training to serve as a control group against which extension efforts could be measured. The training types were:

1. Printed material only
2. Lectures only
3. Lectures and printed material
4. Lectures, printed material and electronic visual aids.


Trained agronomists visited the groups on a regular basis to give lectures and/or supply printed information. MIPH focused on common problems faced by basic grain producers and suggested cost-effective means for overcoming them.


Methodology


The effectiveness of technology is ultimately gauged by the increased efficiency or profit of farmers. However, a necessary intermediate step in that improvement is the actual adoption








44

of the technology, the simple yes or no decision of the farmer. Technologies exist that reduce costs and alleviate environmental strain, but often do not spillover into widespread use.

Determining what influences the decision to adopt under various circumstances holds policy relevance. Education and extension budgets are limited. Identifying the characteristics of farmers that make them amenable to new technologies and the means by which those technologies can be taught will permit scarce rural education and extension funds to be invested more effectively. Evaluating adoption rates requires far less data than evaluating actual production responses and may be conducted at early stages of extension and education programs to monitor the relative effectiveness of promotion techniques.

A common procedure in technology adoption studies is to examine dichotomous adoption decisions as a function of factors related to the farm production system or the farmer's background. Most studies enumerate site visits to measure the impact of extension services (Birkhaeuser, Evenson and Feder, 1991). A few have attempted to directly identify the source of the farmer's knowledge regarding a technology (J.K.Harper, et al. 1990; Martin and Taylor 1995). This study analyzes IPM adoption rates of HARCs as a function of human capital factors and different types of extension programs in which HARCs had agreed -o participate. The link between extension and farmer contact is measured directly by the group extension method and thus need not be asked.

Analyzing the decision to adopt or not adopt a given technology .n a regression equation requires the specification of a binary dependent variable. Probit or logit models (Maddala 1983, Takeshi 1981). have been used extensively to investigate factors th.t influence technology adoption. The basic linear form of these models is given by:








45


(3.1) 9i = 3'xi + C

where = 1 if adopt


=0 otherwise
It follows that
(3.2) Prob(9i = 1) = Prob (ei > -3'xi)
= I F(-O'xi)

where F is the cumulative distribution function for e.

The estimation procedure of the logistic function is maximum-likelihood because OLS yields biased estimates (Burrows 1983; Domenich and McFadden 1975). Several studies have employed logit regressions to examine technology adoption in agriculture in the United States (Schaible and Whittlesey 1991, Harper et-al. 1990; and Zepeda 1990) and a few in developing countries (Martin and Taylor 1995: Palanigounder 1989; Jamison and Lau, 1982).

Two regressions are run. One for adoption prior to pest infestation, and one for adoption after infestation. Farmers were never asked directly whether or not they adopted a recommended technology. Rather, enumerators gave farmers open ended questions regarding pest management practices. A response was scored as a correct adoption if it corresponded to recommendations made in the extension program. Adoption practices were queried regarding six prevalent pests.


Data and Model Specification


Farm size, human capital, labor availability and land tenure are considered important characteristic variables that explain adoption rates (Feder et al. 1985). The following logit model includes three sets of variables which influence farmers' decisions to adopt new technologies. The first set (variables associated with parameters 01 to 0,) relates to -he propensity to adopt technologies, the second set (variables on parameters 0, to /..) represents experience with








46

advanced inputs, and the third set (variables on parameters 01 to 0,,,) relates to the mode of technology promotion employed by MIPH:


(3.3)Adopt = 0, + 01LT30 + 0,GT50 + 0,Primary + 0,Class + 0,Sold + 0,Household +
0,Paraiso + 0,Seed + flHerbicide + 31dnsecticide + (%Fertilzer + 012Lecture
+ fPublication + fi1Lecturepub + 0 Lectureaid + 3,16Prevent
Where:


= 1 if farmer adopted the technology, 0 otherwise. = 1 if farmer's age less than 30, 0 otherwise = I if farmer's age greater than 50, 0 otherwise. = Number of years of primary schooling. = 1 if farmer attended literacy class, 0 otherwise. = Proportion of output sold. = Number of household members. = 1 if producer is from the region of El Parafso, 0 otherwise. = Years of experience with hybrid seeds = Years of experience with herbicide = Years of experience with insecticide = Years of experience with fertilizer = 1 if group received extension lectures without additional teaching aids, 0 otherwise = 1 if group received lectures accompanied by electronic visual aids, 0 otherwise

= 1 if group received both lectures and printed extension publications, 0 otherwise = 1 if group received printed extension publications and no personal lecture, 0 otherwise


Adopt LT30 GT50

Primary Class Sold Household Paratso Seed Herbicide Insecticide Fertilizer Lecture Lectureaid Lecturepub Publication








47


Results


The logit regression is displayed on Table 3. 1. The model appears to "fit" the data well, correctly classified estimates (based on a fifty-fifty classification scheme) amounted to 93.46 percent and the model 2, which tests the overall significance of the model, is significant at the one percent level.

The higher the proportion of output sold (Sold) shows a strong influence to adopt technology. Primary education has a positive influence (Primary), but not as strong as attendance at (literacy) Class. The Class coefficient could indicate more a personality type than a transformation in thinking to adopt new technologies; farmers amenable to becoming literate might be similarly amenable to adopting new production techniques. Class would thus represent a predisposition to new technologies than an actual change in thinking to adopt technology. The positive sign on Household supports the Boserup hypothesis in that as constraints tighten on production systems, more advanced methods are adopted. Farmers over the age of 50 seem less amenable to adopting new techniques than their younger compaileros. The regional variable Paraiso had a positive influence on adoption, a result that could be attributed to better extension services or a more activist campesino union prominent in the region (ANACH).

The only significant result among variables attempting to represent experience is shown on the coefficient of hybrid seed. One of the primary goals of MIPH was to help farmers with a history of pesticides misuse. Pesticides had been aggressively promoted by retailers and some government extension agents. IPM technologies, by virtue of their not relying solely on pesticide use, present additional obstacles to optimal technology








48

diffusion. Integrated pest management requires different modes of technology transfer than traditional pest prevention strategies because commercial pesticide p -omoters benefit from the latter and are often in conflict with the former(Agudelo and Kaimowitz 1991).

In terms of extension methods, both Lecture and Lectureaid registered positive and significant influences on adoption. Both forms of extension provided personal contact, an important if expensive extension method demonstrated influential in Martin and Taylor (1995). Publication, which provided only printed materials is insignificant. However, it is somewhat perplexing in that Lecturpub, which also provided personal extension, is insignificant.








49


Table 3.1 IPM Adoption: Pre-infestation maximum likelihood estimates


Variable


Constant
LT30 GT50
Primary Education
Class Sold
Household
Para[so
Experience with:
Seed
Herbicide Insecticide
Fertilizer
Type of extension:
Lecture
Lectureaid Lecturepub Publication


Wald Statistic


-3.437
-0.320
-0.663 0.089
0.4028 0.542 0.057
0.374


0.127
-0.069
-0.0386
-0.059


0.524 0.632 0.060
-0.322


Significance


0.422 0.274 0.246 0.046 0.1834 0.198
0.034 0.194


0.034 0.041 0.0369 0.037


0.249 0.262 0.282 0.291


-2 Log Likelihood Model Chi-Square Improvement Goodness of Fit


Chi-Square df
1102.132 2432


80.088 80.088 2473.165


15 15
2432


Predicted
0 1
0
0 8 0


Percent Correct 100.00% 0.00%


Overall 93.46%


Coefficient


66.239 1.3632 7.2979 3.6389
4.8266 7. 4896 2 912 3.7263


14.2069 2.8689 1.0907 2.5177


4.4292 5.3276
0.0452 1.2242


Std. Error


0
0.243 0.0069
0.0564 0.028 0.0062 0.0879 0.0536


0.0002 0.0903 0.2963 0.1126


0.0353 0.0158 0.8317 0.2685


Significance


1.000 .000 .000 .000


Observed


Variable Coefficient








50


Summary


Technology adoption is a critical first step for increasing the efficiency of HARC production. Improving the diffusion of advanced technologies is becoming increasingly important in Central America as traditional slash and burn methods of production deplete the agricultural resource base which must serve to feed rising populations.

IPM techniques offer a comprehensive "proxy" to examine technological packages because they incorporate various aspects of the production system. IPM techniques also offer a more precise means for examining the effectiveness of public extension programs because socially optimal remedies of IPM are often at odds with those that are commercially optimal.

This chapter examined IPM technology adoption as a function of human resource factors. Three sets of influences were considered, extension type, experience with inputs, and the demographic factors that predispose producers to opt for new techniques. The model yielded notable insights. The proportion of output sold is positively correlated with the decision to adopt technology. Older farmers appear less likely to adopt, but literacy class and, to a lesser extent primary schooling, are shown to be positively correlated with adoption. Lectures, both alone and accompanied with visual aids positively influenced adoption.














CHAPTER 4
TECHNICAL AND ALLOCATIVE EFFICIENCY: COLLECTIVE VS INDIVIDUAL


One of the most important economic policy issues facing restructuring and developing countries concerns the relative efficiency of collective and individual enterprises. Cooperatives" theoretical potential has been superseded by the observed and dismal le'el of their failure. The term "collective" is rarely heard in the popular and business media without pejorative antecedents: "inefficient," "lethargic," "bureaucratic-heavy," etc.. Many of the economic problems of formerly communist countries and of poorly performing developing countries are attributed to the waste incurred by collective organizations.

Inefficiency obviously exists in cooperatives, but its precise nature is unclear. Furubotn and Pejovich (1970) predicted that cooperatives would degenerate in a capitalist environment as workers/owners have relatively less incentive to make long-run capital investments. However, cooperatives in France, Italy, and parts of the former Yugoslavia have notable success records. And recent studies from Northern Italy (Bartlett, et al. 1992), and the former Yugoslavia (Boyd, 1987; and Piesse, et al., 1996) suggest that cooperatives are more efficient than private enterprises.







'There are several types of cooperatives. Some organize econo:nic agents for specific mutually beneficial activities such as marketing or input purchases. Unhess otherwise noted, the term cooperative in this dissertation refers to full production cooperatives in which resources are communally owned, labor is pooled in production and revenues are shared.

51








52

Conventional wisdom attributes inefficiency of collectives to "shirking," or the abdication of personal work responsibilities. However, it is not clear whether shirking is most detrimental at the level of the worker or at the management level.

Two articles laid the groundwork for the issue of shirking in labor-managed firms. Alchian and Demsetz (1972) concentrate on the worker level, contending that individuals of a collective enterprise lack adequate incentive to monitor coworkers because they do not receive the residual claim awarded to managers of private capitalist firms. Even if one member is appointed the task of monitoring, they argue, the monitor/manager has no authority to hire and fire, and has no incentive to efficiently utilize and maintain fixed capital because the individual portion of capital returns is less than the personal trade-off between labor and leisure. Alternatively, Jensen and Meckling (1979) submit that shirking inefficiencies are most problematic at the level of management. They consider it "naive" to believe that managers of collectives would take the same pains to "seek out high pay-off new projects, to weed out projects which have negative pay-offs, to control waste and shirking, etc." without an additional claim on returns.

The distinction between worker shirking and management shirking is an important one. The alleviation of widespread worker shirking involves substantial monitoring costs and may be an exogenous social characteristic unresolvable by policy modifications. The success of the Israeli Kibbutz, for example, as well as Amish and Hutterite communities in the United States, is often attributed to pre-existing religious bonds that preclude labor shirking. Management inefficiencies, on the other hand, are more easily overcome through restructuring and incentive realignments.

Different sources of shirking generate characteristically different inefficiencies. The selection of inputs is determined through management, but the actual input use is exercised by








53

workers. Suboptimal selection of inputs results in allocative inefficiency. However, any input combination can be used in a technically efficient manner, and is a function primarily of worker responsibility2.

If "net-shirking," shirking so extensive that it over-rides gains achieved by economies of size, occurs at the worker level as posited by Alchian and Demsetz (197'), technical efficiencies will be lower on collective systems than individual systems. Allocatih e inefficiencies may be caused by two factors. Allocative inefficiencies could well be the result of management disincentives, as maintained by Jensen and Meckling (1979). However, allocative inefficiencies could also result from friction in input distribution systems, which are notably inefficient in Honduras.

This chapter compares technical and allocative efficiencies of HARC individual and collective maize production systems. Most research on issues related to :ollective vs. individual production, beginning with Ward's "firm in Illyria" (1957) is theoretical in nature. Empirical studies are rare because data sets comparing collective and individual enterprises are lacking. A recent article by Carter, et al. (1996) examined Honduran agrarian re orm cooperatives. The authors estimated a standard OLS production function regression which included a dummy variable on collectivity which suggested that collective organization had a positive effect on production. The same study showed that 90 percent of surveyed H XRC members did not consider shirking a problem. This chapter is intended to contribute to tie empirical side of the discussion.





'To the extent that workers are not properly trained and motivated, technical inefficiency may also be attributed to management. However, in the case of HARCs, management makes no decisions regarding the training of workers. HARCs exist in large part to provide training to farmers who have demonstrated motivation through political activism.








54

The first section of this chapter reviews the theoretical unde-pinnings and statistical methodologies used to estimate and analyze efficiencies. The second section presents and discusses the statistical results.





Productive Efficiency


The terms productivity and efficiency are often used interchangeably, but subtle differences are noted in the contexts in which they are used and thus in their evolving definitions. Productivity is defined, in its simplest form, as output obtained per unit of input. Productivity measures are usually ratios of output to total or partial input. Such ratios are relative to each other and thus cannot comprehensively differentiate sources of productivity changes. Measures of total factor productivity (TFP), calculated as monetary values or as weighted combinations of physical inputs, fail to distinguish differences in technology as well as the effectiveness with which the technology is implemented.

Efficiency measures, by contrast, are derived by measuring the variation in the input/output relationship to a technical maximum. That maximum is specified by a neo-classical production function and thus represents the maximum output orfrontier I vel of output attainable from a given set of inputs. The distance between the observed input/output relationship and the frontier production function reveals inefficiencies associated with any irput combination.

Frontier measurements of the observed technical maximum offer the most suitable vehicle for accomplishing objective three3 in Chapter One of this study and illuminating policy options. Productivity measures, because they are devoid of theoretical constructs and because they are


3Objective #3 is: Compare the technical and allocative efficiencies of individual production systems vis vis collective production systems.








55

often partial measures, do not offer adequate means to compare efficiencies across groups of producers. A farmer who owns one donkey may appear more efficient than one who owns ten even though the one-donkey owner has made gross overinvestments in machinery and equipment. Frontier productions, on the other hand, simultaneously encompass all inputs in a theoretical framework from which efficiency estimates may be derived.

The inclusion of all relevant factors, including human capital and institutions, conceivably places everyone on the optimal frontier, rendering frontier and "average" functions indistinguishable (MUller. 1974). However, production functions can be reliably estimated in terms of scarce material resources whose impact can be quantified and measured in marginal terms. Human capital and institutions, on the other hand, while important to the production process, do not lend themselves to precise evaluation for two reasons. First, they are a nebulous concepts that cannot be described by cardinal numbers. Years of schooling and courses completed yield exact numbers, but there is no way to calibrate the quantity and quality of knowledge that actually contributes to production. Institutions that govern resource use are important policy alternatives for improving the efficient use of scarce material resources, but are similarly incapable of numeric characterization. Second, there is no rigorous theory to guide analysis of the marginal relationships human capital and institutions hold with production.

Frontier functions are, however, useful tools in determining how human capital investments and organizational alternatives directionally alter the efficient use of scarce resources. Thus inquiry can be made into the human capital attributes and institutional parameters of producers operating close to the frontier. Are they more educated? Does family background inspire greater efficiency? Are they better trained and experienced? What rules govern their actions?








56

Such information is useful to policy makers wanting to identify human resource investments and rules and regulations which have the highest payoff. Knowledge of human capital distinguishes human resource investments that enhance the "value of the ability to deal with disequilibria" (Schultz, 1975) caused by changes in technology. Similarly, organizational designs may be identified that contribute to the efficiency of over-all material resource use.




Technical Efficiency


The most widely used efficiency measures are rooted in the writings of Debreu (1951) and Farrell (1957). The Debreu-Farrell measure of technical efficiency is defined as the equiproportionate reduction of all inputs that produces a demonstrated optimal level of outputs. Conventionally, it is the ratio of observed output to optimal output for a given set of inputs. Thus, unity represents 100 percent efficiency and a fraction less than one indicates a measurable level of inefficiency.

The conventional base within which efficiency is evaluated is the production technology. Technology establishes the limits at which inputs are capable of producin; given levels of output. Let inputs x and outputs q be represented respectively by: x = (x,, x2, x3,...x.) E R'

q = (q1, q2, q3,...q) E R".

The input requirement set of the production technology is (4.1) R(q) = {x: (q, x) is attainable),

which represents all combinations of inputs, efficient and inefficient capable of producing qm. R(q) is shown in Figure 4.1 as the shaded area.







57









x
2






R (q>









0 x
1

Figure 4.1 Input requirement set








58

The boundary set of inputs in R(q) shows the least amount of inputs necessary to produce q. with the current level of technology. Along the boundary, reduction in any input would reduce output or require additional use of a substitute input in x. The boundary set is defined by the isoquant (4.2) 4(q) = {x: x E R(q) and Xx f R(q) if 0 X < 1},

which excludes infeasible input combinations in R". The concept of efficiency requires 4(q) to be convex, or at least quasi-convex relationship, as efficient input combinations use as little of each x, as possible to produce q.

The Debreu-Farrell measure of technical efficiency can be formally interpreted as (4.3) TE(q, x) = min{O: Ox E R(q)} L.'

Optimum efficiency yields a technical efficiency measure of unity, which is identical to the isoquant

(4.4) D(q) = {x: TE(q, x) = 1}.

If one assumes only one product is produced, the production technology may be described using a production function, and its associated isoquants. A production function is a mathematical form that relates the maximum possible output attainable from given quantities of a set of inputs,

(4.5) q = q(x) = max{q: x G R(q)}.

The Debreu-Farrell technical efficiency measure is expressed by the ratio (4.6) TE(q,x) = q/q(x).

The observed level of output, q, cannot be greater than the maximum level of output q(x), delineated by the input-output combinations of the most efficient produce-s. Clearly, then, TE(q,




'The Debreu-Farrell technical efficiency measure is the inverse of the distance function (Shephard 1953, 1970) where
D(q, x,) =max{O: (x/0) E R(q) > 1}.








59

x) 1 because the production function is a technological frontier that can be achieved, but not exceeded, by technically efficient producers. Allocative Efficiency


Technical efficiency is derived solely from the input-output relationship; prices, and thus any notion of cost minimizing behavior, are absent. Allocative efficiency, defined as the optimal combination of inputs to produce a given level of output, is residually obtainable from costminimizing frontier. Given input prices associated with the input vector x (4.7) w = (wI,w,,w3,. ..wJ E R",

and a cost minimizing frontier is represented by (4.8) c(qw:r) = minf{wTx: 1/TE(q,x;r) 5 1}

where r is a vector of parameters representing optimum technology. The ratio of the frontier cost function to the actual cost incurred yields a measure of economic efficiency (Farrell, 1957) (4.9) EE = c(q,w;r)/x*NVT

where x^ is the vector of actual inputs and wT is the price vector. Allocative efficiency is the ratio of economic efficiency to technical efficiency (4.10) AE = EE/TE = wTxE/TE(q,x)).



AE 1 as EE 1 = TE*AE, economic efficiency can only be unity if full technical and allocative efficiency are achieved.

Technical and allocative efficiencies are illustrated in Figure 4.2.







60














x 2

A





B

C E
I00




0 Xl


Figure 4.2 Farrell technical and allocative efficiency








61

The unit isoquant I' represents a frontier production function utilizing inputs x, and x2. A technically efficient producer is represented by B, which lies on the production frontier. A producer operating at point B or any other point along the frontier cannot reduce any one input without either increasing another input or reducing the level of output. Point A, on the other hand, represents an inefficient producer, who utilizes more inputs than B to produce the same level of unit output. The producer at point A can reduce input use without any reduction in the unit level of output represented by 10, The technical efficiency measure is then: (4.11) TE = 11 B L /11 A 11 5.

Many producers, given differences in managerial and work skills, are likely to fall short of the technically efficient frontier. The amount by which a producer lies below the optimal production frontier can be regarded as a measure of inefficiency and may be accounted for by human capital and institutional differences.

The Debreu-Farrell technical efficiency is a "radial" measure in that, if the isoquant is weakly convex, producers located on the portion of the isoquant where the slope is zero or infinite are considered technically efficient. The Debreu-Farrell definition of technical efficiency is thus not as restrictive as that proposed by Koopmans. Koopmans (1951) defined technical efficiency as the state where increasing one output requires a decrease in another output or an increase in at least one input; and the reduction of one input requires an increase in another input or results in reduced output. Clearly, more efficiency is attainable if reducing one input results in no reduction of output, a necessary condition for Koopmans, but rot for DebreuFarrell. However, while slack6 may pose some problems for mathematical programming estimates, it


'In vector notation || X = (Exj)-, Vi.
6" Slack" refers to the range of an isoquant which may have a slope of zero or infinity. In such a range, a producer may be considered technically efficient, even hough he could reduce one input quantity an maintain the same level of output.








62

does not hamper econometric techniques because functional forms (e.g. Cobb-Douglas) preclude slack.

The concept of allocative efficiency, a measure of cost mini nizing performance, is depicted in Figure 4.2 as well. In order to minimize costs, productioA must be set at a level where the ratio of input prices equals the marginal rate of technical substitution, or point E on Figure 4.2. Allocative efficiency is measured by the distance separating the price line and the efficient isoquant:

(4.12) AE= C II/IIBII.

Allocatively efficient producers adjust input mixes where the marginal rate of technical substitution is equal to the price ratio. Producers can be technically efficient but allocatively inefficient. Allocative inefficiency may be a result of management deficiency or it may appear when input markets are distorted. Institutions are thus more important to allocative efficiency than to technical efficiency. Both technical and allocative efficiency approach unity as they achieve optimum or "frontier" efficiency.

Two empirical methodologies have predominated Debreu-Farrell efficiency measurement, non-parametric frontiers and parametric (predominantly econometric' frontiers. They are discussed in the following two sections.


Nonparametric Frontiers


Farrell's approach (1957) is considered non-parametric and deterministic because the convex hull of input-output ratios is constructed by mathematical programming techniques. Mathematical programming techniques obtain the convex hull of the input requirement set that represents the smallest input combinations for a given level of output. Data Envelopment Analysis, a management science and operations research approach originally proposed by








63

Charnes, Cooper and Rhodes (1978, 1981) and recently discussed in Seiford and Thrall (1990), dominate recent nonparametric applications.

Data envelopment analysis (DEA) was originally proposed as a technique for evaluating the efficiency with which a group of operating units transforms inputs into outputs (Charnes, Cooper, & Rhodes, 1978). DEA has been applied primarily to public sector institutions where prices are absent or unreliable. Consequently, most DEA studies involve technical efficiency only, although allocative efficiency can be calculated given sufficient price data.

DEA is formulated as a fractional linear program, continuing along the same math programming path initiated by Farrell. In concept, the procedure maps input-output data to find the subset of most efficient producers who comprise the frontier against which other producers are compared. The most efficient units in the group define a production function that is linear in a "piecewise" fashion. Input-output relationships are linear at each piece of the function between efficient units, although they are not necessarily linear across all efficient units. The function describes a hyperplane "efficient surface" equal in dimensions to the number of inputs and outputs. Input-output relationships for the remaining units are then evaluated relative to this efficient surface. DEA has always been an attractive alternative fir measuring efficiency because it imposes no functional form on the data.

The basic DEA approach is given by the specification of a transformation function T, restricted by constant returns to scale and strong disposability: (4.13) T = {(x,q):XO !x, q ! QO, OER+

where x is an nxl vector of inputs, q is an mxl vector of outputs and k denotes the number of producers. X is the nxk matrix of observed inputs and Q is the mxk matix of observed outputs. The vector 0 serves as a measure of intensity for any activity (xi, qj). Technical efficiency measures for each producer i, are obtained by minimizing the ratio of :ctual output to frontier








64

output subject to observed input/output relationships. They are calculated with mathematical programming techniques:

(4.14) TEi(xi,q) = Min X

st:
XO x
QO;>qi/\


where Xi represents the level of inefficiency of the ith producer.

The program produces a scalar efficiency measure for each unit by selecting weights 0 that maximize the ratio of a linear combination of the unit's outputs to a linear combination of its inputs. This efficiency measure is constrained so that the weights selected must be feasible and cannot result in an efficiency ratio greater than that observed for the most efficient unit in the group.

Later developments have unencumbered Farrell's original programming approach from the assumptions of constant returns to scale (Banker, et.al., 1984) ana strong disposability of inputs and outputs (Fare, et.al., 1985, 1987). Nonparametric frontiers are advantageous in that they only require that the functional form of the technology be nondecreasing and concave. DEA methods are particularly useful in modeling operational processes that do not conform to standard market assumptions, such as in nonprofit firms or regulated industries. Nonparametric methods may also be used to test for cost minimization or profit maximization (Varian, 1984) and have demonstrated the potential for economies of scale that went undetected by econometric estimates (Banker et al., 1986). However, nonparametric efficiency measures are severely limited by the lack of statistical properties. Moreover, they attribute all measuremert error to inefficiency, allowing nothing for random uncontrollable events, and by the same token are very sensitive to outliers.








65


Parametric Frontiers


The predominance of parametric frontiers employ statistical methods to analyze the transformation of inputs into outputs. Prior to the emergence of frontier functions, the conventional model that was estimated took the form (4.15) q = q(x,r) - e

where q(x,T) is the production function, q and x are vectors of outputs and inputs respectively, r represents technology and e is the random error. OLS necessarily assumes the expected value of the disturbance term. e, is zero because it estimates parameters of variables by minimizing the sum of the squared errors. However, neo-classical production theory defines the production function as the maximum output obtainable from a given set of inputs. In the absence of random error, e > 0 because observed levels of output cannot exceed the theoretical maximum.

Aigner and Chu (1968) proposed a "parametric deterministic" procedure to calculate the frontier with mathematical programming techniques. They generalized a function with a onesided error term which required all output levels to lie on or below the calculated frontier. (4.16) q ; q(x; #)

where # represents the frontier "estimates". The resulting linear programming function took the form:








66


n

(4.17) i1
s.t.
[q(q,; )-qJ : 0 i=1,..,n
A 0.








The quadratic formula is the square of the bracketed term of the objective function. The measure of inefficiency is given by the ratio of the actual level of output to the frontier level of output.

Although the programming approach yields parameters, it precludes the testing of statistical confidence in that no standard errors can be computed. The estimation procedure also imposes a structure on the technology and is sensitive to outliers.

Schmidt (1976) showed that OLS estimation of Aigner and Chu's formulation yields best linear unbiased estimates of the slope coefficients, but not of the intercept. He also showed that Aigner and Chu's linear programming estimates are maximum likelihood if the error is nonnegative and has an exponential distribution


(4.18) fle) = 1exp(-e/



where p is the distributional mean and p2 is the variance. A half-normal distribution of the (non-negative) error yields maximum likelihood estimates for the quadratic programming technique (Schmidt, 1976):








67


(4.19) Ae)= 2 exp(-e2/2a)





e 0


However, Schmidt acknowledged that both maximum likelihood estimators violate the regularity conditions in that the range of the observed dependent variable is dependent on the parameters being estimated.

A major advantage of estimating frontiers with econometric techniques is that they render statistical properties to efficiency measurements. Parametric frontie- estimations impose a structural form on the technology, but unlike nonparametric methods they provide statistical measures, account for uncontrollable shocks and are less vulnerable to outliers. Early parametric techniques (Schmidt, 1976; see also Forsund et al., 1980) yielded "full 'rontier" measures that, like nonparametric measures, attributed all deviations from the frontier to inefficiency (e 0).

Attributing all deviation from the frontier is unrealistic in that it does not account for random uncontrollable events beyond the purview of management. The "stochastic frontier" (Aigner Lovell and Schmidt, 1977; Battese and Corra, 1977; and Meeusen and van den Broeck, 1977) allows for random deviation from the frontier owed to measurement error or events beyond the control of the producer. The error term of the production function (4.5) in the stochastic frontier is comprised of two components:

(4.20) e = (v - u)

where v has a symmetric distribution which captures random effects and exogenous shocks across firms; and the one-sided error, u > 0, captures technical efficiency of a firm relative to the stochastic frontier. Thus the estimated frontier accounts for stochastic characteristics that are








68

likely to affect any production system, isolating systematic effects in the measurement of technical inefficiency.

If u is assumed to have a half-normal distribution the associated log-likelihood function is


(4.21) Sf(a,P, a,,X) = -Nlnc; - x+ [1nT(-eX
S a 2 o




where X = au/u, a- = oa2/u, and 4, is the cumulative distribution function of the standard normal distribution. As or - oo, X - 0, systematic inefficiency increases relative to random inefficiency.

Assuming u is exponentially distributed yields the likelihood function


(a, P,,(P) = N[1n(+ ((02]
(4.22) +- 2








which is parametized in o and a,.

Although Aigner et al. (1977) characterized the variances of u and v within the residual e, they were not able to break the residual into its two components for each observation. Efficiency scores were calculated as averages for the entire sample where a. = V'(2/7r). Decomposition of the variances for each observation, a distinguishing attribute of mathematical programming techniques, remained beyond the scope of statistically generated frontiers until Jondrow, Lovell, Materov and Schmidt (1982) derived the conditional distribution (ui I e). By








69

specifying a functional form for the distribution of u given the composed error term e, Jondrow et al. demonstrated that point estimates of efficiency are obtainable for each observation. The expressions for the expectation of u given e of the half-normal and exponential models are


0), $(ejA/;) ell
(4.23) E[u,|Iel]= [ --] and
(1 +X2) '(-e,)4o) a





a,*[(.ej - V G /O,]
(4.24) E[u,e] = (e, - p ab + - respecziveLy.
T[(e, -< 4P /a





These indirect measures of u are unbiased. However, they are not consistent because with a mean truncated at zero, the variance of the coefficients can never be zeto.

Stochastic frontiers have been used in LDC's to measure the effectiveness of credit programs (Ekanayake, 1987; and Taylor, Drummond and Gomez, 1986). Several studies examined extension programs (Kalirajin and Shand, 1985; Kalirajan, 1984; Kalirajan and Finn, 1983; and Bravo-Ureta and Evenson, 1994) and education (Kalirajan. 1990; and Pinheiro, 1992). The stochastic frontier has also been used to identify firm and managerial characteristics that influence efficiency (Seale, 1990).


Technical Efficiency vis A vis Technology Adoption


A philosophical distinction needs to be made regarding technology in the empirical examination of technology adoption and technology in a production frontier. The previous chapter examined technology adoption and the factors that influence it. Certain IPM technologies







70

were identified as superior based on engineering designs and experiments. Technical efficiency is based theoretically on the same concept regarding input-output relationships. However, technical efficiency is examined based on the observed best practice among farmers in the sample. It is possible then that farmers appear to achieve optimum efficiency even though they did not employ the optimum possible technology because they use fewer of the inputs of suboptimal technologies. Graphically, the actual isoquant may lie between the best practice isoquant IP and the origin in Figure 4.2.

Most technical efficiency studies recognize that it is virtually imossible to construct the technically efficient frontier from engineering knowledge. Even simple agricultural production systems are too complex. It was that insurmountable challenge that lead Michael Farrell (1957) to propose the construction of the isoquant based on the observations of best practice producers. The only assumption necessary in input-input space is convexity, which is virtually synonymous with efficiency.

Technology adoption studies can assume recommended technologies are optimal because they focus on specific, comprehensible aspects of production. To the extent to which technologies are developed and recommended based on a broad assessment of all inputs available to producers, and the most efficient producers adopt those technologies, the observed frontier and the actual or potential frontier are identical. The MIPH program endeavored to design programs that both incorporate farmers' available inputs and to communicate them to farmers. The IPM technologies developed by MIPH are relatively more comprehensive than most in that they incorporate several production facets related to pest control. However, MIPH technologies cannot represent the entire gamut of possibilities available to farmers.

In sum, the coefficients on technology adoption (3.3) and the coefficients on the frontier production functions in the following sections of this chapter have slightly different interpretations








71

with respect to technology. The former refers to available technology as determined by engineering studies related directly to the physical input-output relationship; the latter refers to an observed technological optimum as demonstrated by "best practice" producers.


The HARC Stochastic Frontier


Most problems associated with parametric frontiers concern the data to which the method is applied. Relevant variables concerning human capital and institutional characteristics are rarely available. When relevant variables are available, they often cannot be included in the estimations because they present degrees of freedom problems due to limited samples.

Field surveying for this study was carried out in a manner designed to overcome these problems. Data were gathered on several aspects of HARC operations to expand the explanatory power of the models and sufficient observations were obtained to avoid problems with degrees of freedom. The stochastic frontier function thus appears to be the best suited for evaluating various aspects of HARC efficiency.

Standard criticisms of technical efficiency methods (Milon, 1987) do not apply to HARCs or the environment in which they operate. Contractual relations, with the exception of internal labor contracts which are documented, do not vary across HARCs, unmarketed factor inputs are included in the analysis and the means and ends of the HARCs are arguably homogeneous. Externalities, while important to broader welfare considerations than considered here, would not likely have much of an influence on efficiency measurements because harmful inputs are not used nearly as extensively in HARC basic grain production as they are in developed countries.

Data on collective and individual production systems were used to estimate HARC frontier parameters for maize production. The Cobb-Douglas model is selected as the functional form for its convenient properties. The general form of the Cobb-Douglas is







72


k m n
(4.25) q = A j x, 0 n hi1H O|U
i=1 i=1 i=1







where q is a producer's output. A is a given level of technology which "shifts" the function in response to technological changes. x, represents the set of i = 1...n inputs and the B's are the corresponding input coefficients. The standard production function estimates output q, solely as a function of physical inputs x,. However, Jensen and Meckling (1979) suggested an extended form of the production function which recognized that production did not occur in a physical vacuum. Knowledge hi (human capital) and "organizational forms" oi also influence the level of output by their parameters ai and yj respectively.

The Cobb-Douglas has several convenient properties. If 04034 1 and EA =1 the form is concave. Thus, assuming the firm minimizes costs and factor supply and product demand functions are continuously differentiable on their domains, the input demand and output supply functions are continuously differentiable everywhere on their respective domains, a very useful property for interpreting results. The Cobb-Douglas is homogeneous and thus provides a means for examining returns to scale in that =1 characterizes constant returns to scale, Eoi>1, increasing returns to scale; and EA <1, decreasing returns to scale. However, the elasticity of substitution remains unity for all levels of output, weakening scale observations. Most important for this study, the dual cost function, necessary for the calculation of allocative efficiencies, can be derived directly from the Cobb-Douglas production function.

Although considered restrictive in some instances, the Cobb-Douglas was developed for and has been used extensively in agriculture in both developed and developing countries.








73


Moreover, functional form has been shown to have minimal impact on efficiency estimates (Kopp and Smith, 1980).

The systematic element is seen in the right-hand side of 4.25 in that ui represents onesided efficiency disturbance :


(4.26)


k m n
q = Afl x 'l h 'f o7'u - q *u,
i=1 i=1 i=1


0 u !1


and q* frontier output. The specific model estimated for maize production is:


(4.27) InMaize where: Variable


0. + fl nLand + 02lnLabor + 33lnSeed + 4lnFertilizer + N3lnHerbicide + 36lnLandprep + 0,Collectivity + f3Paratso Region + flLecture + 0,,Publicatio + 0,,Lectureaid + 3 ,Lecturepub + e



Coefficient


Technology (A) Land Labor Seed
Fertilizer Herbicide Landprep Collectivity Paraiso Region


Lecture


Publication


Constant
Land measured in manzanas* Labor measured in work days Seed measured in pounds Fertilizer measured in quintals Herbicide measured in pounds Total cost of land preparation Degree of collective work arrangements= I if producer is from the region of El Parafso, 0 otherwise.
= 1 if group received extension lectures without additional teaching aids, 0 otherwise = 1 if group received printed extension publications and no personal lecture, 0 otherwise







74


Lectureaid = 1 if group received lectures accompanied by electronic
visual aids, 0 otherwise

Lecturepub = I if group received both lectures and printed extension
publications, 0 otherwise

Manzana = 0.705 hectare
-Parcels used completely in the collective mode are scored as one, those planted prior to parcelization are scored as one half, and those for which the only collective activity land preparation am scored as one fourth. Completely individual production is registered as zero.

0,, is the constant term on the technology variable A in 4.24. Different levels of technology "shift" the function. Land, Labor, Seed, Fertilizer, Herbicidt, and Land Preparation, represent continuous variable inputs and thus are in log form.

Human capital and organizational form variables are included in the production function because they may have a direct impact (Battese, Coelli and Colby., 1989). All extension techniques are included in the model as dummy variables. The variable for the control group, which received no extension assistance, is necessarily excluded to avoid a singular matrix. Collectivity, the variable representing organizational form (Jensen and Meckling, 1979), is calibrated according to the point at which collective operations are yielded to the individual responsibility of each HARC member. In some HARCs land is prepared collectively, but planting and all subsequent cultivation is conducted individually. A few HARCs with large individual parcels do not partition land until planting and initial applications of fertilizer are completed. The Collectivity variable is included in the model for maize to provide additional means of examining the extent to which collectivization at various degrees influences production.

Production functions are estimated with the maximum likelihood technique (equations 4.20 and 4.21). The half-normal and exponential distributions are assumed for the one-sided error. The distinctive effect each distribution has on the frontier is not well known (Bauer, 1990), but Greene (1990) suggests that there is not much difference between the two. Maximum







75

likelihood estimates and technical and allocative efficiencies were estimated using the LIMDEP software program. Technical efficiencies from 4.24, are:


(4.26) TE = = c'.





Allocative efficiency (4.10) is calculated by analytically deriving the dual cost frontier (Kopp and Diewert (1982) and Bravo-Ureta and Rieger (1991)) to obtain a measure of economic efficiency (4.9). It is not necessary to estimate the cost function as a regression equation. Rather, the coefficients from the production function are incorporated into the cost function which is a dual representation of the production function.

Total cost is a function of input prices and output, the minimization of the costs of producing given levels of output. The cost function reveals all the economically relevant information about the technology (Shephard, 1970; Cornes, 1992). Varying price vectors reveal, through the assumptions of cost minimization and convexity, ranges of input mixes that are observed technically efficient. The dual cost frontier is: (4.27) c(qw) = kwaqIX.




The cost function parameters a and m, are analytically derived directly from the estimated parameters of the production function (4.24) where: (4.28) 1=mp, m=([ 13)- k=[ IIPD]~"
ai~m~i. k=m H







76


Results


Information on the study area and data are in Appendix 1. The human capital survey is in Appendix I. Ordinary least squares (OLS) estimates are displayed on Table 4.1. Tables 4.2 and 4.3 display maximum likelihood (ML) estimates of the stochastic frontier where half normal and exponential distributions are assumed for the truncated distribution component, (u). The Cobb-Douglas model fits the maize production data well. The R2 of the OLS is 0.88. All but one of the standard physical input variables, herbicide, are significant at the 0.01 probability level for all three regressions.

The value of X = a./o, is 5.1 and significant at the 0.05 level. This suggests that relatively more of the variation in the model is attributed to random, not systematic influences. As X approaches infinity, the average and frontier functions converge

Coefficients of the Cobb-Douglas production function represent output elasticities for each input. The average function indicates that a one percent increase in land, ceteris paribus, would generate a 0.46 percent increase in output. Frontier estimates indicate a similar increase in land would yield a 0.56 increase in output, 0.09 higher than the average. The coefficient on Seed is higher in the average function than in the frontier functions, though the difference is less notable. There appears no appreciable difference between average and frontier functions for the coefficients on Labor, Fertilizer, Herbicide, and Land Preparation.

The variables Collectivity and Para[so Region are positive and significant. The parameters on extension variables are inconsistent with the results of those estimated in the technology adoption model of the previous chapter. Lecture and Publication and Publication only, the only two extension variables that were not significant in the logit regression that measured technology adoption, are both positive and significant in the frontier function.








77


Table 4.1
Average function: Maize
ordinary least squares regression



Variable Coefficient Std. Error t-ratic ProbIt|>x

Constant 2.295 0.231 9.938 0.000
Land 0.464 0.066 7.028 0.000
Labor 0.129 0.039 3.287 0.001
Seed 0.175 0.058 3.010 0.003
Fertilizer 0.035 0.012 2.F39 0.004
Herbicide 0.016 0.007 2.222 0.026
Land Preparation 0.044 0.017 2.533 0.011
Collectivin 0.112 0.016 6.900 0.000
Paraiso Region 0.151 0.041 3.691 0.000
Lecture 0.052 0.060 0.861 0.389
Publication 0.212 0.069 3.055 0.002
Lecture and Visual Aids 0.039 0.058 0.672 0.501
Lecture and Publication 0.171 0.062 2.756 0.006

Observations: 405
R-squared: 0.877 Adjusted R-squared: 0.873
F[ 12, 392]: 233.246
Log-likelihood: -153.314 Restr.(B=0) Log-l- -577.919
Amemiya Pr. Criter.: 0.821 Akaike Info.Crit.: 0.133








78


Table 4.2
Frontier function: Maize
maximum likelihood estimates
half normal distribution




Variable Coefficient Std. Error t-ratio Prob ItI > x

Constant 2.806 0.194 14.493 0.000
Land 0.556 0.052 10.784 0.000
Labor 0.122 0.028 4.317 0.000
Seed 0.116 0.043 2.690 0.007
Fertilizer 0.023 0.009 2.540 0.011
Herbicide 0.011 0.006 1.693 0.090
Land Preparation 0.049 0.013 3.738 0.000
Collectivity 0.085 0.014 5.997 0.000
Paraiso Region 0.099 0.034 2.872 0.004
Lecture 0.020 0.059 0.336 0.737
Publication 0.129 0.062 2.088 0.037
Lectureaid 0.093 0.054 1.717 0.086
Lecture and Publications 0.131 0.059 2.209 0.027
loru 2.856 4.057 0.704 0.481
ou/av 5.098 2.400 2.124 0.034
V02V+U2U 1.031 0.526 1.961 0.050

Log-Likelihood: -117.0200


Variance components: a2(v) = 0.03935


or2(U) = 1.02273








79


Table 4.3
Frontier function: Maize
maximum likelihood estimates
exponential distribution




Variable Coefficient Std. Error t-ratio Probjt| >x

Constant 2.800 0.178 15.747 0.000
Land 0.559 0.049 11.360 0.000
Labor 0.123 0.027 4.557 0.000
Seed 0.115 0.042 2.732 0.006
Fertilizer 0.022 0.009 2.511 0.012
Herbicide 0.010 0.006 1.688 0.091
Land Preparation 0.050 0.013 3.921 0.000
Collectivity 0.086 0.014 6.258 0.000
Paraiso Region 0.095 0.033 2.854 0.004
Lecture 0.015 0.056 0.271 0.787
Publication 0.125 0.060 2.039 0.037
Lecture and Visual Aids 0.097 0.053 1.830 0.067
Lecture and Publication 0.127 0.057 2.230 0.026

(P 3.524 0.280 12.:~75 0.000
dv 0.198 0.017 11.547 0.000

Log-Likelihood: -114.7641


Variance components: U2(v) = 0.03907


OF2(U) = 0.08053








80


It would be imprudent to argue that any of these dummy variable estimates represent precise measurements of the influence each bears on production. However, they do provide valuable directional indicators. Thus the absolute values are less relevant than the sign of the variables or relative extremities.


Technical and Allocative Efficiencies


Technical and allocative efficiency measures are presented in Tables 4.4 and 4.5. Collective efficiencies are the calculated efficiency measures for the collective parcel of each HARC, and individual parcel efficiencies are presented as averages for each HARC. The standard deviations of individual technical efficiency averages are less than a third of the average for all but two cases, suggesting that technical efficiency does not vary substantially within HARCs. The uniformity of efficiencies within HARCs may be explained by the communication provided by cooperatives as they are in part established to facilitate communication across large numbers of farmers. Empirical evidence (Martin and Taylor, 1995) attests to the facilitating role cooperatives play in communication. It was also observed throughout the course of fieldwork that new inputs and new techniques were duplicated by other farmers within HARCs, in some cases reenforcing errors.

The most salient feature regarding the efficiency measures is that individual parcels are no more efficient than collective parcels. In fact, efficiencies based on the half normal (Table 4.4) and exponential (Table 4.5) distributions show that collective parcdls are more technically efficient than individual parcels for 11 of the 16 HARCs that employed both modes of production. Thus one of the greatest theoretical arguments, and easily the most touted political arguments against collective forms of enterprise, shirking, shows no emDirical basis.











Technical and allocative efficiencies
for maize
half normal truncated distribution


Technical


HARC


Ideas en Marcha El Boqueron Empalizada El Benque Los Bienvenidos El Esfuerzo Los Peregrinos Esquilinchuche San Nicolas Los Almendros La Esperanza Santa Cruz Cayo Blanco Zopilotepe Guaymuras


Individual


.80
(.07)'
.41 (.16) .71 (.15)
.71 (.19) .76 (.13) .81 (.06) .78 (.08) .79 (.19) .52 (.55)
.70 (.10) .73
(.14) .77 (.11) .75
(.14)


Efficiency Allocative Efficiency
Unpaid Labor w=0 w=5

Collective Individual Individual C


.37 .42
(.10) (.10)
.78 .08 .09
(.09) (.09)
.87 .30 .32
(.24) (.23)
.18 .22
(.13) (.15)
.91 .14 .20
(.11) (.09)
.33 .33
(.05) (.05)
.77 .31 .35
(.12) (.10)
.30 .37
(.20) (.20)
.50 .14 .17
(.18) (.23)
.79 .15 .24
(.11) (.07)
.79 .17 .24
(.13) (.12)
.28 .37
(.15) (.15)
.84 .10 .12
(.06) (.06)
.79


.91


electivee


Table 4.4


81










Technical Efficiency Allocative Efficiency
Unpaid Labor
w=0 w=5

HARC Individual Collective Individual Individual Collective




La Concepcion .81 .91 .26 .30 .46
(.10) (.11) (.08)
San Juan de Linaca .85 .90 .33 .39 .63
(.06) (.14) (.12)
La Puzunca .78 .24 .28
(.08) (.13) (.09)
Tempiscapa .74 .20 .17
(.12) (.10) (.08)
La Providencia .68 .56 .20 .24 .16
(.21) (.16) (.16)
19 de Abril .70 .74 .16 .18 .26
(.21) (.11) (.10)
El Coyolar .86 .32 .37 .36 .05
(.07) (.07) (.08)
El Plomo .77 .81 .22 .20 .29
(.14) (.12) (.07)
Los Dos Naranjos .82 .87 .31 .31 .43
(.06) (.05) (.06)
Los Venecianos .76 .21 .22
(.11) (.11) (.12)
La Libertad .72 .54 .19 .19 .10
(.15) (.12) 11)
Montahuelas .80 .34


' Standard errors are in parentheses.


82


Table 4.4












Technical and allocative efficiencies
for maize
exponential truncated distribution


Technical Efficiency


Allocative Efficiency Unpaid Labor


w=0


w=5


HARC


Ideas en Marcha El Boqueron Empalizada El Benque Los Bienvenidos El Esfuerzo Los Peregrinos Esquilinchuche San Nicolas Los Almendros La Esperanza Santa Cruz Cayo Blanco Zopilotepe Guaymuras La Concepcion


Individual Collective Individual Individual Collective


.82 (.06)1
.41 (.16) .72 (.15) .72
(.20) .77 (.13) .82 (.05)
.79 (.07) .80 (.19) .52 (.55) .71 (.10)
.74 (.14) .79 (.11) .76
(.14)


.82 (.10)


.79 .87




.91




.78




.51 .80 .79




.84 .80 .91 .92


.46 (.09) .11
(.12) .37
(.24) .23
(.14) .17 (.13)
.41 (.05)
.39 (.13)
.34 (.21) .15
(.20) .20 (.14) .22 (.15)
.34 (.17)
.12 (.08)


.31
(.12)


.52 (.08)
.12 (.12) .40 (.23) .27 (.16) .25
(.11)
.41 (.05)
.45
(.09)
.42 (.21) .18 (.26) .32 (.09) .31
(.12) .45 (.14) .14 (.06)


.31 .38 .51




.46


.18

.48 .32


.31 .31

.46 .51


.37 (.08)


Table 4.5


83








Table 4.5


Technical Efficiency


Ailocative Efficiency


Unpaid Labor


w=0


w=5


Individual Collective Individual Individual Collective


San Juan de Linaca La Puzunca Tempiscapa La Providencia 19 de Abril El Coyolar El Plomo Los Dos Naranjos Los Venecianos La Libertad



Montahuelas


1 Standard errors are in parentheses


HARC


.46


.68


.86 (.06) .80 (.08)
.75
(.12) .69
(.22)
.71
(.21) .87
(.06) .79
(.14)

.83 (.05)
.78
(.10)
.74
(.14)


.91







.56 .76

.32 .82 .88




.55



.81


.38 (.16)
.30 (.13) .25
(.10)
.24
(.18)
.20 (.13)
.44 (.07) .27
(.13) .38 (.05)
.28 (.13)
.24 (.14)


(.13) .35
(.08)
.22
(.08) .29 (.17) .23 (.11)
.43 (.08)
.24 (.07) .38 (.05)
.27 (.13)
.24 (.12)


.23 .36 .06 .36


.49


.15



.43


84








85


Two of the five remaining HARCs show individual parcels are only scarcely higher. Individual technical efficiencies for Los Peregrinos and San Nicholas (the only for which the standard error is higher than the average) exceed their collective measures by one and two points respectively. Differences are much greater in the remaining cases where collectives are more technically efficient; at least nine points for seven HARCs, two show five point differences and the remaining two HARCs register differences of four points.

Allocative efficiencies are markedly lower than technical efficiencies for both collective and individual systems, and vary proportionately across HARCs with technical efficiencies. Curiously, allocative efficiencies increase for 16 of the HARC individual production systems, in both the half-normal and exponential distributions, when the standard wage of five lempiras is imputed for free labor.' The average allocative efficiency for the imputed wage is 0.05 greater than the same average when no wage is imputed, a difference that is significant at the .01 level of probability. Higher allocative efficiencies for the case of imputed wages suggests that there is an opportunity cost for labor on individual parcels vis A vis other opportunities available to farmers. If labor had been over-employed on individual parcels, allocative efficiencies would decrease when the standard wage is imputed. It would suggest that the shadow price of labor for HARC households is less than the average wage.

This result is somewhat contrary to that of Nguyen and Martinez (1979), who found that productivity of the ejido sector in Mexico declines when the market wage is imputed for free labor. The difference here is that allocative efficiency is a more precise measure, reflecting


7Since paid and hired labor was aggregated into one input, wage rates were calculated as total cost of labor divided by the amount of both free and hired labor. WLen all labor was free a nominal .01 value, .002% of the standard L5.00 wage rate, was imputed for all free labor as it would be mathematically impossible to calculate allocative efficiency with a zero value for any free input. It can be considered a minor opportunity cost.







86

optimum input mixes as opposed to monetary input-output relationships. It may be that HARC members have competing coop-related responsibilities or opportunities that preclude them from devoting free labor to individual parcels. Work outside the cooperative is limited in rural Honduras.

Averages of technical and allocative efficiency differences between collective and individual systems for all HARCs and only those with mixed systems are displayed on Table 4.6.

Table 4.6 Differences' in efficiency averages between collective and individual parcels



Mixed2 HARCs All HARCs

Half Exp Half Exp

Technical Efficiency
Difference 0.00 -0.00 -0.02 -0.02
(0.01), (0.06) (0.55) (0.66)
Allocative Efficiency
wage = 0
Difference 0.09 0.08 0.06 0.05
(1.82*) (1.53) (1.32* (1.03)

wage = L5.00
Difference 0.04 0.03 0.01 -0.00
(0.86) (0.59) (0.19) (0.05)

Collective average less individual average
2 Efficiency scores exclusively for HARCs that had both individual and collective or "mixed" production systems.
3 t statistics are in parentheses
Significant at the level of .1 probability level.
Significant at the level of .05 probability level.








87

There is, on average, no statistically significant difference in technical efficiency between collective and individual parcels. The only significant difference in mean efficiency scores occurs in allocative efficiency when free labor receives no direct remuneration. Collective parcels in this instance are more allocatively efficient. The significance is attributed to the low level of allocative efficiency when free labor is not imputed a wage.


Summary


The most significant result of this chapter is that collective parcels appear more technically and allocatively efficient than individual parcels. The chapter begins by reviewing the theoretical and empirical developments of the Debreu-Farrell efficiency measures. A frontier production function is estimated, rather than a conventional production function, because the former reveals the observed maximum that provides a benchmark against which efficiency can be measured. Three functions are estimated, an average OLS function and two maximum likelihood functions, one with a half-normal distribution and another with an exponential distribution. The estimated frontier function has a two component arror that differentiates random (stochastic) error from systematic (inefficiency) error.

The explanatory value of all three models is high. All but one uf the coefficients of the average function are similar to those of the frontier functions. The land coefficient is higher in the frontier function than in the average function, suggesting that the most efficient producers obtain a higher return from land than most producers. Collectivization ;ontributes positively to production, as the collectivity variable is positive and significant.

Technical efficiency is calculated based on the most efficient input-output relationship revealed by the frontier. Allocative efficiency, a measure of optimum resource allocation, is calculated based on input prices vis A vis the frontier. The results challenge conventional wisdom








88

in that collective producers appear to be more technically and allocatively efficient than individual producers. More important, the results indicate more precisely the nature of HARC failure. The main problem appears attributable more to managerial problems (Jensen and Meckling 1979) than to the popular notion of "shirking" (Alchian and Demsetz, 1972).

Finally, individual producers are more allocatively efficient when a standard wage is imputed for free labor, indicating that there is an opportunity cost for labor devoted to individual parcels. Collective allocative efficiency is also significantly higher than allocative efficiency of individual producers when labor is imputed a standard wage - a significance that disappears when the wage is not imputed.




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E362DQH5T_YWMYAP INGEST_TIME 2015-04-13T20:57:39Z PACKAGE AA00029921_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

AGRARIAN REFORM COOPERATIVES IN HONDURAS By MICHAEL J. MARTIN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1996

PAGE 2

Copyright 1996 by Michael J. Martin

PAGE 3

This dissertation is dedicated to my Honduran friends who were patient and kind enough to share their rich lives and culture with me. It is also dedicated to my parents, whose unquestionable love and concern turned a Central American sojourn into a larger home.

PAGE 4

ACKNOWLEDGMENTS This dissertation is the product of the collaboration and assistance among many friends in Honduras and the University of Florida. I am especially grateful to my Chair, Professor Timothy Taylor, who generously contributed his guidance and support throughout the entire research process. The door to his abundant professional faculties was always open. I am also very grateful to my Cochair, Professor Uma Leie, whose amiably shared insights in the realm of international development have been invaluable in broadening my perspectives. Professor Chris Andrew, in whose class I presented the embryonic stages of this dissertation and who helped in the conceptualization and identification of the research problem, deserves much gratitude, as does Professor James Scale for sharing his rich technical skills and philosophizing on techniques and context. I extend a special appreciation to University of Florida President John V. Lombardi, who took time out of his busy schedule to serve on the committee. My sincere thanks go to Professor Keith Andrews, Director of the Pan-American School of Agriculture in Honduras, for providing fieldwork support and guidance. All the extension agents and staff at Zamorano have my gratitude for their assistance and friendship. I am indebted to the enumerators for turning tedium into fun and for relating so successfully with farmers. The farmers who participated in the survey merit particular appreciation for their time and trust. I am thankful for my brothers, Jim, Dan and Steve who provided invaluable encouragement that reminded me of the benefits of cooperation. I am also indebted to friends here in Gainesville who provided moral support when it was crucially needed. iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS iv LIST OF TABLES viii LIST OF FIGURES x ABSTRACT xi CHAPTERS 1 INTRODUCTION 1 Objectives 4 The Arrangement of the Dissertation 5 2 EVOLUTION OF FORCES THAT INSTITUTED HONDURAN AGRARIAN REFORM COOPERATIVES 7 The Pre-Colonial Period 8 Agriculture 8 Land Tenure and Labor 9 The Colonial Period 10 Agriculture 10 Land Tenure and Labor 11 Independence 14 Honduran Independence 15 Honduran Land Tenure 1821 1898 17 Honduran Land Tenure: 1898 1940 19 Campesino Unions and the Design of HARCs during the Cold War 20 Cuba and Bananas 20 The Land Reform Law of 1962 22 Decretos Leyes (Legal Decrees) Nos. 8 ife 170 25 Agrarian Reform Amid Political Tumult: 1975 1988 28 The Land Titling Program 33 Collectivization 34 Incentives in Collectives 34 Structural Contradictions of Collectives in Agrarian Reforms 36 HARC Collectivization: Underlying Forces 38 V

PAGE 6

f 3 INITIATING EFFICIENCY GAINS: TECHNOLOGY ADOPTION 40 Introduction 40 Determinants of Technology Adoption 41 Methodology 43 Data and Model Specification 45 Results 47 Summary 50 4 TECHNICAL AND ALLOCATIVE EFFICIENCY: COLLECTIVE VS INDIVIDUAL 51 Productive Efficiency 54 Technical Efficiency 56 Allocative Efficiency 59 Nonparametric Frontiers 62 Parametric Frontiers 65 Technical Efficiency vis k vis Technology Adoption 69 The HARC Stochastic Frontier 71 Results 76 Technical and Allocative Efficiencies 80 Summary 87 5 TRADITIONAL AND ADVANCED TECHNOLOGY: A COMPARISON OF BEANS AND MAIZE 89 Beans: A Traditional Crop 90 General Efficiency Comparisons: Maize and Beans 98 Summary 102 6 THE INFLUENCE OF HUMAN AND SOCIAL CAPITAL ON TECHNICAL AND ALLOCATIVE EFFICIENCY 103 Introduction 103 Theoretical Underpinnings of Human Capital 104 Empirical Applications of Human Capital Components 105 Education 106 Training and experience 108 Health 109 Religion 110 Social Capital Ill Empirical Model and Data 113 Maize 114 Beans 123 Training and Education: Complements or Substitutes? 131 Summary 143 vi

PAGE 7

7 THE NATURE OF HONDURAN AGRARIAN REFORM COOPERATIVES AS PRIVATE ENTERPRISES 145 Introduction 146 The problem 147 Cooperative Enterprises and the New Institutional Economics 148 Why Firms Exist 149 Cooperatives in the NIE 150 Contracts 151 Contracts Related to HARC Basic Grain Production 152 The Standard Administrative Chart 153 Transaction Association by Activity 156 Internal Contracts: The Black Box of the Firm 158 Labor-labor contracts 159 Administration-labor contracts 159 External Contracts: Association with Support Agencies 160 Campesino unionizing 161 Land tenure 163 Technical assistance 163 Credit 164 Marketing 165 Vigilance Committees 166 The Design of Efficient Institutions 167 Credible Commitments 168 Non-credible Commitments 168 An Empirical Observation: The Guanchfas Cooperative 169 Summary 171 8 SUMMARY AND RECOMMENDATIONS 175 Design of the Analysis 175 Summary of the Results 176 Recommendations 178 Future Research 179 The Potential Role Private Capital 179 The Potential Role of NGOs 180 APPENDICES I DATA AND STUDY AREA 181 II FIELD SURVEY 188 III REFERENCE LIST 199 IV BIOGRAPHICAL SKETCH 211 vii

PAGE 8

LIST OF TABLES Table page 2.1 Distribution of Farmland by Farm Size in Honduras, 1974 29 3.1 IPM Adoption: Pre-infestation maximum likelihood estimates 49 4. 1 Average function: Maize ordinary least squares regression 77 4.2 Frontier function: Maize maximum likelihood estimates half normal distribution ... 78 4.3 Frontier function: Maize maximum likelihood estimates exponential distribution ... 79 4.4 Technical and allocative efficiencies for maize half normal truncated distribution . . 81 4.5 Technical and allocative efficiencies for maize exponential truncated distribution . . 83 4.6 Differences in efficiency averages between collective and individual parcels 86 5.1 Average function: Beans ordinary least squares regression 93 5.2 Frontier function: Beans maximum likelihood estimates half normal distribution ... 94 5.3 Frontier function: Beans maximum likelihood estimates exponential distribution ... 95 5.4 Technical and allocative efficiencies: Beans half normal and exponential truncated distribution 96 5.5 Differences in allocative efficiency between an imputed standard wage and zero wage for unpaid labor 98 5.6 Average Technical and Allocative Efficiencies for Maize and Beans 99 5.7 Coefficients of variation 101 6.1 Personal and household characteristics maize 117 6.2 Physical capital maize 119 6.3 Social capital maize 120 viii

PAGE 9

6.4 Extension methods maize 121 6.5 Experience maize 122 6.6 Personal and household characteristics beans 125 6.7 Physical capital beans 127 6.8 Social capital beans 128 6.9 Extension beans 129 6.10 Experience methods beans 130 6.11 Lecture only maize 133 6.12 Publication only maize 134 6.13 Lecture and publication maize 135 6.14 Lecture and visual aids maize 136 6.15 Control group maize 137 6.16 Lecture only beans 138 6.17 Publication only beans 139 6.18 Lecture and publication beans 140 6.19 Lecture and visual aids beans 141 6.20 Control group beans 142 ALL Asentamientos by name, membership and land access 184 AL2 Comparison of sample averages with national averages 185 ix

PAGE 10

LIST OF FIGURES Figure page 2.1 Adjudicated land: 1962 85 26 4. 1 Input Requirement Set 57 4.2 Farrell Technical and Allocative Efficiency 60 7.1 Administration chart 155 7.2 Activity chart 157 X

PAGE 11

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy AGRARIAN REFORM COOPERATIVES IN HONDURAS By Michael J. Martin December, 1996 Chairperson: Timothy G. Taylor Major Department: Food and Resource Economics Honduran agrarian reform cooperatives (HARCs) perform poorly as economic enterprises. HARCs' inefficiency is generally attributed to their collective mode of production and to internal mismanagement. In this study, a stochastic frontier production function is estimated with cross-sectional maize production data on over 400 individual producers and 28 HARCs. "Debreu-Farrell" efficiency estimates indicate that collective production is more technically and allocatively efficient than individual production. Frontier estimates on individual bean production data show much higher allocative efficiencies than those of maize. Input requirements for maize are more varied and advanced than those of traditional bean production and require adjustments in input mixes. That farmers are not optimally adjusting input mixes is attributed to ineffective input distribution systems. The most prominent single factor influencing HARC efficiency is commercialization. Group mean differences show that farmers who sell more of their output demonstrate higher technical and allocative efficiencies for both maize and beans. Commercialization is also a xi

PAGE 12

prominent factor in inducing farmers to adopt IPM technologies, as demonstrated by a bivariate logit regression. Technical and allocative efficiencies are compared for a variety of human and social capital group means. Extension is shown to be a substitute for education at low levels of education in improving technical and allocative efficiency. However, at higher levels of education, different types of extension methods, such as publication circulation, appear to complement general education. Education, literacy, health and experience show positive influences on efficiency. The greater efficiency of collective parcels suggests that the internal dynamics of the institutional "black box" of the firm are sound. However, transactions with support agencies are undependable as evidenced by ineffective distribution and marketing systems. Failure of implicit transaction contracts results from disparate incentives between HARC members and agency personnel. Collaboration with the private sector, which shares HARCs' economic aspirations, offers a means to capitalize on the economies inherent in the cooperative structure and efficiencies spurred by commercialization. The few cooperatives which contracted private capital show that HARCs can evolve into cooperatives that produce high quality products and generate investment returns that significantly increase member income and achieve social goals. xii

PAGE 13

CHAPTER 1 INTRODUCTION The Honduran economy deteriorated significantly during the "lost decade" of the 1980s (see Hayes 1989) in spite of large infusions of U.S. financial aid. Per capita GDP, which grew annually at 2.2 percent between 1961 and 1980, fell at an average annual rate of 0.5 percent over the next ten years. Gross domestic investment and the real minimum wage also fell over the same period. The share of external public debt to GNP was 1 13.8 percent in 1991, twice the 1980 level {World Development Report, 1993). Most of those funds were invested in state and parastatal agencies that are in the process of being privatized under pressure from the IMF and the World Bank. In 1991 Honduras' per capita GNP stood at US$ 580 (World De\elopment Report 1993), exceeding only that of Haiti in the western hemisphere. Over half of the country's 5.3 million inhabitants live in rural areas where absolute poverty, the inability to afford enough food to meet minimal nutritional requirements, is more pronounced. About 20 percent of the population is unemployed {Europa Yearbook, 1991), but casual observation suggests a high degree of underemployment as well. Infant mortality stands at 49 per 1,000 births and stunting afflicts 34 percent of all children between the ages of two and five (World Development Report, 1993). At least 12,000 Honduran children die each year due to preventable illnesses, 25 percent of Honduran families suffer from protein deficiency and 62 percent show inadequate caloric intake (Barry and Norsworthy, 1990). Seventy-three percent of the population is literate (World 1

PAGE 14

2 Development Report, 1993), but functional literacy the capacity to read sufficiently well to cope with modern instructions is probably lower. Unlike its Central American neighbors, no powerful oligarchy emerged in Honduras' history as the dominant political and economic force to beget extremes of wealth and poverty. Still, incomes and land ownership are skewed. The richest 20 percent earn 63.5 percent of total national income while the poorest 20 percent earn 2.7 percent {World Development Report, 1993). Four percent of Honduran farms comprise 56 percent of total farmland, while 64 percent of the farms comprise nine percent of the farmland (Barry and Norsworthy, 1990). Such distribution may be acceptable in developed countries, but the distress of resource maldistribution in an agrarian based economy is inversely related to the size of a nation's resource base and directly related to the size of its population. It is interesting to note that only one in five Honduran farms is worked by its owner. The majority of farmers work under sharecropping or tenancy arrangements, or on municipal lands termed ejidos (Kurian, 1987). In an attempt to redress the problem of land maldistribution and mral poverty, Honduran governments have made sporadic attempts at land reform. Most farmers who wanted to obtain land under reforms were obligated to join an asentamiento (precooperative) and work collectively with other land reform participants in order to qualify for credit. By 1984, about 294,422 hectares of land had been allotted to land reform groups. Total membership in the groups, however, declined from 61,176, counted at the time of land apportionment, to 48,129 in 1984 (Instituto Nacional Agrario. 1985). Cooperatives are ostensibly preferred because they permit independent farmers to jointly make investments in production, storage and marketing activities that they would be unable to make individually, and because they provide a sense of security to members who individually can be devastated by the uncertainties of agriculture. Cooperatives also serve as educational

PAGE 15

3 institutions and provide an efficient structure for development agencies to perceive production constraints and rural needs so that responses can be formulated in the form of credit and technical assistance. Cooperatives are thus intended to achieve both economic and social goals. However, Honduran agrarian reform cooperatives (HARCs) often fall into bankruptcy or never achieve a position of financial independence even though ample financial and material support have been provided. In 1982, 57 percent of the loans to ±e reform sector were delinquent (Stringer, 1989). Blame for cooperative failure is commonly placed on members, and accusations are heard of corruption, non-cooperation and apathy. But such problems, or at least the potential for them, are present to some degree in all firms. Given the extreme poverty in Honduras, and the physical and mental frailty that inevitably accompanies it, cooperative failure may result from an underinvestment in human capital and institutional flaws which fail to account for human capital deficiencies. HARCs have a notorious record of inefficiency, owed principally to the high rate of loan defaults. The precise nature of the inefficiency, however, has never been rigorously analyzed. An abundance of conjectural notions attribute inefficiency to ignorance, corruption or culture. Even development plarmers and social scientists are quick to discount cooperatives as infeasible forms of association, in spite of their theoretical advantages. It is possible that cooperatives do capture some forms of efficiency better than, or at least as well as, other forms of organization, but due to persistent bankruptcy they are rejected in whole. Do cooperatives fail because production is inadequate? Are cooperatives allocatively inefficient, investing in inputs that do not generate a justifiable value with respect to their marginal addition to production? Or are they both technically and allocatively efficient, but experience financial collapse because they are poorly managed? Distinguishing services that coops can perform successfully vis ^ vis those which they cannot will suggest specific aspects of

PAGE 16

4 policy measures that hold a higher probability of solidifying sustainable cooperative enterprises. Objectives This dissenation endeavors to examine how HARCs can become more "efficient." The concept of efficiency is multi-faceted, related to physical input-output relationships, human capital investments, internal firm organization, market operations and relations with public and private organizations. The lines between each facet are not always conveniently distinct for the purpose of analysis. Efficiency concerns of HARCs are intuitively characterized in three areas: technology adoption, a requisite for improving technical efficiency; production efficiency in terms of both physical input-output, or "technical," efficiency and price or "allocative" efficiency; and institutional efficiency, or the rules that govern HARC operations. All are examined vis k vis human capital investments upon which HARCs both depend and are mandated to augment. Although HARCs have had a notably poor record in achieving financial viability, little research has been done concerning the possible means of rectifying characteristic problems. Thus the primary objective of this dissertation is to identify efficiency weaknesses in HARC operations. Additionally, this dissertation contributes to the general knowledge of ajricultural development concerning aspects of cooperative organizations that depend on human capital investments and broadens the scope of empirical cooperative and human capital studies. The following objectives are designed to identify aspects of institutional behavior and human capital investments characteristic of Honduran cooperatives so that policies may be drawn to improve economic performance: 1. Review the evolution of forces that institutionalized HARCs.

PAGE 17

5 2. Examine how human capital, extension methods and demographic characteristics influence the adoption of integrated pest management (IPM) techniques. 3. Compare the technical and allocative efficiencies of individual production systems vis ^ vis collective production systems. 4. Examine how human capital, extension methods and demographic characteristics influence technical and allocative efficiency of individual production. 5. Examine the complementarity of education and varying extension methods in improving technical and allocative efficiency. 6. Synthesize the results in a new institutional economics (NIE) rubric in a manner which yields viable policy alternatives. The Arrangement of the Dissertation The chapters below probe three principle underlying facets of HARC operations. Chapter 2 reviews the evolution of forces that instituted HARCs. How human capital influences the adoption of advanced technology, an intermediate step in improving and modernizing agriculture, is examined in Chapter 3. Chapter 4 presents a stochastic frontier model and empirical results of collective and individual production systems. Collectivization has been indirectly mandated within the reform sector as it has traditionally been a requirement for HARCs to receive credit and support services from government agencies and because ownership of individual parcels has been precluded by land reform legislation. Technical and allocative efficiencies are compared for collective and individual parcels. A stochastic frontier estimation of bean production is presented in Chapter 5 and compared with maize estimations presented in Chapter 4. Beans are grown using a more traditional technology than maize. Human and social capital deficiencies are often cited as sources of inefficiency within HARCs. Human and social capital are examined directly in Chapter 6 as they pertain to

PAGE 18

6 allocative and technical efficiency. Chapter 6 further examines the complementarity of education and extension techniques in improving technical and allocative efficiency. Finally, empirical results are synthesized in Chapter 7 and interpreted within the broader discrete institutional framework in which HARCs operate. The gathering of data involved extensive fieldwork, dozens of cooperative meetings and contact with government support agencies. Relating empirical results to general observations of the operating environment is a further means of capitalizing on direct fieldwork. The intention is to identify short-run modifications in HARC governance to account for human capital shortages that generally require long-run investments.

PAGE 19

CHAPTER 2 EVOLUTION OF FORCES THAT INSTITUTED HONDURAN AGRARIAN REFORM COOPERATIVES (HARCS) Agrarian structures in Central America have evolved over the last 500 years in response to political and economic shocks. Unlike North America, where agricultural settlements completely supplanted Indian communities, Hispanic colonization incorporated indigenous peoples to extract scarce labor and foodstuffs. Aspects of current Central American agricultural organization can be traced back to colonial and even pre-coloniaJ eras. Placing the current system within the context of political history reveals institutional transactions (Bromley, 1989) that served to minimize transaction costs (Coase, 1937 and Williamson, 1985) for interests that influenced Honduran agrarian reform cooperatives (HARCS). Institutions that ostensibly serve HARCs wan-ant reevaluation in the post-Cold War world to identify structures that misaligned incentives with enforceable responsibilities. Transaction costs may be reduced by reorganization within the constraints of available technology and capital investments, major focuses of this study. The purpose of this chapter is to review the forces that both motivated and constrained Honduran agrarian reform throughout history. Those forces explain peculiarities in the design of HARCs, relative to the organization of classical economic firms, that provide a context within which to interpret empirical observations. 7

PAGE 20

The Pre-Colonial Period 8 The ancient ethnic conflicts persist in Guatemala and southern Mexico are relatively absent in Honduras. However, a review of pre-colonial institutions offers insight into the Spanish colonial framework that was established to control indigenous populations and the post-colonial tenure systems that emerged. Agriculture A variety of Indian cultures inhabited Honduras when the Spanish initiated their conquest in the sixteenth century. Historical evidence is available for the Mayan society, which stretched from western Honduras north to the Yucatan. Information is scarce regarding Indian societies that resided in the central and eastern regions.' However, a few reliable observations about the pre-colonial Honduran agricultural economy may be obtained from previous studies. Maize and beans, which continue to form the basic staples of Central American diets, were cultivated in all areas of Honduras along with manioc and sweet potato. Except for a few small irrigated systems in the Mayan region, farming in Honduras was rainfed. Remnants of gardens and orchards have been found, especially along the banks of rivers where flood waters brought natural fertilization. Only the mute dog and the turkey were domesticated for subsistence. Tribes in eastern Honduras were more reliant on hunting .ind gathering and were generally smaller and more dispersed than those in western and central Honduras. Human carriers and canoes formed die only modes of transport. Central American Indians used no animal traction and, in spite of impressive Mayan intellectual achievements in 'A discovery of a burial tomb of "glowing skulls" was recently made in the area, but research results are still being processed.

PAGE 21

9 mathematics, calendrics, architecture and astronomy, never developed the wheel. Trade was thus limited primarily to clothing goods such as fabrics and feathers and light agricultural products. Swidden agriculture, a constrained form of which is still practiced in Central America, was the predominant farming method. This is an extensive system of cultivation which generally requires at least two years of fallow for soil rejuvenation. Land was cleared by hand with stone axes and cultivated with hoes and digging sticks (the barreta, which is .;till in use). Typically, farmers abandoned one site after cultivation and slashed and burned the debris from another, the fertility of which had been restored by time and natural vegetation. The length of the rejuvenation cycle varied across regions, but at least twice as much fallow land was required to sustain output. Institutional land constraints have deprived most contemporary Central American farmers of the rejuvenation portion of the cycle, although debris from the previous season's crops are still burned. Rising populations have further tightened those constraints, contributing to acute social and political conflicts. Land Tenure and Labor The Mayan Empire was never ruled by one authority, but their ;enters were larger and more advanced in terms of abstract knowledge, technology and economic diversification than other contemporary indigenous populations in Honduras. The degree ( f centralization and the bases of economic support and activity varied across settlements and time. The most fundamental organization of ancient Mayan society was, as it continues to be, the nuclear family. Morley et al. (1983) argue that as class distinctions widened, the Mayan elites coerced peasant classes into a feudal relationship whereby the peasants were apportioned land to cultivate, the product of which was shared with the nobles. In exchange for their igricultural production,

PAGE 22

10 as well as military service and non-agricultural labor, the peasants may have received protection from outside raiders. The chiefdoms in western and central Honduras were primarily dependent on subsistence agriculture. They were not as diversified as the Maya, but nobles employed commoners or slaves for the goods and services they consumed (Torquemada, 1723). The smaller tribes in eastern Honduras were more egalitarian; the only division of labor was based on sex and age. Men undertook the clearing of lands and hunting and fishing, while agriculture and child rearing was primarily entrusted to women. The Colonial Period Although Honduras was discovered by Columbus by 1502, colonization was not effectively imposed for decades. Other conquests in the new world offered greater reward at less expense, owed principally to Honduras' rough terrain, relative lack of processed precious minerals and lack of broad social structure. The riches of the Aztec and Inca made the opportunity costs of conquering the scattered tribes of Honduras prohibitive to the conquistadores. Nonetheless, colonization eventually came and obliterated aboriginal chiefdoms and tribes in Honduras. Newson (1986) speculates that by the end of colonial rule, over 90 percent of the Indian population had been annihilated by slavery, forced labor and disease. Miscegenation further eroded Indian communities, giving rise to the predominance of Honduras' mestizo population. Agriculture The Spaniards introduced new crops and agricultural methods, but the technology of production and the crops grown on Indian lands changed little. Indigenous foods of maize and

PAGE 23

11 beans became a fundamental part of the colonialists' diet. The Spanish either produced their own conventional grains of wheat and rice, or extracted them as tribute from Indians who had to be supplied with implements and draft animals. Converse to current production patterns, wheat production expanded in colonial Honduras to meet domestic market demands, though rice did not. Rice now constitutes a staple and is produced throughout Central America. Virtually no wheat is grown in Central America, though wheat products are becoming increasingly popular, owed principally to PL-480 shipments. Colonists also successfully introdu;ed bananas and citrus products, commodities that now form the backbone of the Honduran export economy. Agriculture in colonial Honduras was dominated by livestock production. Cattle, horses and mules found a compatible new environment in the disease and predator-free high grasslands of the Honduran savanna. The mule and horse industry flourished to satisfy demands in agricultural production, mining and, most importantly, in transport. Indigo, hides and tabasco were produced for export to Spain, but were minuscule relative to mining operations (Newson, 1986). Land Tenure and Labor The design of colonial institutions was endogenous to the accessibility of indigenous populations and the nature of social structures, particularly with respect to labor hierarchy and land tenure, that confronted the Spanish. Tribal divisions and warfare, which had been accelerated by the Quichi overextending its dominance to other Central American tribes, also provided incentives for tribal leaders to collaborate with the militarily s iperior Spaniards. According to Newson (1986), the Spanish extracted slaves from easily accessible lowland areas in Central America to work in colonial mining operations. In the relatively remote highlands, the existence of a jurisdictional structure provided the conquittadores with a low cost

PAGE 24

12 means to secure a ready labor supply. The Crown granted encomiendai to colonists that in essence conferred property rights over the Indians while ostensibly providing the Indians with protection and instruction in the Catholic faith. In exchange, colonists were allowed to levy tributes in the form of goods or money and, initially, labor, on the Indians. Agriculture initially played a subsidiary role to mining in the colonial economy of Honduras. Colonists turned increasingly to agriculture as urban living expenses spiraled and opportunities for wealth creation declined in urban areas (Macleod, 1983). High food prices also served as an inducement to establish agricultural enterprises and supply urban dwellers (Frank, 1979). Tributes, exacted through encomenderos that often exceeded legal bounds, constituted the most important source of Crown revenue. Rampant violations of colonial duty and the abdication of protection obligations forced the Crown to transfer tribute exaction to local administrators who established labor quotas (repartamientos^) in indigenous villages. By that time, disease and forced slavery had taken a serious toll on Indian populations, and thus on the available labor supply. Colonists then divided up Indian labor for not just public and religious institutions, which had been the original intention of the repartamiento, but also for private individuals. The work was so arduous and the pay so minimal, laborers suffered malnourishment and were unable to work their own land adequately. Poor diets for the Indian workers and their families further weakened their resistance to prevailing epidemics. Encomiendas and repartamientos served as viable tools of enforcement in areas where Indian states and chiefdoms already existed because most Indians were accustomed to paying ^" Encomienda" comes from the Spanish word encomendar, which means "to entrust." ^"Repartamiento" is derived from the Spanish word repartir, which means "to divide" or "to distribute. "

PAGE 25

13 tributes to authorities and because the hierarchical structure provided a low cost means of exercising control over large populations. In areas absent of large social structures where nomadic tribes subsisted on swidden farming, hunting and gathering, the incorporation of Indians into the colonial economy was cost prohibitive. The Crown charged Catholic Missions with the task of training Indians how to function in the culture diat had over-powered them. However, small disparate tribes were only brought under control when their attacks caused costly interference in Spanish commerce. In such instances Indians were routinely exterminated or captured for slavery. Pre-colombian Indian land holdings were technically honored by royal decree. In reality, however, Spaniards either ignored the decree or circumvented it by various means. The very act of proclaiming property rights to indigenous peoples was a defacto expropriation of the land they inhabited. Ultimate ownership rested with the Crown and was exercised on its behalf through judicial appointees for mutual economic benefit. Land was offered to Spaniards as an incentive for colonization and the conquest of the Americas and came to be a valuable source of government revenues necessary to protect royal property. Martinez Peldez (1975) contends that the Crown extended property rights to die Indians as a means of ensuring their presence to pay the agricultural tributes needed to supply mining activities and urban lifestyles. In 1591, the Crown instituted two portentous cedulas (decrees) affecting land tenure. The first, which ostensibly would have benefitted the Indians, ordered all lands illegally usurped to be returned to the Crown. The second cidula, however, granted that illegally owned lands could be rightfully purchased. Colonists, by virtue of their much greater wealth and knowledge of Spanish law relative to the Indians, were able to circumvent the first cidula to benefit from the second. Redistribution of power or influence was confined to Spanish appointees or

PAGE 26

14 entrepreneurial Creoles who accumulated enough capital to invest in lucrative official positions (Brading, 1986). This skewed distribution of land prevails in Honduras. The encomienda system lasted longer in Central America than in any other area of Spanish America (Newson, 1986). The current demographic complexion of Honduras began to take shape as the Indian population was surpassed by the mestizos through slavery, disease and miscegenation. Royal Decrees offered no property rights to mestizos as it had to Indians. Latifundios expanded through the exploitations of mestizos, who were forced to devote portions of their production in exchange for usufruct privileges (Martin6z Pel^ez, 1975). Independence Latin American revolutions in the 19th century were inspired by their North American counterparts and facilitated by Napoleon's incursions into Spain. However, while much of the independence rhetoric was similar, the Spanish aristocracy continuec". to rule. Apart from forswearing formal titles that eliminated colonial taxes, little else changed for the majority of the people. Central America was polarized between two forces that would continue to squabble, at great expense to economic development, into the twentieth century. The conservatives arose from Spanish aristocracy and sought to maintain colonial institutions and social classes without the burden of royal tribute. The liberals, comprised mostly of elite bureaucratic professionals that formerly served the aristocracy, saw economic benefit in reforms that expanded economic opportunity. A major source of contention between liberals and conservative-, concerned the Church as a formal component of government. The Church conferred moral justification on die privileges enjoyed by elite conservatives and was instrumental in quelling the masses. Perhaps

PAGE 27

15 most destructive to long-run stability and development was the disp;ite on education. The conservatives wanted to maintain an elitist system of education under the auspices of the Church, while the liberals promoted secular, and ultimately mass, education. Like the Crown, conservatives claimed and exercised the right to make ecclesiastical appointments (Barnadas, 1986). Liberals sought to reduce trade barriers and eliminate monopoly rights that conservatives held by virtue of their royal commission. Liberals also wanted to ban the nepotism of conservatives in governmental, commercial and ecclesiastical positions. Both parties were comprised of elite, urban, intellectual interest", and encompassed only a small portion of the total population. Notions of broad-based utilitarianism, promoted most notably by Jeremy Bentham, provided a philosophical rallying force to revolt against Imperial Spain. However, the exact terms of "utilitarian" economic design did not result from such popular consensus. Honduran Independence Honduras experienced three independences in the frenzied rebellions that swept the Spanish Americas in the nineteenth century. It was first liberated from Spain as part of what was to be the Kingdom of Mexico. Anarchy in Mexico afforded the Audencia de Guatemala an opportunity to institute its own autonomous government, Provincias Unidades del Centra de America. With the exception of Chiapas, all the states of the Audencia revolted against Mexico and joined the Provincias. Central American unity, however, was hampered by the surge of hemispheric independence that permeated all levels of society. Motivations for local participation and autonomy were born of nascent ideals and economic opportunity. Powerful regional caudillos

PAGE 28

16 (political/military strongmen) in each state provided conservatives the foixes needed to defeat the liberals, led by Francisco Morazan, but they ultimately dissolved Central America into five independent states. Independence for Honduras was thus a by-product of war strategy rather than a popular consensus based on cultural heritage or economic principles. Elections notwithstanding, caudillos ruled Honduras from 1839 until the mid twentieth century. The private acquisition of public and Church lands, a hallmark of liberal economic programs intended to stimulate exports and growth, aroused a symbiotic alliance between colonial elites and the Church on the one hand, and Ladino and Indian communities on the other. Conservative forces respected Church property rights and supported preservation of ejido lands for subsistence farmers, who constituted the majority of the population. Economic growth slowed under conservative rule, the country returned to traditions that reflected traditional HispanicCatholic principles, and subsistence farming gained in prominence vis k vis export agriculture. Apprehensions about capitalism also arose from the lack of government capacity to administer a broad-based development agenda. Liberals attempted to op-in alarming proportions of property to foreign investment. A debilitating shadow over liberal economics was cast by William Walker, a U.S. citizen who, in the name of democracy and economic freedom, proclaimed himself president of Nicaragua. His group of predominantly Mississippi Valley mercenaries recruited after the Mexican-American War expected to receive land in return for their services and were extremely unpopular in Honduras, which had proclaimed war on the United States to defend Mexico. They represented the detrimental excess of economic liberalization: foreign usurpation of domestic resources and political control. Walker, who was executed in Honduras after a failed filibuster, severely discredited Liberal policies throughout Central America and intensified anti-American sentiments that prevail tc this day.

PAGE 29

17 Honduran Land Tenure 1821 1898 The land tenure patterns that evolved in post-independence Latin America differed substantially from those in North America. The United States and Canada opened their frontiers to settlers, allowing for a more equitable distribution of wealth and ensuring property rights to people within the political and economic system (ethical considerations regarding the usurpation of Indian lands aside). Property rights were much more tenuous in Latin America, owed to the continuation of ejido lands and absentee landlordship of large holdings. Agricultural technology lagged in Latin America, restricting labor productivity and the consequent labor and food surpluses exigent to industrial growth. The industrial revolution in the mid 19th century that transformed the global economy influenced Honduras' Central American neighbors much more than Honduras. Recorded history and data on the 19th century Central American economy are sparse, but some policies and especially the reactions to them reveal distinguishing features of Honduras. Coffee plantations arose in Guatemala, El Salvador and Costa Rica in response to rising world demand, but did not take hold in Honduras until the twentieth century. Liberal reforms regained prominence among Honduran political philosophers who also contended that, unlike the laissez faire reforms promoted by Francisco Morazan, government had a pivotal role to play in economic development. Lacking a bourgeois class, Honduran economic strategists considered government necessary to stimulate growth and secure integration with the world economy (IHDER, 1980). A new constitution in 1880 mandated the state to establish credit for Honduran enterprises and provide the infrastructure necessary to attract foreign investment. Liberal reform began in 1876 when President ]os6 Maria Medina, under economic liberalization pressures from Guatemala, yielded the chief post to Marco Soto (Yankelvich, 1988).

PAGE 30

18 Soto considered ejido structures inefficient and in 1877 instituted an agrarian law that slated coffee, cacao and Indian rubber as favored exports (Guevara, 1983). The intent of the reform was to propel Honduran agriculture into modern specialized production. Central to the agrarian law was the dictate that municipalities sell any national lands to bidding entrepreneurs at a "just price." The government took an active role in designing and monitoring the agricultural system. Landowners and laborers of "modern farming units" were exempted from military and civil obligations and the tariffs on imported agricultural inputs. The Government established strict guidelines regarding such things as fencing, crops and laborers that constituted modern farming units. Participating farmers were required to make written reports to demonstrate their compliance every six months. Interventions to stimulate the production of export crops failed. Reasons for the failure are not clear. Coffee, which initially was considered the crop of greatest potential, began a longterm decline in world demand soon after the program was mounted (Guevara, 1983). Also, labor was not as inexpensive relative to other Central American countries, partially because some ejido lands remained inalienable (IHDER, 1980; and Quinones and Argueta, 1978). Rugged terrain and poor transportation and communication systems also hindered commerce. Government support for transportation infrastructure was hampered by the enormous debt it incurred from trying to build a transcontinental railroad earlier in the century. In 1892 the total debt was over 73 times annual export volumes. Consequently, foreign investors were reluctant to invest in Honduras. The failure to attract upstanding foreign investors rendered Honduras vulnerable to riskier international investors who promised great returns but did nothing more than further deplete the treasury. The bulk of Honduran agriculture thus remained a decentralized system of traditional subsistence farming into the 20th century (Guevara, 1983).

PAGE 31

19 Honduran Land Tenure 1898 1940 Banana exports began growing along the north coast of Honduras around 1880. Most of the production came from small, independent producers who sold their produce to intermediaries close to the coast or with river transport vessels. Banana production grew in popularity among small commercial farmers. Unlike coffee, the favored export of the government, bananas could be grown with relatively little capital investment and along rivers that enabled transport. Most important, bananas presaged short-run gains. The liberal reforms' hapless attempts to establish a bourgeois class in the 19th century were dealt a decisive blow by the emergence of international banana companies. Twentieth century commodity production, storage and transportation required capital investments beyond the reach of most Hondurans and certainly beyond that of small-scale farmers engaged in banana production. Ultimately, foreigners established large banana plantations that sold directly to exporters, thus undercutting both small Honduran producers and intermediary transporters (Guevera, 1983). Honduran land reforms during 1898-1961 reserved public lands for family parcels. The Honduran Institute of Rural Development (IHDER, 1980) identified two characteristics of land reforms during that period. First, the reforms established inalienable property rights, the intent of which was to impede land concentration and proletarianization of rural labor. The 1924 decree stipulated that family parcels be 20 hectares (Stokes, 1947). Second, because reform parcels were small and used primarily for subsistence crops requiring only about ..00 labor days aimually, land reforms provided an abundant, low wage seasonal labor force. The motivations for the various decrees during this period have been attributed to the government's desire to spur economic growth and generate tax revenue (Villanueva, 1968) and to ameliorate lay-offs in the banana industry (IHDER, 1980).

PAGE 32

20 Most of the pre-Cold War land reform debate was rooted in disputes between campesino unions and large American fruit companies. Land reform initiated in the basic grain sector is distinct in many ways from land reform in the banana sector. Integral to both sectors were the campesino unions that fused political strength among the landless to pressure for land redistribution. Campesino Unions and the Design of HARCs during the Cold War The first campesino unions appeared around the late 1920s. Financed and influenced by the communist party (Posas, 1981), they were characterized by the inflaiDmatory Marxist rhetoric that began fueling the reactionary backlashes to a broad range of grassroots social organizations characteristic of, and crucial to, open democratic societies. Campesino unions were influenced by national and international forces which, particularly throughout the Cold War, had enormous leverage over the design of HARCs. Those forces arose from political exigencies that, unlike the forces that forge autonomous cooperatives, did not necessarily hold economic efficiency or even viability as a guiding principle. Cuba and Bananas Two events in the 1950s had dynamic impacts on Honduran agrarian reform, the Cuban revolution and the lay-off of over 18,000 laborers on banana plantations. Labor organizing was already afoot in Honduras prior to the Cuban revolution. Labor and farm organizations were inspired significantly by Cuba, as were academic institutions that could provide scarce technical and logistic support for mobilization activities. Massive floods and an unprecedented strike in 1954 in which 25,000 workers participated gave United Brands incentive to restructure by closing operations on marginal lands and by making labor-saving capital investments. The company

PAGE 33

21 reduced its labor force by 69 percent between 1954 and 1963 (Posas, 1987). Those workers were forced to return to subsistence production, but they returned with a new form of human and social capital embodied in the capacity to organize for worker rights. Campesino unions modeled their organization on urban and banana company union mentors, whose objectives and optimal strategies were different from landless farmers. Urban and plantation workers tend to be more concentrated than campesinos, allowing for relatively low costs of coordinating activities. It was easier to define and achieve consensus on the goals among urban and plantation union members than among campesinos. Association with and support of internationally polarized "labor" organizations also drew campesinos into conflicts about which they had little understanding and to which, by virtue of their geographical dispersion and illiteracy, their interests were more vulnerable. Two principle campesino unions emerged in Honduras at the inception of the Cold War. The first, the National Federation of Honduran Campesinos (FENACH^ was an offshoot of the unionizing apparatus of the 1954 strike. Although FENACH was spawned by autonomous local organizations largely devoid of revolutionary ideology, two of its three principle leaders were militant communists (Posas, 1987). Within two months of the founding of FENACH in 1962, another, less radical, campesino organization appeared, the National Association of Honduran Campesinos (ANACH). ANACH could claim some lineage to the grassroots 1954 organizations, but it also received support from the American Institute for Free Labor Development (AIFLD), an AFL-CIO affiliate that received funding from the U.S. Agency for International Development (US AID) and from its parent organization, the Inter-American Regional Organization of Workers (ORIT). ANACH was clearly a union established by foreign funding as a moderating alternative to FENACH.

PAGE 34

22 Both ANACH and FENACH applied for personeria juridica, which technically confers legal recognition on individuals or organizations. In practice it is used as an institutional mechanism to support or leave unmolestedthose in the government's favor. ANACH was granted personena juridica two months after solicitation. FENACH, v/hose principle disputes were with the Tela Railroad Company (a subsidiary of United Brands), was never granted personeria juridica (Posas, 1981). One would be hard-pressed to argue, given the collapse of communism, that Marxist intellectuals' abstract notions of an ideal society, and particularly its attainment, were in sync with the aspirations of uneducated laborers and farmers. On the other hand, opponents of communism, however defmed, reacted in Honduras and throughout Lstin America on several well documented occasions by means contrary to the democratic ideals which they espoused. Cold-War reactions were also evidenced more subtly in development assistance programs. Honduran agrarian reform cooperatives became tools of partisan international concerns which relegated the long-term benefit of the HARCs subordinate to the attainment of political goals. The Land Reform Law of 1962 The Cuban revolution prompted anti-communist forces within Latin America and the United States to respond to landless farm workers as a means of quelling violent uprisings in the western hemisphere. The Alliance for Progress, accorded at the Inter /vmerican Conference at Punta del Este in 1961, stipulated that aid recipient countries must mount land reform programs in order to qualify for assistance. Though motivated by political exigencies, land reform also had economic underpinnings. The redistribution of idle lands was consistant with the Kaldor-Hicks compensation principle in terms of wealth in that resources began generating substantive economic gains; a potential transfer

PAGE 35

23 could compensate absentee landowners for the negligible utility derived from owning but not utilizing the land. Redistribution also constituted a stricter Pareto improvement in terms of income in that the former owners derived no income from idle lands. The Alliance induced democratically elected President Ram()n Villeda Morales to establish the National Agrarian Institute (INA) in 1961. INA was mandated to respond to land petitions and coordinate land distribution. It was also charged with providing support services, principally the organization and support of cooperatives. The Villeda government passed the first land reform law in 1962 that slated not only government lands for distribution, but idle private lands as well, thus requiring all lands to serve social purposes. Unused lands were ostensibly subjected to progressive taxation or expropriation, but ambiguously written laws and ineffectual enforcement capacity mitigated the threat to vigilant landowners or land holding companies (IHDER, 1980). Property rights have never been clearly defined in Honduras. To the extent that a real -estate market exists in Honduras, INA and the judiciary serve as its clearing houses. Latifundistas could also legally activate idle lands by establishing cattle ranching operations, a convenient activity for absentee landlords that resulted in a threefold increase in beef production while domestic consumption of beef declined (Monthly Review, 1985). Transferring pastures to crops would apparently be a much more efficient use of the land in terms of income, employment and foreign exchange earnings or savings (Garcia et al. 1988). Growing popularity of structural reforms, legitimized by the Alliance for Progress, pressured Villeda on one side while the landed status quo and banana companies pressured him from the other. The opposing forces destabilized a balance that was vulnerable to interference by a virtually autonomous military. Villeda's attempts to weaken the military through institutional reform enraged commanders who were encouraged by dictators in Nicaragua and the

PAGE 36

24 Dominican Republic to overthrow the government (Morris, 1984). Subsequently, aspirations for land reform were cut short by the coup d'etat in October, 1963, led by Colonel Oswaldo Ldpez Arrellano. Campesino unions' expanding influence had apparently posed a greater threat than the 1962 agrarian reform. The military did not reverse the agrarian reform law, but weakened INA and destroyed FENACH by ransacking its offices and assassinating its leaders. Consequently, the slow rate at which land was being distributed was halved (Bueso, 1987). L(3pez's tenure as dictator and later as a not-so-freely elected president was marred by traditional bureaucratic corruption. Forces arose to effectively challenge the traditional caudillo politics. Labor unions and business organizations recognized mutual benefit in extending government participation to all sectors of the Honduran public. A severe hurricane and the infamous "football war" with El Salvador mobilized Honduran society o work collectively for the first time, propagating the wide-spread recognition of governmental rights. The agrarian reform of Peru had also inspired younger military personnel, many of whom came firom campesino families and empathized with agrarian concerns. In 1971 Colonel L6pez ostensibly relinquished control of the political forum to democratic elections. The campaign of 1971 brought Ramdn Ernesto Ciuz to the presidency for a brief interlude from military rule. Cruz represented the conservative National Party, which owed its strength to its alliance with, and the campaign management of, the military. The Cruz administration prohibited land "invasions", a dangerous and final recourse campesinos had to demonstrate the acuteness of their poverty and thus their resolve. Six campesinos who had invaded land in the department of Olancho were cssassinated and several others wounded and incarcerated in early 1972. Campesino unions contended land seizures were "recuperations" rather than invasions because such land was actually govermnent or ejidal land

PAGE 37

25 illegally expropriated by latifundistas through forged documents or outright bribery. Sensing their alienation from the real political processes, ANACH planned a massive hunger march and other actions that threatened social stability. In December the military restored order by peacefully relieving Cruz of the presidency and reinstalling General Oswaldo L(3pez Arellano as chief of state. L(5pez returned to power cognizant of the need to broaden his political base. At a huge rally shortly before the coup he emphasized that "The Armed Forces are composed of workers and campesinos , the Armed Forces are not enemies of the workers and the campesinos." Ldpez, who arose from campesino stock, proceeded to carry out the most extensive agrarian reform in Honduran history. Decretos Leves (Legal Decrees) Nos. 8 & 170 Within a month after assuming power, Ldpez promulgated temporary agrarian reform legislation. Law Decree Number 8 (popularly referred to as Decreto No. 8). Decreto No. 8 was billed as an emergency two-year measure necessary to respond to campesino needs until longterm legislation could be designed. Unlike previous agrarian laws that were rendered ineffective through myriad loopholes and ambiguities, Decreto No. S bestowed clear, decisive power on INA to grant access to lands that were not fulfilling a social purpose. Consequently, the pace of land adjudication accelerated dramatically (Figure 2.1). The purpose of Decreto No. 8 was to incorporate campesinos into the economy and give willing workers the opportunity to improve their livelihoods.

PAGE 38

Tota I and Arab I e 30, QQD 40,000 20,000 10, 000 / ' * \ / ' ^ \ / ' * \ / / \ \ / \ 1 \ V — 1 ' ^ \ / \ ^ ' ^ \ — 'A y'l / / y ~ " 1 1 \^^^r ' 1 1 ^.....^..^-^ 1 1 1 1 1 1 1962 196'^ 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 YEAR Tota I Arab I e SourceIHDER, 1980. re 2.1 Adjudicated land: 1962 85

PAGE 39

27 The declared "temporary" nature of Decreto No. 8 and the swiftness with which it was implemented reduced the potential for opponents to organize and contributed greatly to its impact. More than twice as much total and arable land was adjudicated under the scheduled two year enforcement of Decreto No. 8 than the previous ten years under the agrarian reform of 1962. Campesino mobilization is considered the pivotal factor that precipitated the reform. 92 percent of the groups had to pressure INA in extraordinary ways to obtain land. Over 23,000, or 15 percent of all rural landless families, (about 140,000 people) were granted land-use rights to 108,496 manzanas (IHDER, 1980). Upon expiration of Decreto No. 8 a new agrarian decree, Decreto No. 170 was enacted. Decreto No. 170 consisted of 180 articles that in practice marked a return to legal sidestepping and thus a deceleration in reform. Decreto No. 170 granted permanence to the temporary landuse permits granted under Decreto No. 8 and maintained the authority to expropriate idle lands. However, Decreto No. 170 also denied legal recourse to campesinos that obtained land by invasion, a principle means of procuring land under lethargic government agencies. Concomitant with the surge of land adjudications was the growth in public credit to serve the increasing number of asentamientos'* in the reform sector. The low rate of loan repayment to public lending institutions has come to be one of the major criticisms of HARCs. However, large farmers have exhibited equivalently high default rates (Lele 1974. IBRD 1975). In 1980, the total amount of non-payment in the nonreform sector was more than three times higher than in the reform sector (IHDER, 1980), though the non-reform sector repaid 63 cents to the Lempira compared to 56 cents for the reform sector. *" Asentamientos are considered "pre-cooperatives" in that they have not been awarded personerfa juridica, although they invariably belong to a campesino union that is legally recognized. Technically, for a group to qualify for status as a coopera.ive, it must be granted a personerfa juridica. Legal recognition is reserved for those organizations that have demonstrated their good citizenship to the government.

PAGE 40

28 Administrative corruption is not uncommon in Honduras. General Ldpez was ousted from power after the Wall Street Journal alleged he had accepted a millicn-doUar bribe that saved United Brands over 75 million dollars in taxes. Loan defaults become more understandable, if not excusable, under such circumstances when looked at from the perspective of a politically and economically weak campesino. Agrarian Reform Amid Political Tumult: 1975 1988 L(3pez was succeeded by General Juan Alberto Melgar Castro in a peaceful coup d'etat in April 1975. Melgar's power abided in anti-reform forces intent on curbing the pace of land distribution. Over one-half of rural agricultural families were still landless after the reforms achieved under the Ldpez regime, 56 percent of the land was still claimed by less than five percent of the families (Table 2.1).

PAGE 41

29 Table 2.1 Distribution of Farmland by Farm Size in Honduras, 1974 1 Oillt 01Z.V \HCXj . / No of Farms % No. of has. % Under 1 jj, III 1 / . J 7^ S49 Z 1 , JtZ. n 8 1 2 1 O /CCA io,o5(J 1 Q Q fir 2-5 1/CA ZD. 6 L\JJ,0\JJ 5 10 28,264 14.5 201,274 7.7 10 50 34,390 17.6 729,361 27.7 50 100 4,433 2.3 301,228 11.5 100 1,000 3,304 1.7 763,673 29.0 1,000 + 169 0.1 395,330 15.0 Total 195,341 100 2,62),859 100 Source: Stringer, 1984.

PAGE 42

30 Land invasions continued with little effective response from the government. Landed interests belonging to the National Federation of Agriculturalists and Stock Raisers of Honduras (FENAGH) began an intimidation campaign against campesino unions and land reform proponents. Tensions were manifested in the summer of 1975 when, again in the Department of Olancho, ten leaders of the National Union of Campesinos (UNC), two priests and two women were murdered at the hands of ranchers and military officers. UNC has endured more violent intimidation than any other campesino union in Honduras. Frustrated with the Melgar government's unwillingness to provide basic protection and its reluctance to enforce provisions pertaining to both idle and ejidal lands, the three national campesino unions forged an alliance. The United Campesino Front (FUNC). Melgar responded to FUNC with his plan "Operativo Reldmpago Juan A. Melgar Castro." The plan resulted in the distribution of 8,722 hectares to 3,160 families. Melgar's reluctance tc either implement legal reforms or decisively reverse them angered campesinos and the "Superior Council of the Armed Forces" (CONSUFFAA) that installed him. Melgar eventually dismissed heads of INA and the Labor Ministry who had been appointed by Ldpez Arellano and were proponents of reform, but he was too late to demonstrate control. CONSUFFAA announced Melgir's resignation in 1978 and installed General Policarpo Paz Garcfa. Paz ruled until 1982, during which time few land requests were processed. Characterization of campesino unions was often based on association, not unlike the McCarthy years in the United States. UNC sprung from rural developm;nt programs sponsored by the Catholic Church and assistance from Christian Democratic Party. After the Second Vatican Council in 1964 the Catholic Church placed new emphases on ministering to the poor, which unsettled traditional power structures throughout Latin America. UNC was a part of the General Central of Workers (CGT), an activist organization backed by Christian socialists in

PAGE 43

31 Latin America. Although founded by 1972, UNC had to wait until 1984 to obtain its personeria jurfdica (Posas, 1987). In another instance, Efrafn Diaz Galeas, a former president of ANACH, riled financial sponsors by soliciting assistance from UNAH faculty, whom the sponsors considered communists. The Executive Committee of ANACH yielded to threats of funding cuts by dismissing Diaz. Dfaz later reappeared as the president of the Federation of Agrarian Reform Cooperatives (FECORAH). a campesino union that was also delayed fojr years in obtaining its personeria jurfdica. ANACH, by contrast, long the largest campesino union in Honduras, had been relatively unperturbed by the military, owed principally to its alliance with AIFLD. Generous operating budgets financed by the AFL-CIO, USAID, and a number of private corporations ranging from IT&T and Mobile Oil to Sterling Drug and Bacardi Co., (SITRAUNAH, 1979) gave ANACH the capacity to bid organizers away from grassroots unions or hire talented organizers otherwise uninterested in union mobilization. ANACH organizers were thus more financially beholden to foreign interests than to the campesinos whom they represented. Unions that arise from autonomous forces are by nature more democratic in that they must respond to the economic needs of their members. Incentives existed in ANACH, on the other hand, to subordinate the economic welfare of the HARCs to political priorities. This is not to deny that the AFL-CIO had sincere ideological interests in preventing unions from being overcome by communist control. However, the principle source of funding for AIFLD was the US government (92 percent in 1969). Several members of the original AIFLD Administrative Board headed companies that had large financial interests in Latin America. The strings attached to ANACH were unmasked in 1978 when Antonio Julfn M^ndez accused the sitting president of ANACH, Reyes Rodriguez Ar^valo, of fraudulent reelections.

PAGE 44

32 M^ndez demonstrated the veracity of his charges by congregating a majority of assembly representatives in opposition to Rodrfguez. A year later at an ANACK national convention in a Lion's Club hall in Siguatepeque, M^ndez again outnumbered Rodrfguez supporters 173 to 109. Upon realizing he was at a political disadvantage, Rodriguez abandoned the convention to reconvene in a military battalion. Paz Garcia had shown his support for Rodrfguez earlier by deporting a missionary priest who had denounced Rodriguez's corruption (Posas, 1987). The military did not recognize the majority of ANACH delegates, led by M6ndez and Camilio Padilla, but ratified the Ar^valo faction. However, the M6ndez and Padilla coalition received a valuable endorsement from the Confederation of Honduran Workers. M^ndez's authority would again be unsuccessfully challenged in 1984 by an operative of General Gustavo Alvarez Martfnez, the army commandeer responsible for the civillian crackdown and several disappearances in the 1980s. The failures and tribulations of campesino unions cannot be solely attributed to external forces. UNC and FECORAH arose from other labor organizations and, like ANACH, were always in the process of splintering themselves. The fragmenting of unions was easier to accomplish in rural-based unions than in urban or plantation unions because rural systems are necessarily less centralized. Regional and local campesino union leaders hold more member allegiance than national leaders. Under Melgar, INA embarked on a policy that divided its resources between export crops and basic grains. The Paz regime slated eight percent of INA's total budget to the basic grain sector while designating almost 65 percent to commercial export crops. Bananas account for about 22 percent of all national value-added and occupies less than five percent of area planted to crops. Basic grains on the other hand, contribute only about 14 percent to value added even though it uses about 65 percent of cropland (Stringer, 1987). Critics charged on social grounds

PAGE 45

33 that land reform had become a colonization project through coerced collectivization and die transferring of campesinos from their homes to areas where land was less constraining. On the other hand, allocating labor resources to a higher level of employment has distinctive economic benefits. One persistent aspect of the postWWII reforms was the attempt to propel the agricultural population into modern agricultural production (IHDER, 1980). Neighboring Central American countries had more advanced production and higher growth rates, but the maldistribution of wealth and income contributed to social upheaval (Bulmer-Thomas, 1989). Honduran reforms endeavored to escape that fate. The Land Titling Program In 1981 Roberto Suazo Cdrdova became Honduras' first democratically elected president in 20 years. Suazo campaigned strongly on an agrarian reform platform. The pace of land redistribution increased 75 percent over the previous four years. Nonetheless, at the end of Suazo's tenure over 125,000 rural families still had no secure access to land (Ruhl, 1989). Rural population growth alone fed the landless population faster than the land reform could reduce it. Suazo had not satisfied his political mandate with respect to land reform. In 1981, at the behest of USAID, the focus of land tenure adjustments switched from outright redistribution of public or private lands to land titling. In order to qualify, a farmer must demonstrate he has worked a specific parcel of land, but that his tenure was not secure or legally binding. The primary intentions of the titling program were to provide workers of the land collateral for credit and long term incentives to make land improvjments. The quality of soil and the extent to which crops are commercialized influences participation in the land titling

PAGE 46

34 program (Seligsan and Nesman, 1989). Those factors have also been s'lown to be important in the adoption of technology (Martin and Taylor, 1995). Structural adjustment measures have encouraged privatization mroughout all sectors of the Honduran economy, including the agrarian reform sector. The land reform law of 1992 opened vast tracts of land for sale by cooperatives in the agrarian reform sector. Collectivization Beginning in 1962 land reforms in Honduras reversed the liberal tradition of distributing specific parcels to individual families, opting for a collective form of enterprise. In theory, collectivization improves efficiency by permitting large "lumpy" investments in production, storage, marketing and input purchases that are beyond the capacit) of small, capital-poor farmers. Cooperatives also theoretically reduce monitoring costs associat"xi with large enterprises and lower transaction costs of providing extension and support services to large numbers of farmers. In practice however, collectives have had a poor economic record. Attrition rates have always been high in HARCs, averaging over 30 percent between 1962 2nd 1985 (Bueso, 1987). Loan default rates were also disproportionately higher for HARCs than for individual farmers. More than half of the National Bank of Agricultural Development's (BANADESA) loans to the reform sector were delinquent as of 1982 (Stringer, 1989). Incentives in Collectives The conditions under which voluntary collectivization might occur are the focus of agricultural production cooperative (APC) theory (Putterman 1986, 1989; Bonin, 1987; Carter, 1987) which contrasts the behavior of collective production based on va^ing work and revenue sharing rules. APC theory customarily attributes cooperative failure to "shirking," the

PAGE 47

35 relinquishment of individual responsibility to other members. The problem is that widespread shirking leaves many responsibilities unfulfilled and negatively impacts cooperative performance. The important question, however, is not if shirking occurs, but whether it overrides gains in efficiency arising from worker collectivization. APC theory does not take into account the role of human capital. Two of the collective production systems often alluded to in discussions of APC theory, the Kibbutz in Israel and the Hutterite communities in the U.S., place a high value on human capital investments. Nor does APC theory or the cooperative literature in general distinguish the types of monitoring costs that different cooperatives may be more successful in reducing. Illiterate members cannot carryout simple bookkeeping procedures and are obviously ill-equipped to oversee complicated contracmal obligations. On the other hand, landless farm laborers are capable of monitoring the unspecialized agricultural labor tasks to which they are accustomed. The shirking alleged by Alchian and Demsetz (1972), would be relatively more difficult to accomplish under such circumstances than the financial rent-seeking cited by Jensen and Mecklirg (1979). To the extent that dignity is derived from labor, shirking may be unlikely to occur. Fieldwork is one of the few endeavors available to HARC members, and one in which pride is evidenced in traditional folklore, dance and song. The exact nature of HARC breakdown caimot be easily attributed to shirking at least at the level of manual labor. In fact, a common charge leveled against campesinos is that they are simple people who "love to work with their machetes in their milpcs (cornfields)," but that they are disinterested about the more cerebral matters of managing technologically advanced enterprises. HARCs' dismal financial history may abide in elements other than the technical inefficiency implied by shirking. Even firms that are one hundred percent technically and

PAGE 48

36 allocatively efficient could be ruined by financial mismanagement especially if institutions do not safeguard interest holders from "opportunism with guile" (Williamson, 1985). Structural Contradictions of Collectives in Agrarian Reforms The justification for land reform sits on tenuous ground when cooperatives are promoted as part of the reform to capture economies of size'. Many reform advocates contend that skewed land distribution results in inefficiency because small farmers are consid^jred to be more efficient than large farmers (See Corner and Kanel 1971; Berry and Cline 1979; Cornia, 1985). On the other hand, development planners impose cooperative structures in an attempt to capture efficiencies that are supposedly absent in large operations. The economic argument for agrarian reform cooperatives are thus fallacious unless large collective enterprises could attain greater efficiency than those that are individually operated. This may be a tenable proposition for land reform coopera.ives in Honduras and elsewhere in Central America. Honduran agrarian structures are still emerging from precapitalist modes of production and are often still characterized by absentee ownership and sharecropping practices that maintain traditional production methods. By the same token, cooperatives have lacked the managerial skills in both production technology and administration needed to exploit the potential for economies of size. Scale economies that require specialization exact transaction costs for monitoring because production processes become less personal (Bardhan, 1989), partially, at least, explaining why sharecropping practices persist (Alchi;m and Demsetz, 1972). But the cooperative naturally reduces monitoring costs because members are owners and bear ^ Economics traditionally treats efficiencies obtained from the simultaneous expansion of all inputs as economies of scale. Economies of size, on the other hand, occurs when long term, often "lumpy" and not necessarily simultaneous, investments lower the average total cost of production. Economies of size is used here because it is more comprehensive and thus more precisely represents efficiencies sought by the process of collectivization.

PAGE 49

37 substantial peer pressure on one another. The desire to collaborate in a productive manner was quite noticeable among the majority of HARC members. Size economies in HARCs are likely to become increasingly attainable as technological advances are adopted because long-run capital investments such as farm machinery, input inventory, transportation and storage equipment reduce per unit costs of output. Such investments provide new opportunities for coop members to exploit comparative labor advantages, underscoring the importance of the division of labor to carry out specialized tasks and a managerial workforce which can dedicate its time to identifying and implementing profitable activities. Many of those opportunities depend on labor talents that can emerge only through concomitant investments in human capital. The combination of monitoring cost reduction gained through collectivity, and the division of labor gained through human capital may thus allow for the capturing of size economies beyond the grasp of large individual farms or poorly skilled cooperatives. Beyond efficiencies in production, collectives reduce costs incurred by service delivery agencies. New technologies inevitably require a learning process and are aggressively promoted by input suppliers who exploit the glamour appeal those technologies hold on traditional farmers. Cooperatives are efficient structures for technical and support agencies to monitor operations and provide agricultural training to large numbers of farmers. Budgets in developing countries are inadequate to employ enough support personnel and extension workers to train and monitor individual farming operations. In cooperatives training is conducted through seminars from professional instructors and, more importantly, transferred through other members (Martin and Taylor, 1995).

PAGE 50

38 HARC Collectivization: Underlying Forces It is difficult to discern precisely why the collective was adopted as the predominant form of operation and production in the Honduran agrarian reform sector. Stringer (1984) attributes collectivization to Roberto Sondoval Corea who was appointed Director of INA in 1968. Cardona (1979) ascribed the origins of collectivization to Jorge St. Siegens, a Romanian economist who was contracted at the recommendation of UNESCO to serve as the Technical Director and Professor of the Autonomous University of Honduras (UNAH). Posas (1987) claims Virgilio Carias, Head of the Economic Research Institute at UNAH, initially promoted the Israeli mode of production in the successful and broadly renowned Guanchfas Cooperative on the North Coast. Although the origin of collective production is unclear, one thing is certain: HARCs have lacked proper incentive structures and enforcement mechanisms of contracts, particularly with government support agencies, have been notoriously weak. Property assurances are also firagile. HARC members hold no individual title to specific parcels. In fact, very few HARCs technically own the land which they work collectively. The motives for imposing collective production, through denial of credit and technical assistance to individual farmers, are seen in almost all the official organizations that support the HARCs. The government appreciates collectivization as HARCs can absorb additional farmers who are more reluctant to risk their lives invading land than to arrange an agreement with HARC members for formal admittance (IHDER, 1980). Collectivization also permits large financial transactions to be made between loan agents and a few campesino union officials, outside the

PAGE 51

39 scrutiny of individual HARC members. "Soft-state" governments such as Honduras are vulnerable to nepotism and rent-seeking agency staff that could detrimentally exploit HARCs. The notion of collectivization was also supported by nongovernmental entities in Honduras for political, not economic, reasons. The campesino unions maintain more political control over collectives than they could over individual farmers who, once owning a land parcel, would be less dependent on the union and more difficult to mobilize. Cold War development planners also found collectives (within a free enterprise rubric) to be a means of mollifying communist supporters in allied countries.

PAGE 52

CHAPTER 3 INITIATING EFFICIENCY GAINS: TECHNOLOGY ADOPTION Introduction Technology adoption lies at the foundation of improving production efficiency in the agrarian reform sector of Honduras. Technological change has rapidly become available to HARC members whose production systems, rooted in ancient Mayan traditions, are still practiced throughout Central America. Slash and burn methods of field preparation developed reflexively over generations as farmers sought to maximize efficiency within the physical and institutional parameters imposed on them. Abundant land resources of Mayan civilizations allowed farmers to abandon fields long enough to become naturally rejuvenated. Fields allowed to lie fallow for several years benefit from the natural growth of vegetation and composting which restore essential nutrients to the soil. Unlike their ancestors, however, modern-day Central American farmers are not afforded the luxury of moving on to another parcel of naturally rejuvenated land. Colonial institutions and population growth have tightened land constraints and forced farmers to cultivate the same land every year. Perennial cultivation tires the soil and gradually lowers yields from traditional production methods. Worsening soil fertility has provided impetus for farmers to seek and experiment with various technologically advanced production methods. Expanded markets and complementary technologies have provided further impetus. The capacity of farmers to respond to such pressures often corresponds to their ability to purchase complementary human capital, either in themselves 40

PAGE 53

41 or through hired labor. Small farmers usually lack the funds to purchast, necessary assistance for technology adoption. Environmental strain has also motivated development institutions to promote environmentally friendly, or "sustainable," technologies. High-input strategies of the "green revolution" depend on non-renewable natural resources. International agricultural development institutions are now mandated to simultaneously pursue environmentally friendly and productivity enhancing technologies (GREAN, 1995). Integrated pest management (IPM) technologies are considered crucial to environmentally friendly components of agricultural systems. Integrated pest management strategies require the identification of, and the complex interactions among, organisms that occur in the crop system. IPM attempts to introduce biological control agents that are compatible with management schemes and the ecological dynamics of the agricultural system. The ultimate goal is to develop minimally disruptive methods of managing pests that do not reduce the value of agricultural output. Those technologies are being developed, but perhaps more challenging is the effective promotion of technology adoption among farmers in developing countries. This chapter addresses that task. Determinants of Technology Adoption Education and extension have long been considered indispensable to technology diffusion and the correction of attendant economic disequilibria (Schultz, 197.t). Lack of skill and knowledge lie at the root of production inefficiencies which have restricted production systems in developing countries from reaching their technical frontiers. The level of human capital in Central American farmers in general and Honduran farmers in particular is low. Functional literacy levels in rural Honduras stands at about fifty percent (World Fact Book 1988). Institutional factors also influence the accessibility and appropriateness

PAGE 54

42 of technology. Land tenure arrangements of small farms in Central America are tenuous, increasing the risk associated with new technologies. Credit is available on a sporadic basis, reducing producers' capacity to sustain newly adopted technologies. Farm and household size vary in the region as well, playing important roles in the selection of technologies. Technology transfer in rural Honduras is confronted with some formidable obstacles. Assistance programs to introduce new technologies are often criticized fcr not taking into account a common farmer's opportunity set, production constraints and, perhapj most important of all, his mental preparedness to respond to assistance. If outside technology is introduced to an area that lacks the skills to interpret and screen relevant information, then the "spill in" of the technology will not occur. In developing countries, much of the social science literature regarding technology lies outside economics, focusing on normative notions related to political and social aspects of production (Roy and Clark 1994). Birkhaeuser, Evenson and Feder (1991) review several studies related to the determinants technology adoption and efficiency. Jamison and Lau (1982) indicated that education positively influences productivity and technology adoption of Thai farmers. Khaldi (1975) and Pudasaini (1983) indicate that education improves allocative efficiency progressively more as the rate of technology adoption increases. Martin and Taylor (1995) showed that producers of commercial crops and members of cooperatives are more inclined to adopt new technologies than subsistence and independent farmers. Their results also demonstrated that T&V training methods lave a multiplier effect through personal contact of experts and friends, and are very effective in motivating technology adoption in contrast to impersonal multi media techniques. This chapter examines the adoption of integrated pest management (IPM) techniques developed by the Integrated Pest Management Program (Spanish acronym MIPH) of the Pan

PAGE 55

43 American Agricultural School. MIPH promoted IPM procedures among HARCS in the study area (See Appendix I for data and study area). IPM technologies merit inquiry on three counts. First, they are specific technologies which can be investigated in depth with a few questions (See Appendix II for the survey instrument). The second refers to the word "integrated" in the term "integrated pest management." IPM technologies integrate several other production technologies ranging from soil preparation to the point of harvest. Other technologies such as fertilizer application or row spacing may be components of IPM, but they do not depend on other technologies to the same degree as IPM. Thus IPM technologies serve as a proxy for the adoption of comprehensive systems. Finally, IPM is pertinent to cooperative organizations because IPM techniques often work more effectively at the level of common property (Rook and Carlson 1985; Meister 1980). MIPH randomly assigned four different training types to HARCs in the sample. One group was set aside with no training to serve as a control group against which extension efforts could be measured. The training types were: 1. Printed material only 2. Lectures only 3. Lectures and printed material 4. Lectures, printed material and electronic visual aids. Trained agronomists visited the groups on a regular basis to give lectures and/or supply printed information. MIPH focused on common problems faced by basic grain producers and suggested cost-effective means for overcoming them. Methodology The effectiveness of technology is ultimately gauged by the increased efficiency or profit of farmers. However, a necessary intermediate step in that improvement is the actual adoption

PAGE 56

44 of the technology, the simple yes or no decision of the farmer. Technologies exist that reduce costs and alleviate environmental strain, but often do not spillover into widespread use. Determining what influences the decision to adopt under various circumstances holds policy relevance. Education and extension budgets are limited. Identifying the characteristics of farmers that make them amenable to new technologies and the means by which those technologies can be taught will permit scarce rural education and extension funds to be invested more effectively. Evaluating adoption rates requires far less data dian evaluating actual production responses and may be conducted at early stages of extension and education programs to monitor the relative effectiveness of promotion techniques. A common procedure in technology adoption studies is to examine dichotomous adoption decisions as a function of factors related to the farm production system or the farmer's background. Most studies enumerate site visits to measure die impact of extension services (Birkhaeuser. Evenson and Feder, 1991). A few have attempted to directly identify the source of the farmer's knowledge regarding a technology (J.K.Harper, et al. 1990; Martin and Taylor 1995). This study analyzes IPM adoption rates of HARCs as a function of human capital factors and different types of extension programs in which HARCs had agreed to participate. The link between extension and farmer contact is measured directly by the group extension method and thus need not be asked. Analyzing the decision to adopt or not adopt a given technology in a regression equation requires the specification of a binary dependent variable. Probit or logit models (Maddala 1983, Takeshi 1981). have been used extensively to investigate factors that influence technology adoption. The basic linear form of these models is given by:

PAGE 57

45 (3.1) % = /3'Xi + e where = 1 if adopt 4 = 0 otherwise It follows that (3.2) Prob{3?i = 1) = Prob (e-, > -jS'xd = 1 F(-/3'Xi) where F is the cumulative distribution function for e. The estimation procedure of the logistic function is maximum-likelihood because OLS yields biased estimates (Burrows 1983; Domenich and McFadden 1975). Several studies have employed logit regressions to examine technology adoption in agriculture in the United States (Schaible and Whittlesey 1991; Harper et-al. 1990; and Zepeda 1990) and a few in developing countries (Martin and Taylor 1995; Palanigounder 1989; Jamison and Lau, 1982). Two regressions are run. One for adoption prior to pest infestation, and one for adoption after infestation. Farmers were never asked directly whether or not they adopted a recommended technology. Rather, enumerators gave farmers open ended questions regarding pest management practices. A response was scored as a correct adoption if it corresponded to recommendations made in the extension program. Adoption practices were queried regarding six prevalent pests. Data and Model Specification Farm size, human capital, labor availability and land tenure are considered important characteristic variables that explain adoption rates (Feder et al. 1985). The following logit model includes three sets of variables which influence farmers' decisions to adopt new technologies. The first set (variables associated with parameters /3, to iS^) relates to the propensity to adopt technologies, the second set (variables on parameters 0j to /Sio) represents experience with

PAGE 58

46 advanced inputs, and the third set (variables on parameters 0u to |S,4y relates to the mode of technology promotion employed by MIPH: {3.3)Adopt = /3o + I3,LT30 + I3.GT50 + ^^Primary + fi.Qass + ^^old + ^flousehold + ^jParaiso + ^f,Seed + (3gHerbicide + ^iJnseaicide + ^.uFenilzer + ffi^Lecture + iS^iPublication + jS^^Lecturepub + ^^^ctureaid + ji^f^Prevent Where: Adopt = 1 if farmer adopted the technology, 0 otherwise. LT30 = 1 if farmer's age less than 30, 0 otherwise GT50 1 if farmer's age greater than 50, 0 otherwis j. Primary = Number of years of primary schooling. Class = 1 if farmer attended literacy class, 0 otherwise. Sold = Proportion of output sold. Household = Number of household members. Parafso = 1 if producer is from the region of El Parafso , 0 otherwise. Seed — Years of experience with hybrid seeds Herbicide = Years of experience with herbicide Insecticide = Years of experience with insecticide Fertilizer = Years of experience with fertilizer Lecture = 1 if group received extension lectures without additional teaching aids, 0 otherwise Lectureaid = 1 if group received lectures accompanied by electronic visual aids, 0 otherwise Lecturepub — 1 if group received both lectures and printed extension publications, 0 otherwise Publication = 1 if group received printed extension publications and no personal lecture, 0 otherwise

PAGE 59

47 Results The logit regression is displayed on Table 3.1. The model appears to "fit" the data well, correctly classified estimates (based on a fifty-fifty classification scheme) amounted to 93.46 percent and the model x^, which tests the overall significance of the model, is significant at the one percent level. The higher the proportion of output sold (Sold) shows a strong influence to adopt technology. Primary education has a positive influence {Primary), but not as strong as attendance at (literacy) Class. The Class coefficient could indicate more a personality type than a transformation in thinking to adopt new technologies; farmers amenable to becoming literate might be similarly amenable to adopting new production techniques. Class would thus represent a predisposition to new technologies than an actual change in thinking to adopt technology. The positive sign on Household supports the Boserup hypothesis in that as constraints tighten on production systems, more advanced methods are adopted. Farmers over the age of 50 seem less amenable to adopting new techniques than their younger compaileros . The regional variable Paratso had a positive influence on adoption, a result that could be attributed to better extension services or a more activist campesino union prominent in the region (ANACH). The only significant result among variables attempting to represent experience is shown on the coefficient of hybrid seed. One of the primary goals of MIPH was to help farmers with a history of pesticides misuse. Pesticides had been aggressively promoted by retailers and some government extension agents. IPM technologies, by virtue of their not relying solely on pesticide use, present additional obstacles to optimal technology

PAGE 60

48 diffusion. Integrated pest management requires different modes of technology transfer than traditional pest prevention strategies because commercial pesticide p -omoters benefit from the latter and are often in conflict with the former(Agudelo and Kaimowitz 1991). In terms of extension methods, both Lecture and Lectureaid registered positive and significant influences on adoption. Bodi forms of extension provided personal contact, an important if expensive extension method demonstrated influential in Martin and Taylor (1995). Publication, which provided only printed materials is insignificant. However, it is somewhat perplexing in that Lecturpub, which also provided personal extension, is insignificant.

PAGE 61

49 Table 3.1 IPM Adoption: Pre-infestation maximum likelihood estimaxs Variable Coefficient Std. Error Wald Statistic Significance Constant -3.437 0.422 66.239 0 LT30 -0.320 0.274 1.3632 0.243 GT50 -0.663 0.246 7.2979 0.0069 Primary Education 0.089 0.046 3.6389 0.0564 Class 0.4028 0.1834 4.8266 0.028 Sold 0.542 0.198 7.4896 0.0062 Household 0.057 0.034 2 912 0.0879 Parai'so 0.374 0.194 3.7263 0.0536 Experience with: Seed 0.127 0.034 14.2069 0.0002 Herbicide -0.069 0.041 2.8689 0.0903 Insecticide -0.0386 0.0369 1.0907 0.2963 Fertilizer -0.059 0.037 2.5177 0.1126 Type of extension: Lecture 0.524 0.249 4.4292 0.0353 Lectureaid 0.632 0.262 5.3276 0.0158 Leaurepub 0.060 0.282 0.0452 0.8317 Publication -0.322 0.291 1.2242 0.2685 Chi-Square df Significance -2 Log Likelihood 1102.132 2432 1.000 Model Chi-Square 80.088 15 .000 Improvement 80.088 15 .000 Goodness of Fit 2473.165 2432 .000 Predicted Observed 0 1 Percent Correct 0 2288 0 100.00% 1 160 0 0.00% Overall 93.46%

PAGE 62

50 Summary Technology adoption is a critical first step for increasing the efficiency of HARC production. Improving the diffusion of advanced technologies is becoming increasingly important in Central America as traditional slash and burn methods of production deplete the agricultural resource base which must serve to feed rising populations. IPM techniques offer a comprehensive "proxy" to examine technological packages because they incorporate various aspects of the production system. IPM techniques also offer a more precise means for examining the effectiveness of public extension programs because socially optimal remedies of IPM are often at odds with those that are commercially optimal. This chapter examined IPM technology adoption as a function of human resource factors. Three sets of influences were considered, extension type, experience with inputs, and the demographic factors that predispose producers to opt for new techniques. The model yielded notable insights. The proportion of output sold is positively correlated with the decision to adopt technology. Older farmers appear less likely to adopt, but literacy class and, to a lesser extent primary schooling, are shown to be positively correlated with adoption. Lectures, both alone and accompanied with visual aids positively influenced adoption.

PAGE 63

CHAPTER 4 TECHNICAL AND ALLOCATIVE EFFICIENCY: COLLECTIVE VS INDIVIDUAL One of the most important economic policy issues facing restructuring and developing countries concerns the relative efficiency of collective and individual enterprises. Cooperatives'' theoretical potential has been superseded by the observed and dismal le-el of their failure. The term "collective" is rarely heard in the popular and business media without pejorative antecedents: "inefficient," "lethargic," "bureaucratic-heavy," etc.. Many of the economic problems of formerly communist countries and of poorly performing developing countries are attributed to the waste incurred by collective organizations. Inefficiency obviously exists in cooperatives, but its precise nature is unclear. Furubotn and Pejovich (1970) predicted that cooperatives would degenerate in a capitalist environment as workers/owners have relatively less incentive to make long-run capital investments. However, cooperatives in France, Italy, and parts of the former Yugoslavia have notable success records. And recent studies from Northern Italy (Bartlett, et al. 1992), and the former Yugoslavia (Boyd, 1987; and Piesse, et al., 1996) suggest that cooperatives are more efficient than private enterprises. 'There are several types of cooperatives. Some organize economic agents for specific mutually beneficial activities such as marketing or input purchases. Unless otherwise noted, the term cooperative in this dissertation refers to full production cooperatives in which resources are communally owned, labor is pooled in production and revenues are shared. 51

PAGE 64

52 Conventional wisdom attributes inefficiency of collectives to "shirking," or the abdication of personal work responsibilities. However, it is not clear whether shirking is most detrimental at the level of the worker or at the management level. Two articles laid the groundwork for the issue of shirking in labor-managed firms. Alchian and Demsetz (1972) concentrate on the worker level, contending that individuals of a collective enterprise lack adequate incentive to monitor coworkers because they do not receive the residual claim awarded to managers of private capitalist firms. Even if one member is appointed the task of monitoring, they argue, the monitor/manager has no authority to hire and fire, and has no incentive to efficiently utilize and maintain fixed capital because the individual portion of capital returns is less than the personal trade-off between labor and leisure. Alternatively, Jensen and Meckling (1979) submit that shirking inefficiencies are most problematic at the level of management. They consider it "naive" to believe that managers of collectives would take the same pains to "seek out high pay-off new proje;ts, to weed out projects which have negative pay-offs, to control waste and shirking, etc." without an addiuonal claim on returns. The distinction between worker shirking and management shirking is an important one. The alleviation of widespread worker shirking involves substantial monitoring costs and may be an exogenous social characteristic unresolvable by policy modifications. The success of the Israeli Kibbutz, for example, as well as Amish and Hutterite communities in the United States, is often attributed to pre-existing religious bonds that preclude labor shirking. Management inefficiencies, on the other hand, are more easily overcome through restructuring and incentive realignments. Different sources of shirking generate characteristically different inefficiencies. The selection of inputs is determined through management, but the actual input use is exercised by

PAGE 65

53 workers. Suboptimal selection of inputs results in allocative inefficiem-y. However, any input combination can be used in a technically efficient manner, and is a function primarily of worker responsibility". If "net-shirking," shirking so extensive that it over-rides gains achieved by economies of size, occurs at the worker level as posited by Alchian and Demsetz (197',), technical efficiencies will be lower on collective systems than individual systems. Allocati\ e inefficiencies may be caused by two factors. Allocative inefficiencies could well be the result of management disincentives, as maintained by Jensen and Meckling (1979). However, allocative inefficiencies could also result from friction in input distribution systems, which are notably inefficient in Honduras. This chapter compares technical and allocative efficiencies of HARC individual and collective maize production systems. Most research on issues related to :ollective vs. individual production, beginning with Ward's "firm in lllyria" (1957) is theoretical in nature. Empirical studies are rare because data sets comparing collective and individual enterprises are lacking. A recent article by Carter, et al. (1996) examined Honduran agrarian re.brm cooperatives. The authors estimated a standard OLS production function regression which included a dummy variable on collectivity which suggested that collective organization had a positive effect on production. The same study showed that 90 percent of surveyed H \RC members did not consider shirking a problem. This chapter is intended to contribute to tiie empirical side of the discussion. ^0 the extent that workers are not properly trained and motivated, technical inefficiency may also be attributed to management. However, in the case of HARCs, management makes no decisions regarding the training of workers. HARCs exist in large part to provide training to farmers who have demonstrated motivation through political activism.

PAGE 66

54 The first section of this chapter reviews the theoretical underpinnings and statistical methodologies used to estimate and analyze efficiencies. The second section presents and discusses the statistical results. Productive Efficiency The terms productivity and efficiency are often used interchangeably, but subtle differences are noted in die contexts in which they are used and thus in their evolving definitions. Productivity is defined, in its simplest form, as output obtained per unit of input. Productivity measures are usually ratios of output to total or partial input. Such ratios are relative to each other and thus cannot comprehensively differentiate sources of productivity changes. Measures of total factor productivity (TFP), calculated as monetary values or as weighted combinations of physical inputs, fail to distinguish differences in technology as well as the effectiveness with which the technology is implemented. Efficiency measures, by contrast, are derived by measuring the variation in the input/output relationship to a technical maximum. That maximum is specified by a neo-classical production function and thus represents the maximum output or frontier 1 jvel of output attainable from a given set of inputs. The distance between the observed input/output relationship and the frontier production function reveals inefficiencies associated with any input combination. Frontier measurements of the observed technical maximum offer the most suitable vehicle for accomplishing objective three^ in Chapter One of this study and illuminating policy options. Productivity measures, because they are devoid of theoretical constructs and because they are 'Objective #3 is: Compare the technical and allocative efficiencies of individual production systems vis ^ vis collective production systems.

PAGE 67

55 often partial measures, do not offer adequate means to compare efficiencies across groups of producers. A farmer wiio owns one donicey may appear more efficient than one who owns ten even though the one-donkey owner has made gross overinvestments in machinery and equipment. Frontier productions, on the other hand, simultaneously encompass all inputs in a theoretical framework from which efficiency estimates may be derived. The inclusion of all relevant factors, including human capital and institutions, conceivably places everyone on the optimal frontier, rendering frontier and "average" functions indistinguishable (Miiller, 1974). However, production functions can be reliably estimated in terms of scarce material resources whose impact can be quantified and measured in marginal terms. Human capital and institutions, on the other hand, while important to the production process, do not lend themselves to precise evaluation for two reasons. First, they are a nebulous concepts that cannot be described by cardinal numbers. Years of schooling and courses completed yield exact numbers, but there is no way to calibrate the quantity and quality of knowledge that actually contributes to production. Institutions that govern resource use are important policy alternatives for improving the efficient use of scarce material resources, but are similarly incapable of numeric characterization. Second, there is no rigorous theory to guide analysis of the marginal relationships human capital and institutions hold with production. Frontier functions are, however, useful tools in determining how human capital investments and organizational alternatives directionally alter the efficient use of scarce resources. Thus inquiry can be made into the human capital attributes and institutional parameters of producers operating close to the frontier. Are they more educated? Does family background inspire greater efficiency? Are they better trained and experienced? What rules govern their actions?

PAGE 68

56 Such information is useful to policy makers wanting to identify human resource investments and rules and regulations which have the highest payoff. Knowledge of human capital distinguishes human resource investments that enhance the "value of the ability to deal with disequilibria" (Schultz, 1975) caused by changes in technology. Similarly, organizational designs may be identified that contribute to the efficiency of over-all material resource use. Technical Efficiency The most widely used efficiency measures are rooted in the writings of Debreu (1951) and Farrell (1957). The Debreu-Farrell measure of technical efficiency is defined as the equiproportionate reduction of all inputs that produces a demonstrated optimal level of outputs. Conventionally, it is the ratio of observed output to optimal output for a given set of inputs. Thus, unity represents 100 percent efficiency and a fraction less than one indicates a measurable level of inefficiency. The conventional base within which efficiency is evaluated is the production technology. Technology establishes the limits at which inputs are capable of producing given levels of output. Let inputs x and outputs q be represented respectively by: X = (xi, X., X3,...xJ e R" q = (Qi, 02, q3,---qJ ^ R"The input requirement set of the production technology is (4.1) l(q) = {x: (q, x) is attainable), which represents all combinations of inputs, efficient and inefficient capable of producing q„. R(q) is shown in Figure 4.1 as the shaded area.

PAGE 69

Figure 4.1 Input requirement set

PAGE 70

58 The boundary set of inputs in Kq) shows the least amount of inputs necessary to produce with the current level of technology. Along the boundary, reduction in any input would reduce output or require additional use of a substitute input in x. The boundary set is defined by the isoquant (4.2) $(q) = {x: X e Itl(q) and Xx ^ R(q) if 0 < X < 1}, which excludes infeasible input combinations in R°. The concept of efficiency requires (q) to be convex, or at least quasi-convex relationship, as efficient input combinations use as little of each X; as possible to produce q. The Debreu-Farrell measure of technical efficiency can be formally interpreted as (4.3) TE(q, x) = min{d\ Bx G a(q)} < 1." Optimum efficiency yields a technical efficiency measure of unity, which is identical to the isoquant (4.4) $(q) = {x: TE(q, x) = 1}. If one assumes only one product is produced, the production technology may be described using a production function, and its associated isoquants. A prjduction function is a mathematical form that relates the maximum possible output attainable from given quantities of a set of inputs. (4.5) q = q(x) = max{q: x € a(q)}. The Debreu-Farrell technical efficiency measure is expressed by the ratio (4.6) TE(q,x) = q/q(x). The observed level of output, q, cannot be greater than the maximum level of output q(x), delineated by the input-output combinations of the most efficient producer s. Clearly, then, TE(q, "The Debreu-Farrell technical efficiency measure is the inverse of the distance function (Shephard 1953, 1970) where D(q, X,) =max{e: {xld) G a(q) > 1}.

PAGE 71

59 x) < 1 because the production function is a technological frontier that can be achieved, but not exceeded, by technically efficient producers. AUocative Efficiency Technical efficiency is derived solely from the input-output relationship; prices, and thus any notion of cost minimizing behavior, are absent. AUocative efficiency, defined as the optimal combination of inputs to produce a given level of output, is residually obtainable from costminimizing frontier. Given input prices associated with the input vector x (4.7) w = (w„w„W3,...wJ E R", and a cost minimizing frontier is represented by (4.8) c(q,w;T) = min,{w^x: I/TE(q,x;T) < 1} where t is a vector of parameters representing optimum technology. The ratio of the frontier cost function to the actual cost incurred yields a measure of economic efficiency (Farrell, 1957) (4.9) EE = c(q,w;T)/x\v^ where x'^ is the vector of actual inputs and is the price vector. AUocative efficiency is the ratio of economic efficiency to technical efficiency (4.10) AE = EE/TE = wV/TE(q,x)). AE < 1 as EE < 1 = TE*AE, economic efficiency can only be unity if full technical and allocative efficiency are achieved. Technical and allocative efficiencies are illustrated in Figure 4.2.

PAGE 72

60 Figure 4.2 Farrell technical and allocative efficiency

PAGE 73

61 The unit isoquant 1° represents a frontier production function utilizing inputs Xj and Xj. A technically efficient producer is represented by B, which lies on the production frontier. A producer operating at point B or any oUier point along the frontier cannot reduce any one input without either increasing another input or reducing the level of output. Point A, on the other hand, represents an inefficient producer, who utilizes more inputs than B to produce the same level of unit output. The producer at point A can reduce input use without any reduction in the unit level of output represented by 1°, The technical efficiency measure is then: (4.11) TE = II B II / II All \ Many producers, given differences in managerial and work skills, are likely to fall short of the technically efficient frontier. The amount by which a producer lies below the optimal production frontier can be regarded as a measure of inefficiency and may be accounted for by human capital and institutional differences. The Debreu-Farrell technical efficiency is a "radial" measure in that, if the isoquant is weakly convex, producers located on the portion of the isoquant where the slope is zero or infinite are considered technically efficient. The Debreu-Farrell definition of technical efficiency is thus not as restrictive as that proposed by Koopmans. Koopmans (1951) defined technical efficiency as the state where increasing one output requires a decrease in another output or an increase in at least one input; and the reduction of one input requires an increase in another input or results in reduced output. Clearly, more efficiency is attainable if reducing one input results in no reduction of output, a necessary condition for Koopmans, but rot for Debreu Farrell. However, while slack* may pose some problems for mathematical programming estimates, it 'In vector notation || X || = (Sx;-)''^, vi. *"Slack" refers to the range of an isoquant which may have a slope of zero or infinity. In such a range, a producer may be considered technically efficient, even iiough he could reduce one input quantity an maintain the same level of output.

PAGE 74

62 does not hamper econometric techniques because functional forms (e.g. Cobb-Douglas) preclude slack. The concept of allocative efficiency, a measure of cost minimizing performance, is depicted in Figure 4.2 as well. In order to minimize costs, production must be set at a level where the ratio of input prices equals the marginal rate of technical substitution, or point E on Figure 4.2. Allocative efficiency is measured by the distance separating the price line and the efficient isoquant: (4.12) AE = ||C||/||B|| . Allocatively efficient producers adjust input mixes where the marginal rate of technical substitution is equal to the price ratio. Producers can be technically efficient but allocatively inefficient. Allocative inefficiency may be a result of management deficiency or it may appear when input markets are distorted. Institutions are thus more important to allocative efficiency than to technical efficiency. Both technical and allocative efficiency approach unity as they achieve optimum or "frontier" efficiency. Two empirical methodologies have predominated Debreu-Farrell efficiency measurement, non-parametric frontiers and parametric (predominantly econometric'' frontiers. They are discussed in the following two sections. Nonparametric Frontiers Farrell's approach (1957) is considered non-parametric and deterministic because the convex hull of input-output ratios is constructed by mathematical programming techniques. Mathematical programming techniques obtain the convex hull of the input requirement set that represents the smallest input combinations for a given level of output. Data Envelopment Analysis, a management science and operations research approach originally proposed by

PAGE 75

63 Charnes, Cooper and Rhodes (1978, 1981) and recently discussed in Seiford and Thrall (1990), dominate recent nonparametric applications. Data envelopment analysis (DEA) was originally proposed as a technique for evaluating the efficiency with which a group of operating units transforms inputs into outputs (Charnes, Cooper, & Rhodes, 1978). DEA has been applied primarily to public sector institutions where prices are absent or unreliable. Consequently, most DEA studies involve technical efficiency only, although allocative efficiency can be calculated given sufficient price data. DEA is formulated as a fractional linear program, continuing along the same math programming path initiated by Farrell. In concept, the procedure maps input-output data to find the subset of most efficient producers who comprise the frontier against which other producers are compared. The most efficient units in the group define a production function that is linear in a "piecewise" fashion. Input-output relationships are linear at each piece of the function between efficient units, although they are not necessarily linear across all efficient units. The function describes a hyperplane "efficient surface" equal in dimensions to the number of inputs and outputs. Input-output relationships for the remaining units are then evaluated relative to this efficient surface. DEA has always been an attractive alternative for measuring efficiency because it imposes no functional form on the data. The basic DEA approach is given by the specification of a transformation function T, restricted by constant returns to scale and strong disposability: (4.13) T = {(x,q):Xe
PAGE 76

64 output subject to observed input/output relationships. They are calculated with mathematical programming techniques: (4.14) TE,(Xi,qi) = Min X. St: xeqA 0eR% where \ represents the level of inefficiency of the ith producer. The program produces a scalar efficiency measure for each unit by selecting weights 0 that maximize the ratio of a linear combination of the unit's outputs to a linear combination of its inputs. This efficiency measure is constrained so that the weights selected must be feasible and cannot result in an efficiency ratio greater than that observed for the most efficient unit in the group. Later developments have unencumbered Farrell's original programming approach from the assumptions of constant returns to scale (Banker, et.al., 1984) ana strong disposability of inputs and outputs (Fare, et.al., 1985, 1987). Nonparametric frontiers are advantageous in that they only require that the functional form of the technology be nondecreasing and concave. DEA methods are particularly useful in modeling operational processes that do not conform to standard market assumptions, such as in nonprofit firms or regulated industries. Nonparametric methods may also be used to test for cost minimization or profit maximization (Varian, 1984) and have demonstrated the potential for economies of scale that went undetected by econometric estimates (Banker et al., 1986). However, nonparametric efficiency measures are severely limited by the lack of statistical properties. Moreover, they attribute all measurement error to inefficiency, allowing nothing for random uncontrollable events, and by the same token are very sensitive to outliers.

PAGE 77

65 Parametric Frontiers The predominance of parametric frontiers employ statistical methods to analyze the transformation of inputs into outputs. Prior to the emergence of frontier functions, the conventional model that was estimated took the form (4.15) q = q(x,T)-e where q(x,T) is the production function, q and x are vectors of outputs and inputs respectively, T represents technology and e is the random error. OLS necessarily assumes the expected value of the disturbance term, e, is zero because it estimates parameters of variables by minimizing the sum of the squared errors. However, neo-classical production theory defines the production function as the maximum output obtainable from a given set of inputs. In the absence of random error, e > 0 because observed levels of output cannot exceed the theor'3tical maximum. Aigner and Chu (1968) proposed a "parametric deterministic" procedure to calculate the frontier with mathematical programming techniques. They generalized a function with a onesided error term which required all output levels to lie on or below the calculated frontier. (4.16) q < q(x;^) where j3 represents the frontier "estimates". The resulting linear programming function took the form:

PAGE 78

66 R (4.17) '=1 s.t. i 0 i = i,..,« p ^ 0. The quadratic formula is the square of the bracketed term of the objective function. The measure of inefficiency is given by the ratio of the actual level of output to the frontier level of output. Although the programming approach yields parameters, it precludes the testing of statistical confidence in that no standard errors can be computed. The estimation procedure also imposes a structure on the technology and is sensitive to outliers. Schmidt (1976) showed that OLS estimation of Aigner and Chu's formulation yields best linear unbiased estimates of the slope coefficients, but not of the intercept. He also showed that Aigner and Chu's linear programming estimates are maximum likelihood if the error is nonnegative and has an exponential distribution (4.18) f^e) = -exp(-«/(p) tfiO


PAGE 79

67 (4.19) fie)—^ejj>i-e'l2(^) However, Schmidt acknowledged that both maximum likelihood estimators violate the regularity conditions in that the range of the observed dependent variable is depeident on the parameters being estimated. A major advantage of estimating frontiers with econometric techniques is that they render statistical properties to efficiency measurements. Parametric frontie: estimations impose a structural form on the technology, but unlike nonparametric methods they provide statistical measures, account for uncontrollable shocks and are less vulnerable to outliers. Early parametric techniques (Schmidt, 1976; see also Fersund et al., 1980) yielded "full 'rontier" measures that, like nonparametric measures, attributed all deviations from the frontier to inefficiency (e ^ 0). Attributing all deviation from the frontier is unrealistic in that it does not account for random uncontrollable events beyond the purview of management. Tne "stochastic frontier" (Aigner Lovell and Schmidt. 1977; Battese and Corra, 1977; and Meeusen and van den Broeck, 1977) allows for random deviation from the frontier owed to measurement error or events beyond the control of the producer. The error term of the production functiori (4.5) in the stochastic frontier is comprised of two components: (4.20) e = {Vu) where v has a symmetric distribution which captures random effects and exogenous shocks across firms; and the one-sided error, « > 0, captures technical efficiency of a firm relative to the stochastic frontier. Thus the estimated frontier accounts for stochastic characteristics that are

PAGE 80

68 likely to aifect any production system, isolating systematic effects in the measurement of technical inefficiency. If u is assumed to have a half-normal distribution the associated log-likelihood function is (4-21) ^M,a,X) = -Mna-K + ^[lnT(-^)-^(^)^ where X = aja,, a= a^-la,-, and is the cumulative distribution function of the standard normal distribution. As ^ <» , X ^ 0, systematic inefficiency increases relative to random inefficiency. Assuming m is exponentially distributed yields the likelihood function (4.22) _^ which is parametized in


PAGE 81

69 specifying a functional form for the distribution of u given the composed error term e, Jondrow et al. demonstrated that point estimates of efficiency are obtainable for each observation. The expressions for the expectation of u given e of the half-normal and exponential models are (4.23) EluM-^l-^^^^-^} and (4.24) £[„|,i = (e.-,a,^ + M^5^L:l^^ respectively. T[(«,-
PAGE 82

70 were identified as superior based on engineering designs and experiments. Technical efficiency is based theoretically on the same concept regarding input-output relationships. However, technical efficiency is examined based on the observed best practice among farmers in the sample. It is possible then that farmers appear to achieve optimum efficiency even though they did not employ the optimum possible technology because they use fewer of the inputs of suboptimal technologies. Graphically, the actual isoquant may lie between the best practice isoquant 1° and the origin in Figure 4.2. Most technical efficiency studies recognize that it is virtually imoossible to construct the technically efficient frontier from engineering knowledge. Even simple agricultural production systems are too complex. It was that insurmountable challenge that lead Michael Farrell (1957) to propose the construction of the isoquant based on the observations of best practice producers. The only assumption necessary in input-input space is convexity, which is virtually synonymous with efficiency. Technology adoption studies can assume recommended technologies are optimal because they focus on specific, comprehensible aspects of production. To the extent to which technologies are developed and recommended based on a broad assessment of all inputs available to producers, and the most efficient producers adopt those technologies, the observed frontier and the actual or potential frontier are identical. The MIPH program endeavored to design programs that both incorporate farmers' available inputs and to communicate them to farmers. The IPM technologies developed by MIPH are relatively more comprehensive than most in that they incorporate several production facets related to pest control. However, MIPH technologies carmot represent the entire gamut of possibilities available to farmers. In sum, the coefficients on technology adoption (3.3) and the coefficients on the frontier production functions in the following sections of this chapter have slightly different interpretations

PAGE 83

71 with respect to technology. The former refers to available technology as determined by engineering studies related directly to the physical input-output relationship; the latter refers to an observed technological optimum as demonstrated by "best practice" producers. The HARC Stochastic Frontier Most problems associated with parametric frontiers concern the data to which the method is applied. Relevant variables concerning human capital and institutional characteristics are rarely available. When relevant variables are available, they often cannot be included in the estimations because they present degrees of freedom problems due to limited samples. Field surveying for this study was carried out in a manner designed to overcome these problems. Data were gathered on several aspects of HARC operations to expand the explanatory power of the models and sufficient observations were obtained to avoid problems with degrees of freedom. The stochastic frontier function thus appears to be the best suited for evaluating various aspects of HARC efficiency. Standard criticisms of technical efficiency methods (Milon, 1987) do not apply to HARCs or the environment in which they operate. Contractual relations, with the exception of internal labor contracts which are documented, do not vary across HARCs, unmarketed factor inputs are included in the analysis and the means and ends of the HARCs are arguably homogeneous. Externalities, while important to broader welfare considerations than considered here, would not likely have much of an influence on efficiency measurements because harmful inputs are not used nearly as extensively in HARC basic grain production as they are in developed countries. Data on collective and individual production systems were u^ed to estimate HARC frontier parameters for maize production. The Cobb-Douglas model is selected as the functional form for its convenient properties. The general form of the Cobb-Douglas is

PAGE 84

72 k m n (4.25) q ' AY[x^'Y[h:'Y[c':'u i=l 1=1 j=l where q is a producer's output, A is a given level of technology which "shifts" the ftinction in response to technological changes, represents the set of i = l...n inputs and the fij's are the corresponding input coefficients. The standard production function estimates output q, solely as a function of physical inputs x^. However. Jensen and Meckling {\919^ suggested an extended form of the production function which recognized that production did not occur in a physical vacuum. Knowledge h, (human capital) and "organizational forms" O; also influence the level of output by their parameters and y-^ respectively. The Cobb-Douglas has several convenient properties. and 52/3,= 1 the form is concave. Thus, assuming the firm minimizes costs and factor supply and product demand functions are continuously differentiable on their domains, the input denand and output supply functions are continuously differentiable everywhere on their respective domains, a very useful property for interpreting results. The Cobb-Douglas is homogeneous and thus provides a means for examining returns to scale in that 52/3i= 1 characterizes constant reiurns to scale, 52/3; > 1, increasing returns to scale; and 52/3i< 1, decreasing returns to scale. However, the elasticity of substitution remains unity for all levels of output, weakening scale obser'ations. Most important for this study, the dual cost function, necessary for the calculation of allocative efficiencies, can be derived directly from the Cobb-Douglas production function. Although considered restrictive in some instances, the Cobb-Douglas was developed for and has been used extensively in agriculture in both developed and developing countries.

PAGE 85

73 Moreover, functional form has been shown to have minimal impact on efficiency estimates (Kopp and Smith, 1980). The systematic element is seen in the right-hand side of 4.25 in that ^ represents onesided efficiency disturbance : k m n (4.26) g = AH xf'H ^/'« " « ^^"^^ i=l j=l i=l and q* frontier output. The specific model estimated for maize production is: (4.27) InMaize = (3, + l3,\nLand + p.^nLabor + ^^\nSeed + ^^InFertilizer + ^^XnHerbicide + jS^lnLandprep + ^jCollectiviry + ^^Paratso Region + PJLecture + ^.oPublicatio + ^.^Lectureaid + ^^Xecturepub + e where: Variable Coefficient Tecnology (A) Constant Land Land measured in manzanas* Labor Labor measured in work days Seed Seed measured in pounds Fertilizer Fertilizer measured in quintals Herbicide Herbicide measured in pounds Landprep Total cost of land preparation Collectivity Degree of collective work arrangements" Paraiso Region = 1 if producer is from the region of El Parafso, 0 otherwise. Lecture 1 if group received extension lectures without additional teaching aids, 0 otherwise Publication = 1 if group received printed extension publications and no personal lecture, 0 otherwise

PAGE 86

74 Lectureaid = 1 if group received lectures accompanied by electronic visual aids, 0 otherwise Lecturepub = 1 if group received both lectures and printed extension publications, 0 otherwise 'Manzana = 0.705 hectare "Parcels used completely in the collective mode are scored as one, those planted prior to jarcelization are scored as one half, and those for which the only collective activity land preparation are scored as one fourth. Completely individual production is registered as zero. is the constant term on the technology variable A in 4.24. Different levels of technology "shift" the function. Land, Labor, Seed, Fertilizer, Herbicide, and Land Preparation, represent continuous variable inputs and thus are in log form. Human capital and organizational form variables are included in the production function because they may have a direct impact (Battese, Coelli and Colby, 1989). All extension techniques are included in the model as dummy variables. The variabl j for the control group, which received no extension assistance, is necessarily excluded to avoid a singular matrix. Collectivity, the variable representing organizational form (Jensen and Meckling, 1979), is calibrated according to the point at which collective operations are yielded to the individual responsibility of each HARC member. In some HARCs land is prepared collectively, but planting and all subsequent cultivation is conducted individually. A few HARCs with large individual parcels do not partition land until planting and initial applications of fertilizer are completed. The Collectivity variable is included in the model for maize to provide additional means of examining the extent to which collectivization at various degrees influences production. Production functions are estimated with the maximum likelihood technique (equations 4.20 and 4.21). The half-normal and exponential distributions are assumed for the one-sided error. The distinctive effect each distribution has on the frontier is not well known (Bauer, 1990), but Greene (1990) suggests that there is not much difference between the two. Maximum

PAGE 87

75 likelihood estimates and technical and allocative efficiencies were estimated using the LIMDEP software program. Technical efficiencies from 4.24, are: Allocative efficiency (4.10) is calculated by analytically deriving the dual cost frontier (Kopp and Diewert (1982) and Bravo-Ureta and Rieger (1991)) to obtain a measure of economic efficiency (4.9). It is not necessary to estimate the cost function as a regression equation. Rather, the coefficients from the production function are incorporated into the cost function which is a dual representation of the production function. Total cost is a function of input prices and output, the minimization of the costs of producing given levels of output. The cost function reveals all the economically relevant information about the technology (Shephard, 1970; Cornes, 1992). Varying price vectors reveal, through the assumptions of cost minimization and convexity, ranges of input mixes that are observed technically efficient. The dual cost frontier is: The cost function parameters a and m, are analytically derived directly from the estimated parameters of the production function (4.24) where: (4.26) TE (4.27) (4.28) , Jk=^[in..pf']-'".

PAGE 88

76 Results Information on the study area and data are in Appendix I. The human capital survey is in Appendix II. Ordinary least squares (OLS) estimates are displayed on Table 4. 1 . Tables 4.2 and 4.3 display maximum likelihood (ML) estimates of the stochastic frontier where half normal and exponential distributions are assumed for the truncated distribution component, (m). The Cobb-Douglas model fits the maize production data well. The of the OLS is 0.88. All but one of the standard physical input variables, herbicide, are significant at the 0.01 probability level for all three regressions. The value of X = aja, is 5.1 and significant at the 0.05 level. This suggests that relatively more of the variation in the model is attributed to random, not systematic influences. As X approaches infinity, the average and frontier functions converge Coefficients of the Cobb-Douglas production function represent output elasticities for each input. The average function indicates that a one percent increase in land, ceteris paribus, would generate a 0.46 percent increase in output. Frontier estimates indicate a similar increase in land would yield a 0.56 increase in output, 0.09 higher than the average. The coefficient on Seed is higher in the average function than in the frontier functions, though the difference is less notable. There appears no appreciable difference between average and frontier functions for the coefficients on Labor, Fertilizer, Herbicide, and Land Preparation. The variables Collectivity and Paraiso Region are positive and significant. The parameters on extension variables are inconsistent with the results of those estimated in the technology adoption model of the previous chapter. Lecture and Publication and Publication only, the only two extension variables that were not significant in the logit regression that measured technology adoption, are both positive and significant in the frontier function.

PAGE 89

77 Table 4.1 Average function: Maize ordinary least squares regression Variable Coefficient Std . Error t-ratic Prob|t| >x Constant 2.295 0.231 9.938 0.000 Land 0.464 0.066 7.028 0.000 Labor 0.129 0.039 3.287 0.001 Seed 0.175 0.058 3.010 0.003 Fei^ilizer 0.035 0.012 2.839 0.004 Herbicide 0.016 0.007 2.2.12 0.026 Land Preparation 0.044 0.017 2.533 0.011 Collectivity 0.112 0.016 6.900 0.000 Paraiso Region 0. 151 0.041 3.691 O.OUU Lecture 0.052 0.060 0.861 0.389 Publication 0.212 0.069 3.055 0.002 Lecture and Visual Aids 0.039 0.058 0.672 0.501 Lecture and Publication 0.171 0.062 2.756 0.006 Observations: 405 R-squared: 0.877 Adjusted R-squared: 0.873 F[ 12, 392]: 233.246 Log-likelihood: -153.314 Restr.(fi = 0) Log-1-577.919 Amemiya Pr. Criter.: 0.821 Akaike Info.Crit.: 0.133

PAGE 90

78 Table 4.2 Frontier function: Maize maximum lilcelihood estimates half normal distribution Variable Coefficient Std. Error t-ratio Prob|t| >x Constant 2.806 0. 194 r\ AAA Land 0.556 0.052 10.784 A AAA 0.000 Labor 0.122 0.028 4.317 A AAA 0.000 Seed 0.116 0.043 2.690 A AAT 0.007 Fertilizer 0.023 0.009 2.540 A A 1 1 0.01 1 Herbicide 0.011 0.006 1.693 A AAA 0.090 Land Preparation 0.049 0.013 3.738 A AAA 0.000 Collectivity 0.085 0.014 5.997 A AAA 0.000 Paraiso Region 0.099 0.034 2.872 0.004 Lecture 0.020 0.059 0.336 0.737 Publication 0.129 0.062 2.088 0.037 Lectureaid 0.093 0.054 1.717 0.086 Lecture and Publications 0.131 0.059 2.209 0.027 2.856 4.057 0.704 0.481
PAGE 91

79 Table 4.3 Frontier function: Maize maximum lilcelihood estimates exponential distribution Variable Coefficient Std. Error t-ratio Prob|t| >x Constant 2.800 r\ 1 TO 0. 176 n (\c\t\ U.UUU Land 0.559 0.049 11.360 A AAA U.UUU Labor 0.123 0.027 4.557 A AAA U.UUU Seed 0.115 0.042 1.151 A AA/C U.UUo Fertilizer 0.022 0.009 O CI 1 2.511 Ann U.Ulz Herbicide 0.010 0.006 1 /;oo 1.O50 A AQ1 u.uy 1 Land Preparation 0.050 A A 1 '5 0.013 j.y^i A AAA U.UUU Collectivity 0.086 0.014 6.258 0.000 Paraiso Region 0.095 0.033 2.854 0.004 Lecture 0.015 0.056 0.271 0.787 Publication 0.125 0.060 2.039 0.037 Lecture and Visual Aids 0.097 0.053 1.830 0.067 Lecture and Publication 0.127 0.057 2.230 0.026


PAGE 92

80 It would be imprudent to argue that any of these dummy variable estimates represent precise measurements of the influence each bears on production. However, they do provide valuable directional indicators. Thus the absolute values are less relevant than the sign of the variables or relative extremities. Technical and AUocative Efficiencies Technical and allocative efficiency measures are presented in Tables 4.4 and 4.5. Collective efficiencies are the calculated efficiency measures for the collective parcel of each HARC, and individual parcel efficiencies are presented as averages for each HARC. The standard deviations of individual technical efficiency averages are less than a third of the average for all but two cases, suggesting that technical eftlciency does not vary substantially within HARCs. The uniformity of efficiencies within HARCs may be explained by the communication provided by cooperatives as they are in part established to facilitate communication across large numbers of farmers. Empirical evidence (Martin and Taylor, 1995) attests to the facilitating role cooperatives play in communication. It was also observed throughout the course of fieldwork that new inputs and new techniques were duplicated by other farmers within HARCs, in some cases reenforcing errors. The most salient feature regarding the efficiency measures is that individual parcels are no more efficient than collective parcels. In fact, efficiencies based on the half normal (Table 4.4) and exponential (Table 4.5) distributions show that collective parcels are more technically efficient than individual parcels for 11 of the 16 HARCs that employed both modes of production. Thus one of the greatest theoretical arguments, and easily '^he most touted political arguments against collective forms of enterprise, shirking, shows no empirical basis.

PAGE 93

81 Table 4.4 Technical and allocative efficiencies for maize half normal truncated distribution Technical Efficiency Allocative Efficiency Unpaid Labor w=0 w=5 HARC Individual Collective Individual Individual Collective Ideas en Marcha .80 .37 .42 (.07)' (.10) (.10) El Boqueron .41 .78 .08 .09 .24 (.16) (.09) (.09) Empalizada .71 .87 .30 .32 .32 (.15) (.24) (.23) El Benque .71 .18 .22 (.19) (.13) (.15) Los Bienvenidos .76 .91 .14 .20 .47 (.13) (.11) (.09) El Esfuerzo .81 .33 .33 (.06) (.05) (.05) Los Peregrinos 77 1 .J J .jj (.08) (.12) (.10) Esquilinchuche .79 .30 .37 (.19) (.20) (.20) San Nicolas .52 .50 .14 .17 .13 (.55) (.18) (.23) Los Almendros .70 .79 .15 .24 .38 (.10) (.11) (.07) La Esperanza .73 .79 .17 .24 .24 (.14) (.13) (.12) Santa Cruz .77 .28 .37 (.11) (.15) (.15) Cayo Blanco .75 .84 .10 .12 .25 (.14) (.06) (.06) Zopilotepe .79 .24 Guaymuras .91 .42

PAGE 94

Table 4.4 Technical Efficiency HARC Individual Collective La Concepcion a 1 Q1 10 46 (. 10) c 1 n ( OR) San Juan de Linaca .OJ on .yu . jj (.(JO) ( \2) L,a rllZUnCd . / o 24 28 ( 09.) ( 09) 1 empiscapa 20 17 C 08"> La rroviaencia .00 .JO 16 ( \6) 19 de Abril .70 .74 .16 1 o .18 .26 (.21) (.11) (.10) El Coyolar .86 .32 .37 .36 .05 (.07) (.07) (.08) El Plomo .77 .81 .22 .20 .29 (.14) (.12) (.07) Los Dos Naranjos .82 .87 .31 .31 .43 (.06) (.05) (.06) Los Venecianos .76 .21 .22 (.11) (.11) (.12) La Libertad .72 .54 .19 .19 .10 (.15) (.12) (.11) Montanuelas .80 .34 82 Ailocative Efficiency Unpaid Labor w=0 w=5 Individual Individual Collective ' Standard errors are in parentheses.

PAGE 95

83 Table 4.5 Technical and allocative efficiencies for maize exponential truncated distribution HARC Technical Efficiency Individual Collective Allocative Efficiency Unpaid Labor w=0 w=5 Individual Individual Collective Ideas en Marcha El Boqueron Empalizada El Benque Los Bienvenidos El Esfuerzo Los Peregrinos Esquilinchuche San Nicolas Los Almendros La Esperanza Santa Cruz Cayo Blanco Zopilotepe Guaymuras La Concepcion .82 (.06)' .41 (.16) .72 (.15) .72 (.20) .77 (.13) .82 (.05) .79 (.07) .80 (.19) .52 (.55) .71 (.10) .74 (.14) .79 (.11) .76 (.14) .82 (.10) .79 .87 .91 .78 .51 .80 .79 .84 .80 .91 .92 .46 (.09) .11 (.12) .37 (.24) .23 (.14) .17 (.13) .41 (.05) .39 (.13) .34 (.21) .15 (.20) .20 (.14) .22 (.15) .34 (.17) .12 (.08) .52 (.08) .12 (.12) .40 (.23) .27 (.16) .25 (.11) .41 (.05) .45 (.09) .42 (.21) .18 (.26) .32 (.09) .31 (.12) .45 (.14) .14 (.06) .31 (.12) .37 (.08) .31 .38 .51 .46 .18 .48 .32 .31 .31 .46 .51

PAGE 96

Table 4.5 84 Technical Efficiency Ailocative Efficiency Unpaid Labor w=0 w=5 HARC Individual Collective Individual Individual Collective San Juan de Linaca .86 .91 .38 AC CQ .DO (.06) (.16) (.13) La Puzunca .80 .30 .35 (.08) (.13) (.08) Tempiscapa .75 .25 .22 (-12) (.10) (.08) La Providencia .69 .56 .24 .29 .23 (.22) (.18) (.17) 19 de Abril .71 .76 .20 .23 .36 (.21) (.13) (.11) El Coyolar .87 .32 .44 .43 .06 (.06) (.07) (.08) El Plomo .79 .82 .27 .24 .36 (.14) (.13) (.07) Los Dos Naranjos .83 .88 .38 .38 .49 (.05) (.05) (.05) Los Venecianos .78 .28 .27 (.10) (.13) (.13) La Libertad .74 .55 .24 .24 .15 (.14) (.14) (.12) Montaiiuelas .81 .43 ' Standard errors are in parentheses

PAGE 97

85 Two of the five remaining HARCs show individual parcels are only scarcely higher. Individual technical efficiencies for Los Peregrinos and San Nicholas (the only for which the standard error is higher than the average) exceed their collective measures by one and two points respectively. Differences are much greater in the remaining cases where collectives are more technically efficient; at least nine points for seven HARCs, two show five point differences and the remaining two HARCs register differences of four points. Allocative efficiencies are markedly lower than technical efficiencies for both collective and individual systems, and vary proportionately across HARCs with technical efficiencies. Curiously, allocative efficiencies increase for 16 of the HARC individual production systems, in both the half-normal and exponential distributions, when the standard wage of five lempiras is imputed for free labor.' The average allocative efficiency for the imputed wage is 0.05 greater than the same average when no wage is imputed, a difference that is significant at the .01 level of probability. Higher allocative efficiencies for the case of imputed wages suggests that there is an opportunity cost for labor on individual parcels vis k vis other opportunities available to farmers. If labor had been over-employed on individual parcels, allocative efficiencies would decrease when the standard wage is imputed. It would suggest that the shadow price of labor for HARC households is less than the average wage. This result is somewhat contrary to that of Nguyen and Martinez (1979), who found that productivity of the ejido sector in Mexico declines when the market wage is imputed for free labor. The difference here is that allocative efficiency is a more precise measure, reflecting 'Since paid and hired labor was aggregated into one input, wage rates were calculated as total cost of labor divided by the amount of both free and hired labor. When all labor was free a nominal .01 value, .002% of the standard L5.00 wage rate, was imputed for all free labor as it would be mathematically impossible to calculate allocative efficiency with a zero value for any free input. It can be considered a minor opportunity cost.

PAGE 98

86 optimum input mixes as opposed to monetary input-output relationships. It may be that HARC members have competing coop-related responsibilities or opportunities that preclude them from devoting free labor to individual parcels. Work outside the cooperative is limited in rural Honduras. Averages of technical and allocative efficiency differences between collective and individual systems for all HARCs and only those with mixed systems are displayed on Table 4.6. Table 4.6 Differences' in efficiency averages between collective and individual parcels Mixed* HARCs All HARCs Half Exp Half Exp Technical Efficiency Difference 0.00 -0.00 -0.02 -0.02 (0.01)^ (0.06) (0.55) (0.66) Allocative Efficiency wage 0 Difference 0.09 0.08 0.06 0.05 (1.82**) (1.53*) (1.32*1 (1.03) wage = L5.00 Difference 0.04 0.03 0.01 -0.00 (0.86) (0.59) (0.19) (0.05) ' Collective average less individual average ^ Efficiency scores exclusively for HARCs that had both individual and collective or "mixed" production systems. ' t statistics are in parentheses " Significant at the level of . 1 probability level. ~ Significant at the level of .05 probability level.

PAGE 99

87 There is, on average, no statistically significant difference in technical efficiency between collective and individual parcels. The only significant difference in mean efficiency scores occurs in allocative efficiency when free labor receives no direct remuneration. Collective parcels in this instance are more allocatively efficient. The significance is attributed to the low level of allocative efficiency when free labor is not imputed a wage. Summary The most significant result of this chapter is that collectivs parcels appear more technically and allocatively efficient than individual parcels. The chapter begins by reviewing the theoretical and empirical developments of the Debreu-Farrell efficiency measures. A frontier production function is estimated, rather than a conventional production function, because the former reveals the observed maximum that provides a benchmark against which efficiency can be measured. Three functions are estimated, an average OLS function and two maximum likelihood functions, one with a half-normal distribution and another with an exponential distribution. The estimated frontier function has a two component error that differentiates random (stochastic) error from systematic (inefficiency) error. The explanatory value of all three models is high. All but one of the coefficients of the average function are similar to those of the frontier functions. The land coefficient is higher in the frontier function than in the average function, suggesting that the most efficient producers obtain a higher return from land than most producers. Collectivization ';ontributes positively to production, as the collectivity variable is positive and significant. Technical efficiency is calculated based on the most efficient input-output relationship revealed by the frontier. Allocative efficiency, a measure of optimum resource allocation, is calculated based on input prices vis ^ vis the frontier. The results challenge conventional wisdom

PAGE 100

88 in that collective producers appear to be more technically and allocatively efficient than individual producers. More important, the results indicate more precisely the nature of HARC failure. The main problem appears attributable more to managerial problems (Jensen and Meckling 1979) than to the popular notion of "shirking" (Alchian and Demsetz, 1972). Finally, individual producers are more allocatively efficient when a standard wage is imputed for free labor, indicating that there is an opportunity cost for labor devoted to individual parcels. Collective allocative efficiency is also significantly higher than allocative efficiency of individual producers when labor is imputed a standard wage a significance that disappears when the wage is not imputed.

PAGE 101

CHAPTER 5 TRADITIONAL AND ADVANCED TECHNOLOGY: A COMPARISON OF BEANS AND MAIZE Although maize is a crop rooted in ancient Central American tradition, and is even a central focus of the Mayan religion, technical aspects of maize production have been modified due to land constraints and accessibility of new technologies. Current production processes of maize are much different than those developed by indigenous civilizations. Machinery is used at various levels tilling, seeding, spraying and, though not in harvesting, usually in shelling. Hybrid seeds, which require supplemental fertilizer and pesticide inputs, as well as advanced storage, are in routine use. Much of the input mix for corn production is directly dependent on input distribution systems, the malfunctioning of which can reduce allocative efficiency. Bean production, on the odier hand, still resembles the processes developed centuries ago and uses relatively few inputs. The principle inputs are land, labor and seeds, most of which are available within the family farm. Beans are usually planted in postrera vith a long wooden stick called a barreta between rows of drying corn plants. Corn stalks serve as poles on which bean vines can be strung to dry and to enhance photosynthesis. Bean production also restores nitrogen to the soil which is valuable for the next year's corn production. Bsyond being agronomic complements, maize and beans are nutritional complements. They supply more protein when consumed together than the sum of their individual protein contents. Maize and beans form the basic staples of the Honduran diet. This chapter presents production function estimates of beans parallel to those of maize in the previous chapter. Technical and allocative efficiencies for bean production are also 89

PAGE 102

90 presented. Unlike in the case of maize, beans are produced exclusively on individual parcels; no comparisons can be made with respect to collective production. The main purpose of this chapter is to compare the results of bean production efficiencies with those of maize to draw policy insights relevant to technology and input markets. It is hypoUiesized that technical efficiencies will be higher for beans than for maize because bean production technologies are more traditional relative to those of maize. It is further hypothesized diat allocative efficiencies will be higher for beans than for maize because bean production relies on fewer inputs that do not depend on input markets. Beans: A Traditional Crop The Cobb-Douglas form, which was used in the case of maize, is also used for beans. The general form of the Cobb-Douglas is: i=l i=l where q is output, A is a given level of technology which "shifts" the function in response to technological changes, x, represents the set of i = l...n inputs, h^ represents human capital and the fiiS and a|S are the corresponding coefficients. As in the case of maize, the Cobb-Douglas form is estimated usmg cross-sectional bean data (4.26). The simpler technology employed in bean production is reflected in the production function:

PAGE 103

91 (5.2) InBeans = where: Variable Technology (A) Land Labor Seed Paraiso Region Lecture Publication Lectureaid Lecturepub /3„ + ^^inLand + ^.InLabor + ^^XnSeed + ^J^aralso Region + ^^Lecture Only + ^J^ublication Only + ^jLecture and Visual Aids + ff^cture aiui Publication + e Coefficient Constant Land measured in manzanas' Labor measured in work days Seed measured in pounds = 1 if producer is from the region of El Parafso, 0 otherwise. = 1 if group received extension lectures without additional teaching aids, 0 otherwise = 1 if group received printed extension publications and no personal lecture, 0 otherwise = 1 if group received lectures accompanied by electronic visual aids, 0 otherwise = 1 if group received both lectures and printed extension publications, 0 otherwise * Manzana = 0.705 hectares The variables for beans are similar to those for maize. ^„ is the constant term on the technology variable A in 5.2. Land, Labor and Seed represent continuous variable inputs and thus are in log form. All extension techniques are included in the model as dummy variables. Average and frontier functions are displayed on the following three tables (5.1 5.3). Coefficients on the physical inputs of land, labor and seed are positive and significant for average and frontier estimations. The difference between the frontier and the average functions is less apparent for beans than for maize; aja^. is insignificant. Bean cycles are unpredictable in Honduras, influenced significantly by drought and floods and hence much more encumbered by risk and uncertainty than maize.

PAGE 104

92 A notable distinction between maize and bean functions appears in the relative magnitudes of the labor and land coeftlcients. The elasticity of labor with respect to output is much higher in the case of beans than that of land. Beans are grown in postrera when seasonal demands for coffee harvesting are high. The relatively high labor elasticity value may also be explained by the risky nature of bean production. Farmers expressed the challenges associated with growing beans brought on by erratic weather from one year to the next. The year the data were gathered, rainfall was considered fair-to-good. Risk-averse farmers may have sought alternative forms of employment because annual returns to labor in bean production have been, on average, much lower. The extension variables demonstrate the same pattern that emerged in the technology adoption logit regression in Chapter three. All the extension coefficients that involve personal contact are positive and significant at the five percent level for the average function. However, Publication is not significant in any regression, and Lecturepub is not significant in the frontier functions. These results are contrary to those in the case of maize. Technical and allocative efficiencies for beans are displayed on Table 5.4. All efficiency scores are given by HARC as averages of individual producers.

PAGE 105

93 Table 5.1 Average function: Beans ordinary least squares regression Variable Coefficient Std. Error t-ratio Prob 1 1 1 > X Constant -1.836 0.410 -4.473 0.000 Land 0.165 0.085 1.935 0.053 Labor 0.624 0.083 7.476 0.000 Seed 0.239 0.092 2.586 0.010 Lecture 0.544 0.110 4.947 0.000 Publication 0.238 0.132 1.802 0.072 Lecturepub 0.229 0.110 2.093 0.036 Lectureaid 0.457 0.088 5.211 0.000 Paraiso Region -0. 142 0.090 -1.582 0.114 Observations: R-squared: F[ 8, 169]: Log-likelihood: Amemiya Pr. Criter. 178 0.642 Adjusted R-squared: 37.830 -103.018 Restr.(fl=0) Log-1: 1.259 Akaike Info.Crit.: 0.625 -194.359 0.206

PAGE 106

94 Table 5.2 Frontier function: Beans maximum likelihood estimates half normal distribution Variable Coefficient Std. Error t-ratio Probit! >x Constant -L331 U.Oo4 U.UjZ Land 0.190 n f\oo O.Ooo o 1 *n Z.lT) 1 n (Yx 1 Labor 0.616 0.079 n non l.lol n C\C\C\ u.uuu Seed 0.226 0.100 2.250 0.024 Lecture 0.493 0.123 4.013 0.000 Publication 0.175 0.174 1.007 0.314 Lecturepub 0.181 0.106 1.708 0.088 Lectureaid 0.429 0.083 5.157 0.000 Paraiso Region -0.149 0.091 -1.638 0.101 -0.253 4.687 -0.054 0.957 OU/ffV 1.276 0.980 1.302 0.193 V ff^v + a^u 0.534 0.398 1.343 0.179 Log-Likelihood: Variance components: -101.8082 (t2(v) = 0.10864 ff2(u) = 0.17695

PAGE 107

Table 5.3 Frontier function: Beans maximum likelihood estimates exponential distribution Variable Coefficient Std. Error t-ratio Prob 1 1 1 > x -1.452 0.403 -3.601 0.000 Land 0.188 0.087 2.160 0.031 Labor 0.613 0.078 7.834 0.000 Seed 0.224 0.098 2.291 0.022 Lecture 0.503 0.120 4.189 0.000 Publication 0.174 0.162 1.075 0.282 Lecturepub 0.187 0.105 1.7V8 0.075 Lectureaid 0.431 0.082 5.2^-5 0.000 Paraiso Region -0.139 0.089 -1.573 0.116


PAGE 108

c o Q 5 (A O C S3 «2 3 i .-s D > o . . ^ 1 O CO c c o c X I i 3 X >> s CO o C8 E O Z CO X < X 0^ ON o ON ON IT) ON ON ON CT" C <^ S oo i> O ^ 0* iTT ON ON Pi ON ON o ^ o =^ o o ^ o o o ^: o o <^ S o 7^ ^. ^ Z S ^. ^ ^. ^ ^. ^ ^. ^ \o ^ o ^ s r-. p* 00 00 ov g r~. so o\ ^ o 00 NO oo O '-^ (N OO r~ o\ OO O o O ^ ^ 00 g oo g m rNO G> ON ir D 3 c 1> o E D en 5 Pe z < 03 C« c C/3 o O CO O U J -J C/3 -J CO N C CO D O. u CO -J o JO oa CO U a. o N 3 E CO 3 a

PAGE 109

97

PAGE 110

98 General Efficiency Comparisons: Maize and Beans Table 5.5 shows the difference in means for allocative efficiency between a near zero wage for unpaid labor and an imputed standard wage for unpaid labor. As in the case of maize, imputing a wage for unpaid labor increases allocative efficiency for bean farmers. The difference in group means is statistically significant at the .01 level of probability. Table 5.5 Differences in allocative efficiency between an imputed standard wage and zero wage for unpaid labor Half-normal Exponential Maize Difference 0.04 0.05 (9.54)'*** (9.88)Beans Difference 0.08 0.05 (6.85)'" (4.89)' t statistics are in parenthesis "Significant at the .01 probability level

PAGE 111

99 Contrary to maize, allocative efficiencies for beans are statistically and significantly higher than technical efficiencies at the .01 level of probability (Table 5.6). Allocative efficiencies are understandably lower than technical efficiencies in the case of maize because input distribution systems are notably inefficient. Allocative efficiency may be easier to achieve for low input, traditional farming because it requires the same production-mix decisions that have been made for centuries and throughout a farmer's life. Technical efficiency, on the other hand, may vary in conjunction with a farmer's health or age. Table 5.6 Average Technical and Allocative Efficiencies for Maize and Beans Maize Beans Technical Efficiency Half-Normal .76 .67 (.15y (.12) Exponential -77 .79 (.15) (.10) Allocative Efficiency w 0 Half-Normal .24 .89 (.14) (.15) Exponential .30 .87 (.16) (.13) Allocative Efficiency w = 5 Half-Normal .28 .97 (.14) (.04) Exponential .35 .92 U5) C05) Numbers in parentheses are standard deviations

PAGE 112

100 In general, technical efficiencies are comparable to those of maize, lower for the halfnormal but higher for the exponential distribution start (Table 5.7). The markedly low technical efficiency for beans calculated from the frontier of the half-normal distribution is difficult to interpret. It is statistically significantly different from the technical efficiencies for maize. However, it is also statistically significantly different from the technical efficiency for beans calculated from the half-normal distribution. The most notable result is that allocative efficiencies are strikingly higher for beans than for maize. The difference between maize and beans in this regard may be attributed to differences in the number and accessibility of inputs used in the production process. Maize production depends on hybrid seeds, fertilizer, and pesticides that are not always available in the market. Bean production, on the other hand, primarily employs land labor and seeds, all inputs available to households that do not depend on input markets. This underscores the importance of ensuring that inputs are available to farmers as they are encouraged to adopt more advanced technologies. Variation in efficiency is less pronounced in bean production than in maize production. Table 5.6 shows coefficients of variation (CV)^ for maize and bean production. CVs are less in bean production for all variables, but are notably less for allocative efficiency. This is attributed to the simple and traditional technology employed in bean production relative to maize. Most farmers, regardless of age or strength, are cognizant of optimum input mixes in bean production and have access to necessary inputs. Maize technology is more advanced and depends on inputs that are not universally accessible. The variation drops for both crops when the standard wage of five lempiras is imputed, suggesting that producers are accounting for the opportunity cost of free labor (i.e. all are relatively closer to the optimum). 'The coefficient of variation is the standard deviation as a percentage of the mean.

PAGE 113

Table 5.7 Coefficients of variation Maize Beans Technical efficiency Half normal 19.74 17.91 Exponential 19.48 12.66 Allocative efficiency w = 0 Half normal 58.33 16.85 Exponential 53.33 14.94 Allocative efficiency \v = 5 Half normal 50.00 4.12 Exponential 42.86 5.43

PAGE 114

102 Summary This chapter estimated frontier production functions for individu?J bean parcels of HARC members exactly as in the case of maize. An average OLS function and maximum likelihood functions which assumed half-normal and exponential distributions for the truncated error term, u, were estimated. Technical and allocative efficiencies were calculated based on the optimum input-output relationship observed from the estimated stochastic frontier. The most notable observation results from the comparison of maize and bean efficiency measures. Technical efficiencies of bean farmers are comparable to those of maize, but allocative efficiencies are strikingly higher than maize allocative efficiencies. Unlike in the case of maize, allocative efficiencies of beans are higher than bean technical efficiencies. Moreover, there is less variation in bean allocative efficiencies especially when the standard wage is imputed. The contrasts between maize and bean production may be explained by the fact that bean production involves a simpler, traditional technology for which inputs are fewer in number and more readily available than maize production. Allocative efficiencies for beans are understandably higher because the primary inputs of beans, labor, land and seeds, are easily obtained by HARC households. Allocative efficiencies are also subject to less variation because traditional technologies depend on knowledge learned early in a farmer's life and, unlike technical efficiencies, are less dependent on physical or mental stamina. The low and variable allocative efficiency scores for maize underscore the importance of establishing efficient input distribution systems.

PAGE 115

CHAPTER 6 THE INFLUENCE OF HUMAN AND SOCIAL CAPITAL ON TECHNICAL AND ALLOCATIVE EFFICIENCY Introduction The development of modern technologies necessitates continuous investments in human and social capital so that workers and enterprises can properly employ new production techniques. The mere adoption of new technologies (examined in Chapter 3) does not ensure that potential efficiency will be attained. The optimum investment mix in human and social capital maximizes the technical and allocative efficiencies presented in the previous two chapters. HARCs provide efficient structures through which investments can be made in the mental, physical and social capacity of H ARC members. Among of the main objectives behind the establishment of HARCs are the education and social integration of campesinos who historically have been socially and politically disenfranchised from die educated and economically dominant ruling clais. Human capital, and more recently social capital, are recognized in mainstream economic analysis as integral components to optimally operating economies. Cooperatives are often the first form of secular organization to which campesinos, who are strongly independent and often distrusting, belong. Cooperatives serve as a means through which extension beneficiaries can be concentrated and hence magnify the impact of extension efforts. Perhaps more imponant, evidence indicates (Martin and Taylor, 1995) that cooperatives reenforce extension efforts through members' learning from each other. The importance of human and social capital to HARCs' broad social 103

PAGE 116

104 mission is obvious, but beyond tlie scope of this dissertation. The purpose of this chapter is to focus on how human and social capital influence the efficiency of HARC production. Theoretical Underpinnings of Human Capital Almost all the classical economists recognized the general value of human capital to society. Adam Smith acknowledged in "The Wealth of Nations" that improvements in the skill, dexterity and judgement of workers were fundamental to economic progress and increasing economic welfare. However, the theory of human capital was merely suggested and not developed in any meaningful sense by the classical economists. The concept of human capital appeared sporadically throughout this century'. But the theory of human capital was not established, according to Blaug (1976), until it was given special precedence in a 1962 supplemental volume of the Journal of Political Economy. Gary Becker published an article which served as an analytical structure for much later work in human capital. The issue included an article by Theodore Schultz, who argued that laborers actually operate as capitalists, investing in knowledge and skills that yield commensurate returns in the labor market. He expanded the theoretical framework, suggesting human capital be broadened to include not just men or improved factors, but the economic productivity of education. Concurring with the findings of Edward Dennison (1962), Schultz offered human capital theory as an explanation for the large residuals or "technical change" that resulted from economic growth. A seminal article by Jacob Mincer, which recognized that learning occurs in both schooling and on-the-job training, was also published in this same issue. 'See, for example, Marshall (1920), Dublin and Lotka (1930), J.R. Walsh (1935) and Friedman and Kuznets (1945).

PAGE 117

105 Human capital research diverged onto two complementary paths. The first focused on the relation between human capital and earnings distribution-. The second provided a theoretical groundwork for human capital which could identify the sources of productivity and growth (Griliches. 1964; T.W. Schuitz, 1963). It has been the orientation of human capital studies in LDC's and is the focus of this chapter. Human capital investments generate returns to personal utility that are not reflected in enhanced productivity or increased earnings. A good liberal arts education, for example, not only prepares people to compete and communicate more effectively in industry, it augments their capacity to appreciate and discriminate between a broader range of consumption items. Healthy, educated people have stronger physical senses and mental faculties to sense and esteem the aesthetic aspects of goods and services. Increases in utility resulting from human capital investments are unaccounted in many human capital studies, rendering empirical estimates of human capital benefits based solely on earnings or productivity artificially low. Empirical Applications of Human Capital Components The theoretical bases of how human capital affects productivity is intuitive. Empirically, however, it is difficult to identify adequate measures which represent tne authentic underlying factors. Years of schooling is easily measured, but education is not. Extension and health improve productivity, but both are related to several other factors that cloud measurement. Empirical measures thus provide no reliable precise elasticities but they do often provide valuable directional results against which the theory may be examined. Knowledge is by far the most important and conventional human capital factor. People acquire and store knowledge through various forms of mental activity. Workers may approach 'See Mincer (1958); Becker and Chiswick (1966); and Blaug (1976).

PAGE 118

106 a task equipped with formal education, parental and peer guidance, specific training, and experience. It is virtually impossible to sort out all the factors that influence the knowledge base of any individual, much less the aggregate. In order to facilitate empirical testing, knowledge is considered to emanate from three identifiable factors: education, training and experience. Two other human capital factors not directly associated with knowledge, healti and religion, will also be considered. All five factors, by nature of their being imbedded in each human being, simultaneously influence each other. Education General education refers to basic schooling in the "three R's. " Distinct from the learning acquired through job training and work experience, education expands mental capacity and indirectly contributes to the capital stock of laborers, managers and entrepreneurs (Sheffield, 1975). Education enhances the cognitive ability of workers and dei;ision makers to better understand and cope with complex relations that underlie production processes. Often considered as a useful precursor to direct job training, general education imbues the human input factor with critical thinking capacity; heightening theproductivity of direct training and work experience. Simple skills such as literacy and numeracy are essential to any training program, but even more advanced education in the arts and sciences are generally considered useful to communication and decision making. Farmers better employ advanced technologies if they understand fundamental principles upon which those technologies are developed. In a continuously modernizing economy, where agricultural research alters production techniques and non-farm firms provide inputs or input substitutes previously produced locally, education becomes increasingly important to capacitate farmers to efficiently exploit new

PAGE 119

107 technologies and opportunities. Furthermore, sufficient time is rarely if ever available to reallocate resources so that a new level of equilibrium is reached. Production environments change so rapidly that new changes occur before the old ones can be fiilly adopted. Thus, the equilibrated "full efficiency" state achieved in traditional agriculture (Schultz, 1964) almost always lies beyond the reach of farm people, emphasizing the need to augment the cognitive ability to deal with disequilibria. Empirical evidence of Schultz and others corroborates the link between education and productivity.' Virtually ail human capital studies measure education in terms of schooling. The assumption is necessarily made, usually implicitly, that education is homogeneous. This may be reasonable for education within regions, but it is less so for studies which cross regional, and especially cultural, boundaries. A further problem is that non-measurable factors closely correlated with education may be the root cause of efficiency improvements, but remain unrecognized by models which cannot include them. For example, better educated farmers are more likely to have greater exposure to the evolving social and economic apparatuses of the developed sector than poorly educated farmers, an exposure that could bear a stronger influence on productivity than education. Learning that occurs at home or in collective work arrangements and innate ability are likewise unaccounted for. "Screening" is another problem associated with analyzing the economic effects of education. Are individuals hired for better paying jobs because their education has made them more productive, or does their degree act as a bogus signal of productivity to employers? Fortunately, members of Honduran agrarian reform cooperatives arguably form a homogeneous group with respect to education, social networking and outside opportunities. ' See Jamison and Lau (1982) for a good review. See also Khaldi (1975); Pudasaini (1983); T.W. Schultz (1964); Wu (1977); Huffman (1974); Jamison and Lockheed (1987); Kalirajan and Shand (1985); and Behrman and Wolf, (1984).

PAGE 120

108 Training and experience Training and woric experience, apart from education, are imporant means of improving the broadly defined human capital component of knowledge. Several studies have examined the impact of training on productivity. Recent studies, especially from developed countries suggest a high pay-off from worker training.* Substitutability of training programs with general education is a poorly understood but important relationship, given the relative low cost of training programs. Huffman (1977) and Jamison and Lau (1982) report evidence which suggests that education can be substituted for training or extension. However, complementarity between education aid training may also be observed if the effectiveness of various extension and training methods were examined for farmers of differing educational levels. Agricultural training in LDC's is usually designed to complement farmers' paucity of knowledge. The benefit of general education may not materialize in training programs because those programs were designed for uneducated farmers. Depth of i.istruction and training materials may be unnecessarily uniform and cost ineffective because programs are designed to reach all farmers. The question arises, is it possible to target farmers by demographic characteristics in order to optimize productivity through training? It would be interesting to examine die effects of a variety of training programs across farmers of varying educational levels, although no such empirical studies have been done. The data set gathered for this study includes both variables needed to advance this research, years of schooling and a variety of extension techniques. The Integrated Pest ^For developing countries see Jamison and Lau (1982); Shapiro and Miiller (1977); Behrman et al. (1985); and Kalirajan and Shand (1985). For developed countries see Mincer (1989); and Vaughan (1989).

PAGE 121

109 Management Program of Honduras (Spanish acronym MIPH) offered one of four different types of extension assistance to each HARC in the study. One HARC received lectures only, one lectures and publications, one lectures accompanied by electronic visual aids, and one group of HARCs received IPM publications without lectures. If education enhances extension it may be detected through the influence of varying extension types on different educational levels. Health Health is perhaps the most important human capital factor which influences productivity both directly and, through its effect on education, indirectly. Malnourished laborers do not have the caloric fuel necessary for much of the heavy work of traditional agricultural. Moreover, malnourishment and lack of sanitation make people more vulnerable tc disease, which in turn lowers productivity and the subsistence output on which many depend. The cycle is selfperpetuating. A theoretical link between health and productivity was formulated in Liebenstein's efficiency wage hypothesis (1957). The essence of this hypothesis is that, under certain circumstances which pertain to LDC's, increasing food consumption improves productivity. Starting at a level of daily energy intake that merely covers basal metabolism, productivity increases at an increasing rate, implying that very low wages induce labor deficits because caloric intake is so low. Empirically, there are many problems in measuring the effect of nutrition on productivity or work performance. Specialists disagree on what constitutes the necessary level of consumption because individual metabolisms vary. There is also a problem in measuring the energy required to carry out certain tasks. In LDC's, work is frequently seasonal and can vary a great deal in intensity. There are also lags in noticing the result of nutritional investments. Severely

PAGE 122

110 malnourished people would show dramatic improvement almost immediately, but less undernourished people who may benefit from increasing the protein or vitamin content of their diet show positive results at a slower rate. The most serious malnutrition problems occur among people who become physically limited for life because of early childhood malnutrition. For these and other reasons, empirical studies conducted thus far have yielded little proof to support Liebenstein's theory.' Most assumed measures of current physical stature, such as height and weight or limb circumference, reflect nutritional background.* Religion Economics is fundamentally agnostic. Neoclassical theory goes to great lengths to avoid interjecting even minor personal biases into its analyses. Religion is largely absent from economic, or even human capital, theory. Nonetheless, religion plays an integral role in human capital development and has been the focus of recent research. Religion enhances the explanatory power of earnings variation models formerly constrained to differences in race and sex. Religion bears strong influence on family values, skills, endowments, goals and culture, which in turn are determinants of earnings and productivity. Tomes (1985) distinguishes "religious capital" from human capital in that it includes "ethical and moral codes of behavior governing consumption, the allocation of time and interpersonal relationships." He considers religious capital to be a vital foundation for the standard human capital factors which contribute to productivity and earnings power. Most ' Bliss and Stern (1978); Immink et al. (1982); Behrman et al. (1985); and Audibert (1985). *Immink et al. (1982); Mook and Leslie, (1986); and Jamison, (1986)

PAGE 123

Ill empirical studies regarding religion have examined how religion influences earnings differentials and fertility.' Little dieoretical or empirical research has been conducted on religious capital in LDC's. Religion played a crucial role in the history of Honduras. The CaAolic Church, still the predominant religious organization, underwent dramatic theological changes in the early 1960s that have forced a reassessment of the traditional social structures the Church helped establish in Latin America. "Liberation theology" holds that wealth should be shared more equitably, particularly when extremes of wealth and poverty exist. The church encouraged people to form "base communities" and delegated more responsibility to them to practice their faith. Calling for wealth redistribution and encouraging small community enterprises obviously has special relevance for agrarian reform cooperatives. A number of Protestant "Evangelical" churches have also grown in Central America over the last 20 years. They espouse no unified theology, but generally hold that people should bare sufferings of this world for greater contentment in the afterlife. Unlike liberation theology, Evangelical churches take few political stands, but they do discourage drinking that in excess dampers productivity. Social Capital The term "social capital" has had various meanings throughout recent history and across regions. In communist societies, social capital referred to communally owned plant and equipment. Social capital has also been considered natural resources upon which future generations depend (Toman 1994), as public housing (Spence, 1993), o: as the employment of unproductive labor as an overhead cost (Smith, 1993). 'See Brenner and Kiefer (1981); Becker and Tomes (1976, 1983 and 1984); Chiswick (1983); and Meng and Sentance (1984).

PAGE 124

112 A growing body of literature in sociology and political science is examining various aspects of what is broadly considered "social capital." The concept, as measured by civic involvement and satisfaction with government, has special relevance for cooperatives. Putnam (1993) attributes higher incomes in northern Italy, which also has a successful record of cooperative development, to a higher level of social capital. Social strictures in northern Italy are more "horizontal" than in southern Italy, which maintains a preference for hierarchical system. Helliwell and Putnam (1995) further fmd that social capital has a positive influence on equilibrating income levels and on the rate of percapita income convergence. These results have important implications for HARCs. which hold increases in, and the eqi.ilibration of, income as fundamental goals. HARCs are prominent instruments for the Honduran government in making the same attempt Italy made, namely to transfer technology and make physical and human capital investments in predominantly rural areas. Social capital is worthy of examination to the extent that it is complementary to conventional forms of investment. Cable and Fitzroy (1980) consider social interaction a neglected area of study in labor managed firms. The lack of incentives that property rights theory (Furubotn and Pejovich, 1972) and internal organization theory (Alchian and Demsetz, 1972) contend characterizes cooperatives fails to take into account social dynamics that may overcome individual incentive problems. This chapter examines social capital through interaction between religion, games and music as demonstrated by mcreases technical and allocative efficiency.

PAGE 125

113 Em pirical Model and Data The tables below display differences in mean efficiencies for HARC members grouped by several demographic categories. Corresponding t statistics evaluating the significance of the means differences are presented below each difference. The t test is: X, Xj (6.1) t = -i where X; represents the mean efficiency estimate of group i, and (6.2) ^ where n^ is the number of observations for group i and S'i is the sample variance of group i. If t > t„.ni +n2-2 there is a "significant" difference between the groups. In other words, there is less than an a percent chance that a positive difference between the means is attributed to sampling variability rather than to a difference in population means. Tables 6. 1 6.5 and 6.6 6. 10 below display the differences in means for various groups of maize and bean farmers respectively. Many of the variables fall into one of two mutually exclusive categories (e.g. a respondent was either literate or not). For order categorical variables, comparisons were made for groups above and below the average (e.g. efficiency scores were compared for the group that had more than the average number cf school years with the group that had less than the average number of school years).

PAGE 126

114 Maize Table 6. 1 shows efficiency differences based on various demographic characteristics. The most striking result emerges in Proportion output sold. Farmers vho market on average proportionately more output are more technically and allocatively efficient than farmers who market less. This result indicates that "hillside farmer" programs, which are designed to improve efficiency and environmental protection and which are popular among policy makers, will have limited impact because most hillside farmers are near-subsistence producers. By contrast, producers who supplement their income with Outside work register a significantly positive difference in allocative efficiency, but only when the full wage is imputed. The Labor/land ratio differences are negative and statistically significant for allocative efficiency when the full wage rate is imputed, indicating that farmers with relatively little personal land have a low opportunity cost of labor. The only other significant difference in terms of technical efficiency on Table 6.1 is observed between those above the age of 50 and their younger counterparts. The negative sign indicates that older men tend to be less efficient. Given the rigors of fieldwork, one can understand how younger men would hold an advantage at hard labor. Allocative efficiency is also greater for farmers under the age of 30. That the negative values for older men and the positive values for younger men reduce in significance when the standard wage is imputed suggests that older men have a higher opportunity cost for their labor in individual production than younger men. A positive and statistically significant difference is demonstrated between allocative efficiencies of farmers who are more Literate and have more Primary schooling and Cognitive

PAGE 127

115 capacity,^ indicating that more intelligent, better educated farmers are more capable of adjusting input mixes to take advantage of relative input price differences. Farmers with larger families are less allocatively efficient than farmers with smaller families, but only when the wage for free labor is zero. Larger families are not significantly less allocatively efficient when the standard wage is imputed, indicating that larger families have a higher opportunity cost for labor than smaller families. This result may seem curious in light of the conventional belief that rural children in developing societies are used extensively in farm production. However, children in Honduras now attend school on a routine basis, which divests production of labor that is necessarily supplanted by capital to meet increasing family food needs. Table 6.2 shows the differences in efficiency over varying health factors. Larger farmers are more technically and allocatively efficient than smaller farmers, as demonstrated by Weight, Arm circumference and Leg circumference. It stands to reason that larger farmers would be more technically efficient, given the arduous labor involved. The fact that larger farmers are also more allocatively efficient as well may stem from the fact that smaller farmers often suffered malnourishment at a young age, which impaired mental capacity. Eye sight is the only other factor to register a significant result, for allocative efficiency. It too nay be correlated with greater mental faculty. Table 6.3 displays efficiency differences across various social capital factors. Religion is a prominent form of social capital, the primary initiative behind the establishment of many HARCs and a congregating force in rural Honduras. Catholics appear statistically and significantly more allocatively efficient than noncatholics, perhaps because base communities offer a forum for interaction. Evangelicals are statistically and significantly less technically and 'As measured by Raven's colored matrices test.

PAGE 128

116 allocatively efficient than nonevangellcals. Their allocative inefficiencies become stronger and more significant when the standard wage is imputed. Perhaps the most interesting results are the efficiency differences as they relate to various forms of extension methods (Table 6.4). Those who received a lecture unaccompanied by pamphlets or visual aids {Lecture only) register higher technical and allocative efficiencies. All other significant results are negative, suggesting that supplementary teaching aids are counterproductive tools in extension efforts. The maize technical efficiency differences on extension methods are perplexing in light of the fact that the only extension variable that is not positive and significant' in all three maize production functions is Lecture. The difference could be explained by the fact diat the control group, which was excluded from the production function regressions to avoid a singular matrix, is now included in the group against which Lecture only is compared. Control group shows negative and significant allocative efficiencies, which indicates that those HARCs that did not receive any extension assistance were significantly less adept at adjusting input mixes to an optimum. Finally, Table 6.5 shows efficiency differences for farmers with above average levels of experience. Experience with herbicides and insecticides appears most helpful in improving efficiency. To some extent, no single input should be considered apart from the others in that hybrid seeds require fertilizer and pesticides to maximize potential; it is a package. However, pesticides constitute the most complicated component of the package. Improper pesticide application could override benefits derived from hybrids and fertilizer, thus permitting the only difference in experience to be manifested in pesticides. 'At the 0.1 level of significance.

PAGE 129

a: CO o o j= w in 3 O •T3 B C O 1) Q. re c g II UJ UJ ^ 4) > re o c If-no t~o oo o 00 00 5 in rsi o Ha o o d d d >i o e 2> X UJ re ^ re re o re " E re o u c x> c c X UJ O) ,1> 're 2 117 22 -Jo d ^ _ ^ o =<3 <3 3 C Q ^ o d 0^ 00 ^ O "1 d CN * • * o O O o d d d (N (N O m O o ~ CN m o « * CN CN o\ o ON (N d d d d 00 in oo 1/1 >n oo o o o o d o o d d d d d 8 ^. d o 00 o d c 3 ^ "re > 3^ o I §3 Q a ^ to ^ CN — ^ Z' ^ — o T d — o o o d •a2i > c <3 as; in E; o 9 d CN O CN d ° CN CN ON ° d o NO O (N 6. a: i5 1^ i 1) ,1^ « a Q «

PAGE 130

118 3 C S O o N E s "o •s 3 O C o a. O Hi (£ M U ^ 13 3 I c ^ < DC c 1) c >^ o £| = II W ^ 15 > 'J c £ T3 'J c C c c X " 'J — 1) 1) c N o q d ON o q d <^ o d in 00 O 00 ^ in o =^ o 2^ o o o IT) ON s:^ ON o °° m 00 o d 00 • O ON d -r g 00 ON O NO ^ d o D -J C ^ fc ,0) ^ ^ ••3 I « III go."" q 0 w r! c c 5 * = §1.2? ,

PAGE 131

E a. CO 0. XI H ^ 15 E o B o o > CO o o 'J c 4= U H E 03 C <1> c o c X «5 o c it: S :2 o o t^. o o o c O O ON O ro o o 8 IT) S 2 r4 O — O O O «^ <^ *_. (N o o 00 O O ON d — o oo o o O — CM Op O ON d — oo : O d oo o d oo o ~ IT) O d oo • o d oo r~ * 00 o t-~ d — O ON d — 5J s *^ (— — _ .!> ^ ^ .35 C :^ -fc-I; 03 is -J; o3 0) :^ <~ C: 03 ^ > •> -s: .=« Q w .2? Q ~ ^ .b ^ ^ vr > C •;r > c Q ^ Q ^ X "-3 -1' > ^ 9 Cm .2? 0C--

PAGE 132

120 c •t: ^ W ^ 15 « E 'i g < X (-<1 ^ o d — oo • o o o o«! Q — o p 22 ^ o g II rex) it: W ^ 15 > O o < X cN in o d — OS o o o in o d o o • o d fN in o c o (/3 c c c X X fN ^ o d d fN o d 00 d in o ^ d -r d — 15. 9 ? 1^ 9 9 oo • OO — Ill III St: 0) 0) !j 1) ^ ^ 5l 73 > =0 i) •J c 0) '-) c C i) 0) _3 D Ui ,a> Isc > 'o (5 ^ Q 1> -J 1) -J c 3 ^ t5> -a ,« ^ « -t: 03 ? -t: cs a ^ I a «

PAGE 133

c a> c c c X W ^ 15 E _> CO o o >> o -J c II > E CO o i: o c w "33 o o Ec c c c X UJ i) c a> — o — o \d oo m q Tt; O iri ON q q O Tt 3 5 o O Q c: Q -4 3 bI* D CO o V. 5 Q « 121 1 > 111 III q 5

PAGE 134

122 D N 'to E X U TO c (U c o ^ >< g II Hi 'J 1> W ^ TO TO 'J 5 II c TO TO UJ > TO ~ TO < X c c o £ X >> c o £ u "to o 'c 'J u IJLI TO 1/5 i) '-) c .2> Q 0) (N o o — o oo O CN as * • in o o OO o o O S ii 7, 5111 lit B ^ a w C

PAGE 135

123 Beans Results of how demographic and human capital variables influence efficiency are less significant for beans than for maize. The lack of significance may be attributed to less variation in bean efficiencies (see Table 5.7), and to the fact that the number of bean observations (178) is half the number of maize observations (387). Table 6.6 shows differences in the efficiency of bean production across various personal and household characteristics. The only variable to show a significant difference in technical efficiency is Proportion output sold. Producers who sell more of their output are also more allocatively efficient when the standard wage is imputed. The levels of significance are strong. These results corroborate those of maize: commercial farmers show a higher propensity to operate more efficiently. Producers with high labor to land ratios are also more technically efficient and allocatively efficient when unpaid labor receives no wage. Not surprisingly, and similar to the case of maize, the difference in allocative efficiency becomes negative for farmers with high labor/land ratios when the standard wage is imputed. Literate farmers appear more allocatively efficient than illiterate farmers, but schooling and cognitive capacity show little influence. Unlike the case of maize (Table 5.3), bean production is still based primarily on a traditional technology that, holding to Theodore Schultz's finding in Transforming Traditional Agriculture (1964), cannot be improved by education alone. Older producers are more allocatively efficient when the wage is near zero, but significandy less so when the standard wage is imputed, indicating older farmers see less opportunity cost in devoting free labor to bean production. Similar to the case of maize, allocative efficiency is less for farmers with larger than average size families, as indicated by the negative sign.

PAGE 136

124 Health factors show only a few significant differences in efficiency (Table 5.10). Bean production requires more labor than maize, but usually over half of it is devoted to harvesting and shelling. Hired labor is often brought in for those tasks, thus clouding the effect of health influences. Social capital factors (Table 5. 1 1) again indicate that Catholics are more allocatively efficient than non-Catholics. Differences in efficiency attributed to extension methods (Table 5.12) do not counter the results obtained for maize production. Lecture only and Lecture and publication register a positive influence on allocative efficiency at a near zero wage rate. Control group is again shown to be less technically and allocatively efficient than groups that received some form of extension training.

PAGE 137

73 o n> u 'J O c c o o. CO >> 1/^ c c II U W ^ 15 E CJ O C u c c >^ O g II W W ^ 15 I ^ o o < X >1 c C8 o 5^ X CT3 X 'J C eg u 02 125 is d s ^. d o 8 ^ oo o ON oo d o O d o ~ 2: c^ o «=! o tN o oo o . o 3 5 Q 5 -I ,1> 5 : is d ° is d 9 d ^ oo o d = d 9 o ^ o ^ d 9 Tt NO O oo d — o . d 9 ON 8 ^ d 9 10 <^ q 5 oo • S 2 ^ ^ 9 9 d " o ;r § oo o 8 ^ d o d 9 m O d V S ^ d o — d O 2 ^ _ ^ O <^ d o o o <^ VO 5 ^ " d 9 5 ^ 9 9 o ^ ^ S. o o d 9 d 9 00 * * * 00 • NO p 1 vo : (N in O CNt d CO is d o 8 <^ d « 00 L. d <^ o d On o Q J Q ^ ^ U ^ o 3 D 13 > 5 ao s; •2 c 0) NO 00 m m o 9 r^o : fN ~ OO o • d — "I J3 3 13 >

PAGE 138

3 C o X3 O OS o 12 o j= a> 3 O J= C CO 15 c o U § C c c X 3? B 03 =n 03 < a: D 6X1 >. o c. I " ^ W ^ 15 > 03 CJ o i: =: 03 < X C c G W _ P 03 D X 0) c c 03 CQ o o d =5 00 O 00 p d o o d oo o ~ o O OO o o oo !2 b^ o ^ o d o d ° O o o — d O o d ^ I I Si ^ i: 15 g Q t t I 'J c i) 5 5 -^E T d 9 tt * * CO o Tld d IT) o o o (N O CN o in 1 d d 126 111 111 ill S « ! C S « ~: = C/l t/2 .

PAGE 139

c C i ^ W ^ 15 > O c c S3 D C § II 4: eg ^ 7= > -J o i: o c c c c c X m 73 E 73 O c 1) Q c u 03 O o o 00 IP 8 H o ^ 0 00 o o 00 o o ON o o o d 2: o H o o <^ o o IT) d u 'J i ^ ,i> ^ ^ > -s; ' C > O ^ 9 V d 5 ^ 00 _ o 2^ o ^ d 1^ d 9 ^ — o is d :£! i^ g d 9 — ^ P d o o ^ o g n 8 d o 1^ 0 is d o •J c .2? Q « a: D 3 Q ^ S 0) •J 2i ^ •t: ^ Q t 127 go."* J I i c 5 " i s .H

PAGE 140

V3 c 73 a. (J o 00 sd H _1J ^ ^ 15 i) > c c >> O CJ g II UJ 'G 0) C CO W ^ 15 I i c i: 73 < X o c u 15 'c x: o H c. X 73 X 0) c c 73 1) 03 o o — — ON o rd — o . d o S 9 u c g ^ o — O — O O *_ o 9 5^ oo o ^ o . P O 2 O ON O in 8 9 D ^ -c: ,i> — 5 t c t: 73 o 8 o o 8 o o 8 o o o d o 8 o 1) -J c 73 2 c^ 22 ON o r4 o in IT) 2j o p d 3 Q U oa o § s P 9 ON m On 9 d o c 1) _ -3 ^ ,i> ^ 128 III e i. c 5 « -I w O

PAGE 141

V5 c 1) c >< C D C i II ^ •n 1) W ^ ^ > ^ o i: < I >. o g II W ^ 15 > 58 C 4> C O c X o B 'c •J u r3 V3 4) •J C 1) ?3 1) 03 IP d ^ 00 8 d 9 1^ d o ON o d d — o =° 2 o> o ^ — ON d — ^ — 8 ^ in O o ON 8 o d d — 8 «=. d 2 ^ ON o — o i£2 oo o O o d O o — 9 d o ON I ^ D •J c S *^ Si, D -) C 53 l,1) ^ ,u ^ C So 5 £ ^ >3 a 'J c 129 111 111 ~ o ^ ^ -S !^ « " a " w c 5 « a ?i 2 ts H: -s .2? i«

PAGE 142

(J c c c c. X it; ^ 15 E 0) > o g II W rbX) it CO "J ^ 15 O < X u c c c c 'o 15 X = I c :0 D CO o . O r<^ ° d O CO O o IT) O 0 7^ Cs i> -J c D 00 O o o o o o m * * in lO ON >n p q o p d I d d 1 d as d ON P T o ^ o 00 CN — o o — o _ o d d ' d ON CO On NO s d 'J C :0 > £: ^ !^ _ £: 13 5 r =5 S ^ 130 111 go,"* C 5 « « S ,a ,H c Ir: Pn °* to ,

PAGE 143

131 Training and Education: Complements or Substitutes? Maximizing the knowledge of farmers in a target area requires that scarce extension and educational resources be distributed to the activity with the highest opportunity cost. Evidence of substitutability between education and training (Huffman, 1977 and Jamison and Lau, 1982) is based on homogeneous training programs. The four extension programs designed and implemented by MIPH may have different effects on farmers of varying educational backgrounds. Publications, for example, are not very useful to illiterate farmers. Visual aids may enhance the instructional capacity of a lecture for all educational levels, or they may distract from the substance of the lecture particularly for farmers unaccustomed to them. Tables 6.11 6.15 (Maize) and 6.16 6.20 (Beans) show technical and allocative efficiency differences for different extension techniques at increasingly higher levels of education. Maize farmers belonging to HARCs that received a lecture only (Table 6.11) were more allocatively efficient than other farmers at all levels of education, suggesting that training can be substituted for general education to improve efficiency. However, all the efficiency differences in the control group (Table 6.15) are negative and significant for farmers that have at least one year of primary education. This indicates that complementarity exists between education and training. Complementarity between education and extension can be seen at higher levels of education for other extension types. Perhaps the most interesting result is that farmers with higher levels of education demonstrate statistically significant greater efficiency in groups that received publications only for both maize and beans (Tables 6.12 and 6.17). Publications are a significantly cheaper means of information diffusion than personal contact with extension agents. In the case of maize, lectures and visual aids differences (Table 6.14) are negative and statistically significant with no primary schooling, but become positive and statistically significant

PAGE 144

132 at higher levels of education. Similarly, for lectures and publications (Tables 6.13), differences are negative and statistically significant at low levels of education, but are not significantly different at higher levels of education. These results suggest that complementarity between extension aids and education exists at higher levels of education.

PAGE 145

c o D -J > c re 'J St 1) >. o c o c .2 > ;§ < o o o (A B 'J c 3i G D > C < O c o c o 'c o D re 1i ^ re c re a: c X re i> c o c X re 1125 » * * oo 1092 tt ft * CN CN 1055 * d o ^' o rn OO O ON o OO o o OO r<-i OO m (N O OS O 00 ON q q ON . — m — OO O CN CN O o ON in (N O o _• CN l~ O o CN O o OO r(N o d in ^ d CN « CN # * 00 * CN * • ON • • o CN • NO OO OO OO 00 ON m CN q o in o q o d d d CN d * * • in 0971 * # « 0758 tt « * m 1690 * 1862 tt tt tt TjrCO o m o o o CN in On d CN NO in ^ O ON in m o ON =<1 •S C ,1> ^5 r 60 ^ i re S |C O 60 •S o o re 'in > ^ s 2f D 60 S 3 ^ ft 00 •S o .5 IS 133 111 111 « g i2 i.£P so

PAGE 146

c UJ > 60 _S M < o o E a. o c 1) — UJ <1> o > •= < o E o B UJ 5^ o B 1) UJ 1) 1) C X UJ :t3 E B E 0) B C. X UJ o B s 1) B X UJ c B in 0240 .0259 .0149 o 1 9 9 q d d ' in On O — 9 d o (N o o _ m O ON oo o , o ^ C o 03 0. 0\ — ON O — r<^ o q O 0\ n O 00 00 O (N O d 1 d 1 d 1 1* 9* o «^ o — o as o d O 00 tN o m IT) m oo m O d o o o d d d d o m Tir-~ oo in m m rj o ro r~ o OS o d d d d ON CN O o O ON oo o -s: 3 00 c •J B B 0) 3J ^ <3 ^ = > <3 •S o .5 r so •S "o o s 5 C a _ I Q •5 5 s 134 ^ 1 > ^ X -S g D, 5 " _ w r* = S § 3 io ''^

PAGE 147

D _N "(3 B c 3 O. T3 C ea 3 CJ D -J eg H c 'o E > c •= < o o x: CJ E o. o o c «5 4) o c > 13 ^ = < o E c o nJ C U(1) o =2 W o O c 33 CJ CJ c C X E B B 1) B C. X o B E O Q. X W era E o B X ON O I o ^ o oo . IT) ON m IT) o _• o ~ O . o O o ON o d 00 m o ON oo o <^ o o o _; o o d ' o o o CM ^ O " 00 o — o d ON to 8 o o o CM O O m o 5 o 1 > 00 S § 5 § S CJ B aj 4j C3 _ ^ O) D CJ B aj CO (SO S s 1^ * • 0237 ON oo 0164 .0136 ,0332 NO o 1 o 1 o 1 t o o ID CO o 00 22 S o d oo o d 135 1 > 111 III °-'8 9 (/3

PAGE 148

'J c in ?3 13 ^ 6D C £ < o o en c CT3 4> it a> 'J c u «5 •T3 CT3 3 c CO o D -J X) > o o < o E B 'E >» «« c o £ u 1) c X c 0) c ?3 c D C o o. X O m d O ' oo OS O — P d o • ON o o O 1 ^ : 2 o tN d o — — O • is O ^ IT) m O 9 (N OO o r-O On o O On O — O o o 00 o ^ ON o o IT) m li-i o NO p d oo ro oo O OO O O ON oo 8 d s: o s si, 1) ^ z ^ s; S "S 5 ^ c "I I •S o 4) ^ ^ 2> > ^ • S<3 •5 5 ^ ^ 136 -1 > 111 III eg p " -s ^ I § § I ^ S =2 «^ r

PAGE 149

T3 O 3 _c 'E o O c E o. 3 O o "o C o U IS CO H o c > 05 O :S < o o E c > ±= > U — .E < 'J c £2 UJ 03 'c J 1) 1) 73 c a. X E 03 o II so C X in o c 4J C O Q. X X CN — ON o o oo r<-i o — d o in o d d m J 00 t-~ o ^ d O . o ^. d
PAGE 150

c X! c o 60 B O o OS S Q. o 60 o c o 60 c o o s w 5t: c .2i > W1 a> so = < > o c > D 60 13 ^ = < c '0 'E •J u ED C c c X o c 73 X c u B X c i) c o Cl X C OS in oo o o — oo o * tN o a\ o\ p d '' 1 d d d 2S • 5 ^ o r-~ oo — O o o 8 5 o _ 8 lO o IT) Is O OS O o OS o m o (N * S 2. oo o — in vn O 0643 (N fN .0492 * * (N ,0473 1.23 CN — O m o o o d o in in t-d O O q d oo oo p d in q d OS oo ^ > 60 c o ? ^ > ?U • > « 2j Q ^ so g 2 I i> 3 h 3 138 u > level. u 1 o >> IS at (babi iroba probi g o, o.^ 5 « „ c I § s S -5 c 5,

PAGE 151

e « £ > S3 y •s < o o j= o E o c 2 S > 13 ^ C 0) S3 = < T3 O i> E c is o £ -J a> c "G B CO C o S3 3 S3 H c c X LU S3 c u c X E 73 U c o a. X S3 00 o O O o o o in oo o \o OS o o oo o o \o d o IT) o • (N — 00 O \0 d ^ b 00 o ro ^ O oo ^ 00 O oo 8 o in 04 o O oo o oo M — o OS O O o — in . 00 . m oo 5 ^ o m O m o o o . oo o in o o level. level >> >. >> ttbili '£ '2 a a e probi c S «> .a e it: = Si .2P "5 " .

PAGE 152

c > 1> = < 'J c > O -J »3 CO E > C •J z < 13 E c "c >> o ta c i£ u o •J .1> C c _o CO 3 o. •T3 C 03 •J C UJ 'J c D r3 c D C c. X :0 E C C 33 c 0) c Q. X E o oo d I oo ^ in o ^• d ' — I/O O 00 CO O 00 o o o in o O f*0 OS O CTs o ^ o — CO o o \o 2:: • o d o d o o O ro OO fN ON o in as oo O ^ g ^ d 8 S o 00 * o 00 in m m — o q q o o \o d d d d d o OO 0\ ro vo m ON ON O 00 5 2 o d d d 0236 m ,0149 o rn 6100 q o 1 9 9 i' o 60 •S 60 •S 1 SO o o ex ,0) CO Q t D C 1) 1so •S o 5 ^ ? so •S o S 5 i CO 3 140 .a >. >• 111

PAGE 153

•J s .2 > 15 ^ •J < E > en c 4) > o 1) CO = < T3 O x: n — CN O 00 in o ON m o O ON o so — m O Tt >n so ro oo O oo CN — O o d CO 8 9 vn CN U-) O in o Tiro cn O rso o CN 00 CN in O so -5 o o s: o 1 1.. so o S 0) 5 ?^ • •J c I 5 so •5 o 3 « ! C 5 * ,H i| b >/2 ,

PAGE 154

3 O B O O o c > o CO E 00 c S3 u B > = < o o 6J) C 1> 1/^ 13 ^ c3 ^ i> c c c X CT3 3C c a. X .5 c c o a. X in ID — m o 5 ° o in O (N O \D oo O ON If-? a^ ^ o o o d in o O CN O O d o — as ON oo ON oo o « ON GO (N ON m o O o O d d 1 d 1 O 9 ^ m o in o — NO m O ON o — CO m in O ON m oo (N 1 =^ • > I so •S o o a _ ^ a so •S in 5 3 142 e _ ago = s a « S ,a c h: to 00

PAGE 155

143 Summary This chapter reviewed the theoretical and empirical features of human and social capital studies. Human capital, the development of which is a fundamental goal of HARCs, is undoubtedly important to production. However, the maimer in which it influences production is not guided by formal theory. Empirical human capital studies also encounter specification problems in that conventional components of human capital, such as knowledge and health are presumed to be captured by education and height and weight. Social capital is similarly devoid of formal theory and has been examined in few empirical studies. The general notion underlying social capital is that it appears in forms of social interaction that are not necessarily directly related to work. Thus religion, games and music were considered as forms of social capital in how they influence efficiency. Differences in technical and allocative efficiency of individual HARC maize and bean producers were examined in five groups: personal and household characteristics; physical capital; social capital; and extension and experience. Maize provided more significant differences than beans, because maize data had more variation and degrees of freedom. The most prominent characteristic associated with higher efficiency is the proportion output sold. Producers appear to be more efficient the more they commercialize their output. Lecture only stands out as the most effective extension method as measured by efficiency. Primary schooling and literacy positively influence efficiency. Age shows an inverse relationship with efficiency, especially for maize. Catholic farmers appear more efficient than non-catholic farmers, perhaps because base community organizations provide communication fora. Health, as measured by height and weight and limb circumference, show positive influences on technical and allocative efficiency. Experience, particularly with herbicides and insecticides, appears to enhance efficiency in maize production.

PAGE 156

144 Finally, this chapter examined the complementarity of education for different types of extension methods. The Lecture only group showed a strong correlation with allocative efficiency at all levels of education, suggesting that training and education are substitutable. However, farmers with at least one year of education in the control group appeared statistically and significantly less efficient than their counterparts that received some form of extension and training, suggesting that education and training are complements at higher levels of education. Complementarity at higher levels of education also emerges from the Publications only tables of maize and bean producers and from the general pattern of maize tables for Lectures and publications and Lectures and visual aids.

PAGE 157

CHAPTER 7 THE NATURE OF HONDURAN AGRARIAN REFORM COOPERATIVES AS PRIVATE ENTERPRISES The preceding chapters consider aspects of HARCs in a neoclassical economic framework, which evaluates the production function in static marginal relationships. Broader discrete relationships, on the other hand, form the boundaries of institutions that matter and cannot be captured by mathematical or statistical models. According to Jensen and Meckling (1979), the production function cannot be adequately evaluated without addressing institutions that govern production: ...the production function of the firm depends on the specificaticn of rights and the laws or rules of the game governing contracting. The maximum attainable output of a firm is then not purely a matter of "physical" possibilities given the technology and knowledge; the production function depends on the contracting and property rights system within which the firm operates. The purpose of this chapter is to frame neoclassical findings about HARCs in a broader and dynamic institutional setting that takes into account contracts and property rights. The analysis pursues an understanding of what Williamson (1991) terms "discreet structural alternatives," as opposed to marginal variations. This chapter is intended to make two academic contributions. First, important efficiency questions persist regarding the public or private provision of services such as education, communication and commodity exchange. HARCs and coop support agencies present an unusual mix of public and private enterprises that could contribute to the growing body of institutional economic thought and provide solutions to policy makers assisting HARCs. 145

PAGE 158

146 There is also a noticeable gap in the literature between cooperative theory (Abrahamsen, 1976; Cobia 1989) and organization theory (Chandler, 1990). The latter primarily treats hierarchical management issues of large private firms, and the former describes the various rules and regulations that can govern a cooperative. Williamson (1985) considers institutional aspects of producer cooperatives understudied, and calls for further research to determine their benefits and costs. Observations were made throughout the course of fieldwork (and two years of Peace Corps service) which serve to illuminate the broader implications of the neoclassical analysis and explain areas of debility in HARC operations. Ultimately, this chapter is concerned with examining institutional options which may improve HARC performance. Introduction HARCs in general are conceded to be inefficient as reflected by the high rate of loan defaults. They absorb scarce national reserves, and have prompted influential Honduran sources to call for their dismantling'. With respect to the sample in this study, only eight of the 27 HARCs claimed to hold no outstanding debt for the purchase of short-term variable inputs. Large regional cooperatives that made substantial fixed capital investments have also fallen into insolvency. Desertion rates reflect part of the problem. According to USAID (1978) and INA (1985), HARCS endured a 20 40 percent membership decline between 1968 to 1984. Barham and Childress (1992) attribute the abandonment in part to the large number of people required to invade land, and the relatively fewer number of people constituting the optimum coop size. 'Numerous reports have appeared in the major Honduran newspapers, including El Tiempo, La Tribuna and La Prensa.

PAGE 159

147 Incentives are also geared to encourage abandonment of disenchanted laborers or managers responsible for misinvestments: HARC members have a right to compensation for duties performed if they leave the coop, but they are not beholden to any debt outstanding. The problem The results of this study challenge conventional wisdom (Alchian and Demsetz, 1972; and Jensen and Meckling, 1979) in that collective production systems appear more technically and allocatively efficient than individual systems. Thus, the problem with cooperatives appears to lie in the financial management of the enterprises, not in production. This lends promise to cooperatives as viable enterprises because, whereas production inefficiencies resulting from ineffectual worker incentives and shirking cannot be easily remedied, management problems can be overcome by identifying sources of institutional weaknesses. So the question arises, if HARC failure cannot be attributed to the inefficiency of collective production, where can the causes be found? Perhaps a solution can be found by a more detailed examination of the HARC as a firm, which neoclassical theory consigns to a black box. A great deal of economic activity occurs within firms (Barnard, 1938; Chandler, 1990) and, in the case of HARCs, between firms and government agencies. HARCs are governed by a conglomeration of implicit and explicit rules which vary in their degree of enforceability and thus credibility. Those institutions, regarded by North (1992) as "formal rules (statute law. common law, regulations), informal constraints (conventions, norms of behavior, and self imposed rules of behavior); and the enforcement characteristics of both" provide the context within which marginal insights can be evaluated to clarify misconceptions and/or contradictions.

PAGE 160

148 Cooperative Enterprises and the New Institutional Economics The "new" institutional economics (NIB) is distinct from traditional institutional economics (e.g. Commons 1934, Veblen 1899) in that it takes a positive rather than a normative approach. The body of literature in the new institutional economics has not yet reached an integrated, cohesive theory (Furboton and Richter, 1991). Indeed, the definition of "institutions" varies among the experts: the humanly devised constraints that structure political, economic, and social interactions. They consist of both informal constraints (sanctions, taboos, customs, traditions, and codes of conduct), and formal rules (constitutions, laws, property rights) (North, 1991); sets of ordered relationships among people which define their rights, exposures to the rights of others, privileges, and responsibilities (Schmid, 1972); Institutions can be demarcated into two classes: (1) conventions; and (2) rules or entitlements (Bromley, 1989); regularities in behavior which are agreed to by all members jf a society and which specify behavior in specific recurrent situations" (Schotter, 1981) Conceptual tools also inevitably differ and are usually refined to suit particular circumstances. The rationale for institutional economics, to account lor the shortcomings of neoclassical theory, holds as much intuitive appeal as it does challenge in application. In spite of the widespread acknowledgement of Coase's postulate that the choice between organization within the an economic unit or through the market forms the nature of the firm's existence, empirical applications are sparse (Coase. 1991). However, elements of NIE are emerging which provide means for expanding insights of neoclassical analysis of the "bl ick box" or firm.

PAGE 161

149 Why Firms Exist According to Coase, the existence of a firm rests not on some technological base, which was the conventional wisdom prior to his seminal 1937 article, but primarily on the minimization of transaction costs, the costs of doing business. If transaction costs wers zero, labor and capital could be traded across households or individuals. Using an example from The Wealth of Nations, the manufacturer of pins would find it very costly to obtain labor services through the market from the households of the sharpener and the head maker. Delivery and collection of pin components alone would involve prohibitive costs, but the process of determining prices and negotiating individual exchanges also impose costs. Contracting labor services under one roof where the sharpened pin can be handed directly to the head maker considerably reduces the cost of conducting organizational transactions. Institutions are important to any economic system which involves positive transaction costs because, as Coase (1991) rationalized, "what are traded on the market are not, as is often supposed by economists, physical entities but the rights to perform certain actions...." Campesino unions altered land tenure patterns to broaden member rights to market labor services in a manner formerly restricted by dominant share tenancy systems. Landless farm laborers banded together to obtain for their actions a greater share of the return to the land to which they had devoted their labor and to gain greater control, via cooperatives, over firm management which governs their actions. Institutions then, particularly legal instruments at the disposal of policy makers, are critical to expanding the rights to "perform certain actions," a fundamental goal of both classical economics and HARCs.

PAGE 162

150 Cooperatives in the NIE Most of the empirical NIE literature regarding transaction costs deals with explicit contracts between firms in developed countries (Williamson and Masten, 1995). Implicit contracts are recognized as being just as important, but diey require assessments that are more subjective in nature. Producer cooperatives, which are woven together with implicit contracts, have received little attention in economics literature. The term only appeared in ten references contained in the ECONLIT database (there are of course scores of articles on cooperatives, but they concern autonomously formed firms with clearly defined property •ights). Corporative enterprises have a notable survival record in France and Italy and have been successful in East Asia (Asian Productivity Organization, 1989). Recent empirical evidence from Italian and French cooperatives (Bartlett et. al., 1992; Estrin and Jones, 1992) and the famously successful Mondragon cooperative in the Basque region of Spain suggest that labor managed firms are even more efficient than private firms. The dynamics that contribute to cooperative success or failure are not well understood. Theoretical exercises" regarding cooperative operations are based on restrictive assumptions with respect to objective functions and work rules. It may prove more insightful to consider internal cooperative relations as they are influenced by the official agreements and implicit contracts upon which cooperative alliances depend. Knowledge about the viability of contracts under varying legal and social constraints would lend more precision to the task of designing efficient corporative organizations. ^See Prychitko and Vanek, 1996.

PAGE 163

151 Contracts Transaction costs are inversely related to the efficacy of contracts. The nature of contracts thus matters to collective enterprises such as HARCs that are bound by both internal and external contracts. The ideal transaction is "sharp in by clear agreement; sharp out by clear performance" (Macneil, 1974). Unclear agreements and obscured performance weaken the confidence critical to effective corporative enterprises. The benefits derived from contract are jeopardized in the presence of bounded rationality or "limited cognitive competence" and opportunism or "self-interest seeking with guile" (Williamson, 1985). Both are notably present in HARCs. Contracts can be supplanted by the market when standardized assets are exchanged, even in the presence of bounded rationality and opportunism. The identity of parties related by exchange is unimportant in that assets may be acquired elsewhere; continued relations are unnecessary. However, when asset specificity or non-homogeneous assets, exist, there is the potential for some agents to secure "quasi-rents" generated by the uniqueness of the assets they control. In the case of HARCs, asset specificity characterizes the services and inputs that are ostensibly obtainable from government support agencies and/or private firms. Exchanges in which all three are present, bounded rationality, opportunism and asset specificity, require a governance structure to ensure the integrity of the transaction. Efficiency is served when organizational structure is adjusted to minimize transaction costs which vary according to such things as capital holdings, frequency and risk. According to Williamson (1985), the organizational imperative becomes this: "Organize transactions so as to economize on bounded rationali:y while simultaneously safeguarding them against the hazards of opportunism." (p. 32).

PAGE 164

152 No task could be more appropriate for the viability of HARCs. One tould argue that such an imperative applies to any type of economic organization. However, most organizations which develop autonomously economize transactions in a safeguarding fashion because participants are protecting their interests. Transactions in Honduran cooperatives merit .special attention because the environments in which they operate were established by outside sources whose incentives did not necessarily coincide with those of the coop members. The following sections review those institutional apparatuses surrounding HARCs and HARC production processes. Contracts Related to HARC Basic Grain Production Most of the contracts associated with HARCs are implicit, the oroad outlines of which are documented in plans drawn with support agencies and in minutes ov meetings. At the time data were gathered, various agencies assisted HARCs in drawing up plans for the upcoming planting seasons. Normally, maize is planted in primera. Beans are planted in postrera between the rows of dried maize stalks. HARCs in good credit standing receive loans from the National Agricultural Development Bank (BANADESA). Those who have credit problems solely with BANADESA normally receive credit and assistance from the Natural Resource Ministry (RR.NN.). HARCs that have outstanding loans with both those institutions can obtain credit from the National Agrarian Institute, INA. Loans are usually granted at the beginning of the season on a lump sum basis, but deductions must be solicited and recorded for each input purchase. Typically, an agronomist from one of those three lending/development institutions is assigned to the cooperative/borrower. The decision-making process varies depending on the size and social dynamics of each HARC, but agronomists' recommendations are not commonly disputed. In a few HARCs with marginally qualifiable credit standing, agronomists insist on

PAGE 165

153 "managing" the production as one collective unit to a point where profitable output is assured, after which time land is permitted to be parceled to individuals. After target levels of production are developed, solicitations are made to the credit agency to release ftinds for input purchases, sometimes including labor costs for coop members. Often these decisions are made by the agronomist alone, and, in the case of large HARCs, in conjunction with a few coop administrators. Postrera maize is harvested between September and November (o amortize debt. Initial harvests are stored for household consumption. HARCs have an ostensible contract to sell surpluses at guaranteed prices to the Honduran Institute of Agricultural Marketing (IHMA), the national grain marketing board. Usually, however, they sell to intermediaries, known principally as coyotes. Bean production occurs exclusively on individual parcels. Credit is provided to some HARCs for bean production as well, though for much smaller sums as beans require fewer inputs. Bean output may be sold to IHMA for guaranteed prices as well, but is usually sold to middlemen who own transport vehicles. The Standard Administrative Chart A standard administrative chart illustrating HARC internal and external associations is displayed on Figure 7.1. Ail HARCs incorporate labor, most of which is unspecialized. Management and peer instruction, two elements that represent fundamental reasons for forming cooperatives, are also incorporated. Management gives rural farm laborers power over their lives and, in most HARCs. members are required to serve periodically in official management positions. The mere act of collectivization reduces barriers that formerly prohibited peer instruction, a service that would involve transaction costs among individual, unassociated farmers.

PAGE 166

154 A few relatively large groups own machinery and storage facilities. Smaller HARCs do not produce the volume necessary to justify incorporation of such lumpy investments and acquire them through the rural service market, from parent regional coops, or from government agencies. Three government agencies provide production assistance to HARCs, INA, RR.NN. and BANADESA. All three provide credit and technical assistance. INA is also charged with the very delicate task of providing land-use rights and ultimately, to those few groups which make all the payments, full ownership title to land. RR.NN. on occasion supplies variable inputs, including seed, fertilizer, pesticides and machinery services. The principal private enterprises that serve HARCs are input suppliers and intermediary supply purchasers. The national grain marketing board, IHMA, purchases output at ostensibly guaranteed prices. Explicit contracts are devised for credit and land titling. Technical assistance is a part of the credit contract, but roles of technical advisors are not listed in great detail. Technical agents are usually given authority to control the dispersion of credit funds, a rational instrument intended to protect lender interest, that can be used to influence HARC decisions.

PAGE 167

155 c 2 § 'C c .5 _o c3 e 0 c B w C c « c .2 c S T3 T3 c 3 on OB

PAGE 168

156 Transaction Association by Activity Figure 7.2 shows a transaction map by activity, rather than by organizational entity as in Figure 7.1. The minimization of transaction costs associated with activities constitutes the economizing action which gives rise to the firm. Property rights of activities are consigned to organizations by explicit laws. The use of those activities, to generate independent and efficient HARCs or to absorb HARCs to serve agency interests, depends on the viability of contracts governing agency HARC relations, the institutional environment. Purchase and sale contracts, which represent the exchange of money for inputs or outputs, are displayed by dashed lines in Figure 7.2. Solid lines represent service and support contracts for which no money, outside member or administration fees, are exchanged. Explicit contracts are indicated by black lines, implicit contracts by gray lines. All explic.t contracts are assumed to be transparent and implicit contracts non-transparent. Thin dotted lines depict the coopting of another entity that is enabled by not fulfilling implicit contracts. The purpose of the diagram is to identify implicit contracts and examine their reliability. The activities in Figure 7.2 illustrate the contracts that govern transactions for the "right to perform certain actions" (Coase, 1991). Figure 7.2 thus represents a more fundamental picture of relations of the black box or firm than Figure 7.1. Transactions wiJi particular entities are not necessary to the basic needs of HARCs; transactions that acquire rights to perform essential actions are necessary. Assessing the acquisition of those rights renders a more precise illustration of HARC weaknesses and how they may be resolved. Incentives are of a different nature in the two diagrams. In the standard administration chart, incentives are aligned within each organization. In the activity chart, by contrast.

PAGE 169

157 ling ing rke 1 1 Ma S * e5 o v5 Intern |> '4-» <: X) c o c s <:

PAGE 170

158 incentives are characterized by the nature of the contract that governs the transaction and by the relative position of the activity. If the contract is not a credible commitment, and is not buttressed by a credible threat, then the right to perform the action is jeopardized. The manner in which transaction costs are minimized, through the market or through internal organization, is not the same for private firms as for quasi-autonomous firms. Most of the NIE and transaction cost literature deals with autonomous capitalistic jnterprises (Williamson, 1985) or macroeconomic issues of public welfare (Bromley, 1989). HARCs are micro-enterprises that cannot be considered capitalist or even autonomous. HARCs thus present a third type of transaction cost minimization alternative beyond the market and internal organization: contract with mass social organizations and government entities of potentially dissimilar incentives. The discussion below considers the incentives and viability of ostensible constraints associated with the parties of each activity. Internal Contracts: The Black Box of the Firm The HARC is represented by a thickly lined box, the black box of the firm. Internal contracts are both explicit and implicit. Formal, apriori established regulations, or "rules of the game" provide a broad governance structure for cooperation among HARC members. They are determined by institutional transactions (Bromley, 1989), or what Davis and North (1971) term the institutional arrangement. For example, requirements with respect to meeting attendance, fmes for missed work days and election processes form recognized and enforceable parameters. Many regulations, particularly those governing monetary dealings such as input purchases and output price negotiations, render discretion to HARC central administrators. That discretion is insulated by ignorance when information is lacking to the asemblea general (general assembly)

PAGE 171

159 through illiteracy or high transaction costs of acquiring information. The larger and more complicated the coop, the more elusive is operational information. Labor-labor contracts HARC members contribute labor to an enterprise in which they expect fellow members to share their labor. That collective plots are relatively more technically and allocatively efficient than individual plots suggests that implicit contracts across laborers may hold. The greater efficiency may be attributed to better organization, but if labor commitments are relatively weak on collective parcels it is not enough to override gains in efficiency resulting from better organization. Given the centrality of labor to basic agricultural production, collective or individual, it appears that internal implicit labor contracts are strong. Administration-labor contracts Administration of HARCs takes on many forms. Ostensibly, officers are elected on regular intervals and serve terms established in cooperative guidelines. The elected officials are responsible to formally represent the HARC to outside interests and sign formal, explicit contracts. Often, however, real authority rests with one or a few individuals. When HARCs were initially contacted for participation in this study, some presidents directed inquiries to strong-minded individuals who held no elected posts. Their dominance A'as also often observed through the dynamics of HARC meetings. Ultimate authority rests with the asemblea general, which approves operating plans, but the details are carried out by central administrators. It is a logical arrangement. Even if all HARC members had the capacity to review potential production technologies and financing arrangements, which few do, the time involved in acquiring, sorting and processing relevant

PAGE 172

160 information would entail significant transaction costs. Ideally, administrators would allocate H ARC resources to those activities which have the highest opportunity costs to the collective unit. The viability of implicit contracts between central administrators and labor is difficult to objectively determine. Some HARCs had fallen into unwarranted debt because of unscrupulous administrators, who were subsequently ostracized by the other coop members. However, an important constraint on HARC administrators at the local level is the relatively small amount of credit extended on an annual basis. Unlike large capital investments, which are less open to scrutiny and offer substantial sums for opportunism, short term loans are small and invested piecemeal within the cooperative unit. Thus implicit labor-administrator contracts form more credible commitments. This is not to argue that contractual dealings, including credit and the purchase of inputs, between administrators and technical assistance agencies enjoys the same degree of credibility. Still, the relatively minor gains obtainable from the low credit level, and the credible threat attending localized operations, inhibit administrators from indulging in activities which would jeopardize long-term opportunity and thus Jie welfare of HARC operations. The main weakness of the administration-labor contract is attributed to inadequate human capital. External Contracts: Association with Support Agencies The maimer in which transactions are organized with supporting agencies is just as important to the cooperatives as internal transaction arrangements. Operating designs of HARCs almost always result from external imposition, rather than the autonomous will of the campesinos. In early stages of coop development, paternalism and external assistance are inevitable to generate long run, self reliant and voluntary cooperatives (Lele, 1981). HARCs are constructed under the guidance of strong personalities within the group, the parent camp^.sino organization and

PAGE 173

161 government agencies related to agriculture, agrarian reform and cooperative development. Just one agency, the National Agrarian Institute (INA) "intervenes in project design, feasibility study, technical and administrative assistance for credit acquisition from national and international lending institutions and processing and/or business agreements with national and international companies Large (cooperative) projects.... are managed with almost total government agency support and direction from the formation and transfer of the group and the granting of land to the sale of the product." (Cardonap. 102-103) Other government agencies and campesino unions exert similar influence over the cooperatives, rendering the coops vulnerable to state exigencies or the political priorities of the party in power. The administration-labor contracts are distinct from contracts HARC administration manages with external agents. The bitter disputes between HARCs and support agencies, with both sides charging the other with mismanagement, are not attributable to ineffectual administration-labor contracts. When support agencies and HARCs come to loggerheads, usually the majority of the HARC goes along with the official position. Campesino unionizing The foundations for unions are different from those of cooperatives. Unions are founded on political organizing and require a large number of members to be effective. Cooperatives are founded on capital and must at least achieve solvency if they are to survive and satisfy any other objective. Cooperatives may be large, as demonstrated by those in the United States where they have increasingly taken on a product processing role to provide value-added margins to declining rural communities. But producer cooperatives are by nature relatively smaller than unions because the transaction costs associated with coordinating workers in dissimilar situations are high. Individual producer cooperatives face problems specific to their lo;al areas and situations, problems that cannot be addressed by broad policies of large umbrella organizations. Coop members feel a much stronger sense of ownership of their coops than they do of the campesino unions to which they are also members because communication is local and visible.

PAGE 174

162 Incentives are ostensibly similar for campesino unions and the HARCs they spawned, namely the operation of economically vibrant enterprises. However, the constraints placed on individual administrators of campesino unions vary significantly from those of HARC administrators. HARC administrators are closely scrutinized by members who almost always live in the same community. Decisions made in campesino unions, on the other hand, are open to a cluster of close associates who constitute a relatively small proportion of the entire unit. Coop administrators are more immediately beholden to HARC members' interests than are union administrators. Union power is often built on small clusters of association within the larger union. Problems arise when union administrators exploit HARCs through violation of implicit contracts that serve to incorporate HARCs into serving the larger purposes of union administrators at the expense of HARCs. At that point HARCs begin to lose autonomy and cannot be fully considered "firms." Nor can campesino unions be considered autonomous political institutions, given the national and international forces which sought to influence them. Campesino union management handles large sums of money for both local support organizations and for development assistance on capital intensive investments. The latter is usually worked out with the government agencies that provide technical assistance and, in the case of international funding, breign agencies. Large regional cooperatives, such as CARAOL, made substantial capital investments in deals worked out between campesino unions, government credit agencies and private suppliers. Several concerns were heard among HARC members, who were ultimately responsible for repayment, that provisions were not made for equipment maintenance and, perhaps more importantly, competing investments had higher opportunity costs.

PAGE 175

163 Land tenure Land tenure is tenuous in Honduras, nonetlieless legal documents exist between INA and each official HARC through their campesino unions. Very few HARCs own their land, but are given permission to use land. The challenging task of balancing intense, partisan political dispu:es involving land tenure falls on the National Agrarian Institute (INA). It is INA's responsibility to clarify poorly defined land rights in an agrarian-based economy between traditional landed interests and landless campesinos who claim land is underutilized and not fulfilling a social purpose. Two former heads of INA, Sandoval Corea and Ponce Cimbar, acknowledge INA^ has had difficulties in distinguishing public lands from private lands (CEDOH, 1992). Perhaps because of the high stake political battles, INA has played a rather capricious role in Honduran resource policy and the Honduran economy. Sandoval (CEDOH, 1992) expressed concern about the conflic: of interests that arises when the land tenure institute, INA, gives any funding to campesino groups, large or small. INA evolved from an agency entrusted to settle property rights disputes to a development agency that provided credit and tens of thousands of dollars to campesino unions fo. education and training projects. Land tenure instability has, aside from the obvious long-run problems, short-run consequences for HARC performance to the extent that members' expectations about the viability of the enterprise diminishes their investments in it. Technical assistance Technical assistance, which ultimately governs credit and input supply is managed by one or more of the government support agencies, INA, RR.NN., or BANADESA. The individuals employed to provide technical assistance are not necessarily evaluated according to the success

PAGE 176

164 of the HARC operations for which they are entrusted. Agencies are placed under the auspices of politicians whose authority rests on their ability to remain in office. To the extent that HARC performance contributes to prevailing political priorities, HARC interests will be served. However, at most, HARC interests take on secondary importance and may be discounted entirely to the extent that ignorance and complexity obscures institutional transparency. If contracts do not hold as they are designed, HARCs may be absorbed into the larger, politically geared national institutions that provide technical assistance, and must serve the objectives of those organizations above their own. They may take on the detrimental characteristics of the notoriously inefficient "state-run farms." Oversight of technical assistance is relegated to each HARC. However, administrators responsible for reviewing and negotiating agreements for technical assistance, which includes credit and input purchases, may not be literate. The principal architect of the historic 1968 land reforms, Roberto Sandoval Corea, expressed bewilderment at the manner in which international development agencies supplied funding for projects without effective controls. Sandoval also charged that subsequent state financing of campesino unions had a corrupting influence and thwarted the reform movement (CEDOH, 1992). Corruption, which the Cold War accelerated markedly in Honduras during the "lost decade" of the 1980s (Hayes, 1989), is widely recognized as a pervasive influence in the Honduran economy. Credit Virtually all seasonal credit for variable inputs comes from government agencies. Credit contracts are written so that the cooperative is ultimately responsible for repayment. Almost all loans are unmortgageable and given on a short-term basis for annual operating expenses. No

PAGE 177

165 individual, including an elected official, is technically accountable for repayment. Holding institutions rather than individuals responsible for credit operations presents a credit risk, particularly in the absence of auditing services. Credit from government agencies is usually granted for the upcoming growing season and must be repaid at harvest. Defaulted loans incur substantial transactions costs for both lending agencies and HARCs. Outstanding debts compel HARCs to seek alternative sources of funding with intermediaries, who provide inputs or cash at high interest that is deducted from the forward purchase price. Ties with government agencies under such circumstances jeopardize potential earnings because the government marketing board, IHMA, may withhold payment to clear past debts. By the same token, government lending agencies incur larger transaction costs of monitoring defaulted groups in order to recoup previous losses. Marketing Most HARC output is sold to the national marketing board, IHMA or intermediaries. The harvesting and selling of grain is visible to all members, who uf ually participate in the process. Grain prices paid by IHMA are supposedly guaranteed by futures' notices, but the rules governing maize grading and pricing are inconsistent and have disappointed some HARCs. A marketing incident of a HARC in the sample illustrates the role of transaction costs and the importance of enforceable contract. Rather than sell to an intermediary, the HARC attempted to incorporate transportation into its operations by renting a truck and delivering the maize directly to IHMA. However, IHMA quoted a price considered unsatisfactory by the HARCs members. After waiting at the gate for two days trying to negotiate a higher price and paying truck rental charges, the HARC ultimately sold to an intermediary for a i.igher price than offered by IHMA. The intermediary sold the grain to IHMA, demonstrating clearly that IHMA was

PAGE 178

166 willing to pay a higher price than that offered to the HARC. The incident reveals a breakdown in the contract between IHMA and the HARC and another, illicit, contract between intermediaries and IHMA. The notion that prices are unreliable would discourage HARCs from incurring transportation charges to sell to IHMA, thus reducing the efficiency of HARC operations in general. The theory of transaction costs also serves to explain why IHMA administrators may choose to deal with a few intermediaries rather than several grain producers. Conceivably, the same extra-official commission that motivated the higher price for die intermediary could also be negotiated directly with each HARC. But negotiating one commission with one or a few intermediaries to exercise price discrimination requires fewer transactions be made. It also alleviates the cost of risk associated with several producers who may not wish to engage in such transactions. Intermediaries can provide a valuable and efficient service, but only when they operate in a competitive environment and hold no extra-official price advantage over HARCs. "Information asymmetries" (Williamson, 1985) impose risks that increase the transaction cost of incorporating transportation higher than the cost of acquiring transportation in the market. Intermediaries' potential for monetary gain expands to the extent Uiey can exen monopsony control. Contracts that maintain credible marketing commitments would improve the efficiency of HARC operations and production distribution. Vigilance Committees Vigilance committees technically exist in HARCs to ensure that all contracts, internal and external, are carried out according to their stated purposes. However, vigilance committee members are often illiterate or not functionally literate enough to comprehend all the details

PAGE 179

167 associated with overseeing dealings with which they are unaccustomed. Even when they are capable of reviewing records, it involves another task on top of a heavy labor load, not to mention confrontations with friends. Proprietorship of vigilance is consigned, logically, to each HARC. After all, a main purpose of cooperative enterprises in LDCs is to empower impoverished individuals to look after their own best interests rather than be dependent on traditional elites. However, a contradiction arises when cooperative organizations are entrusted with the responsibility to oversee operations that, because of their complexity, involve significant external assistance and are by the same token beyond the comprehension of most members. HARCs are not kibbutzes, which survive under similar organizational structures because they have impressive levels of human and social capital. The question arises, can proper oversight be given to HARC interests in the absence of an oversight capacity necessitated by the level of technical assistance provided? The Design of Efficient Institutions The administrative flow chart in Figure 7. 1 differs in style and substance from the activity flow chart in Figure 7.2. The administrative chart shows official associations between HARCs and external support and market agencies. The activity chart outlines contracts that exist to facilitate the exchange of rights. The rights of technical assistance and credit are supplied by the government to HARCs in exchange for the right to ameliorate civil tensions by developing lucrative economic enterprises among the politically troublesome disenfranchised. However, policy mechanisms intended to exchange those rights breakdown becausi credible commitments are not made. Either explicit contracts are not enforced, or implicit contracts are based on norms or customs that do not exist.

PAGE 180

168 Williamson (1985) distinguishes credible commitments from credible threats in that the former are "reciprocal acts designed to safeguard a relationship" and "are undertaken in support of alliances and to promote exchange," whereas the latter involve "unilateral efforts to preempt an advantage. " The financial breakdown of HARCs indicates that credible commitments were not made in some areas of operation that were not safeguarded by credible threats. Credible Commitments Credible commitments are valuable assets to an economic enterprise because they involve no ex poste transaction costs. For example, if guaranteed prices are not offered at the time of sale, producers have to incur the cost of legal recourse. Similarly, if borrowers refuse to repay loans, lenders incur costs of pursuing payments. The only notable credible commitments made in HARC operations, the labor-labor contract and the administrator-labor contracts, occur within the black box or corporative enterprise that constitutes a HARC. Fortunately, the labor-labor contract is the most necessary of all credible commitments because it is the one that is virtually impossible to remedy by institutional modifications. The transaction costs of preempting, monitoring and recovering lost labor would not be worth the payment to labor. Conversely, transactions that cover larger cost items such as input purchases, credit disbursements and marketing are more concentrated and thus justify some additional ex ante or ex post transaction costs. Non-credible Commitments Non-credible commitments necessitate credible threats throu:?h various contractual mechanisms. Government support agencies and campesino unions both have incentives to incorporate HARC enterprises to serve their own operations. Indeed, evidence of the regional

PAGE 181

169 cooperatives such as CARAOL and the manner in which technical jobs are doled out to political appointees indicates that commitments between external sources requires some credible threat. Ostensibly, HARCs hold legal recourse to ensure their best interests are being served. However, HARCs do not have the human capital or administrative capacity to carry out the task. HARC administrators are susceptible to being inveigled into options tha: may not necessarily be in the best interests of the HARC. thus jeopardizing the viability of the labor administrator contract. Transparency of operations is virtually non-existent in relations between HARC administrators and government support personnel. Lenders of funds, both national and international have not imposed adequate auditing controls. The requirement of auditing, particularly as an a priori condition of a lending agreement, serves as a viable threat to preempt unwarranted advantage in transactions. An Empirical Observation: The Guanchfas Cooperative The institutional concepts developed above account for why HARCs may fail or remain weak, but do they also correspond to success cases? The Guanchfas cooperative on the north coast of Honduras, which was extolled as a model by all proponents of collectivization, provides an interesting case study. Observers of Guanchfas were quick to attribute its success tc its collective form of operation, ignoring other factors absent in cooperatives sponsored oy the goverrunent or campesino unions. Guanchfas was not initially successful in paying off its loans when it operated as a basic grains cooperative. It became modestly profitable when it abandoned basic grains for plantains, but could not obtain government support for cash crops (McCommon, et al. 1985). Guanchfas then turned to Standard Fruit's independent banana growers program. Standard Fruit agreed to purchase most of Guanchfas' production and provided scarce technical

PAGE 182

170 assistance, credit and inputs. Guanchi'as' success may be attributed to two aspects of the institutional design that prevented systematic leaks common to the basic grain sector. First, Standard Fruit had a financial interest in providing inputs and technical support in a timely fashion and in overseeing compliance. In the basic grain sector by contrast, government extension and loan agents are governed by political rather than economic oversight. Government agents collaborate with agency superiors and coop leaders whose best personal interests do not necessarily coincide with making cooperatives financially sound enterprises. A potential HARC loan default is likely to draw a greater response from a company representative responsible for generating returns to company capital than it would from a government representative. Second and perhaps more significantly, Standard Fruit had control over Guanchfas' output and lived up to purchase agreements. Cooperatives in the basic grain sector face more than one buyer. If they sell directly to the Honduran Institute of Agricultural Marketing (IHMA), outstanding loan charges are deducted from the bill of sale. IHMA also reneged on promises to purchase grain from farmers at guaranteed prices (USAID, 1982), imposing a higher risk burden on cooperative planning. Farmers gain a higher return from coyotes (intermediaries) who offer nominally lower prices than IHMA, but whose real prices, adjusted for lean charges and artificial quality deductions, are higher. The success of the working arrangement with Standard Fruiv is demonstrated most significantly by the increase in income to Guanchfas' members. The average annual income increased from 800 lempiras in 1969, the first year of the program, to 4,484 lempiras in 1980, over three and a half times higher than the average income in Honduras. Investment returns have been turned back into new capital equipment, member housing and social services in education and health.

PAGE 183

171 Standard Fruit signed agreements with two other coops whic'i were also successful, although not as remarkable as Guanchias. The arrangement is mutually beneficial. Workers gain control over production resources, and Standard Fruit maintains stable political relations and portrays a good image. Standard Fruit created a cooperative in March of 1996, not with an existing HARC. but by granting land and capital to a group of employees. Both Standard Fruit and the United Fruit Company intend to turn over more resources to workers to quell political opposition and worker strikes, which have proven quite costly in recent years. Summary The previous chapters demonstrate that HARC breakdown cannot be attributed to labor shirking theorized by Alchian and Demsetz (1992). This chapter attempts do differentiate further the underlying defects by examining HARCs in a broader institutional context. It is also an attempt to contribute to the institutional economics literature, which provides few guideposts for the analysis of cooperatives. Property rights and the contracts which govern the behavior of HARCs and associated agencies formed the core structure of the analysis. Incentives that are potentially present in each were also examined heuristically in light of statistical results and observations made throughout the course of fieldwork. A standard administrative chart was compared to an activity chart which delineated transactions for inputs and services vital to HARC performance. Internal contracts, including implicit contracts, appear to hold. The credibility of Laborlabor contracts is demonstrated by the higher technical and allocative efficiencies of collective parcels over individual parcels. Administrator-labor contracts also appear to be credible, demonstrated by the greater scrutiny that exists at the local level and the minor gains relative to risk associated with breaking agreements.

PAGE 184

172 Contracts between external agents and coop administrators, on the other hand, are less dependable, primarily because they are exercised without proper oversight and across parties of disparate incentives. The contract breaks down because there is a weakness of purview that conventional lenders would require over operations in which their funds are invested. Marketing contracts between IHMA and HARCs have also been unreliable, usually to the detriment of HARCs. Marketing contracts lack credibility because they are unaccountable and involve direct monetary transactions with government agents who art beholden primarily to political sponsors. If all contracts functioned ideally, "sharp in by clear agreement; sharp out by clear performance," (MacNeil, 1974) HARCs would be more independent and commercially vibrant. The problem is that contracts are not clear and are not enforceable, injecting risk and undercontldence in HARC development an undertaking which, by its collective nature, requires stability and assurance of mutual success. Vigilance is ironically relegated to the HARCs, institutions established in large part to develop in the long run the human capital to oversee technically advanced operations. HARCs lack the human capital in which some cooperatives, notably kibbutzes, are richly endowed. Provisions to account for short-run human capital deficiencies may be necessary. Independence of HARCs is threatened to the extent that their incentives do not correspond to those of government support agencies and campesino unions. HARCs become components of a larger government or political enterprise when contracts do not hold. Little wonder they have generated many of the same criticisms leveled against state-run farms. HARCs' identity as firms which minimize transaction costs is obscured because the instruments they rely on to carry out transactions, the contract, is not credible.

PAGE 185

173 The cooperative model presented in this chapter was developed in light of observed failures, but also conforms to cooperatives which have enjoyed substantial success through contracting with a private firm. The most notable, Guanchfas, demonstrated its capacity to function as a profitable, independent enterprise when it contracted with the Standard Fruit Company. Standard Fruit overcame the incentive problems inherent in the governmental and political organizations that ostensibly provide support, and Guanchfas evolved into a cooperative that produces high quality products and generates investment returns that significantly increase member income and achieve social goals. According to the cooperative model presented in this chapter, HARCs have the internal dynamics necessary to survive, but encounter problems with external agents whose incentives do not coincide with those of HARCs. Contracting with a private enterprise whose success depends on the economic success of the HARCs has shown itself to be a viable i olution to the problems that prohibit HARCs from emerging as independent and financially sound firms. The arrangement may appear as an increasingly appealing alternative in the future as Central America is opened to free trade and as autonomous political forces demand the integration of poor people into an economy that provides opportunities for all. This is not to suggest that government support agencies do not have a useful role to play in HARC operations. Guanchfas was aided completely by Standard Fruit because the company was large enough to provide all technical assistance. Most economic opportunities for HARCs dwell in the small-scale supply of products to local markets in expanding towns and cities. Perhaps some of the very skilled agronomists and technicians who have received training in Honduran agricultural colleges could acquire funding to contract with HARCs. Structuring incentives through tax breaks or other means for skilled agriculturalists could inject capital into HARC operations in a manner which safeguards lender interests. They would still depend on

PAGE 186

174 government support for research and technical assistance. More research needs to be done to compare the costs of such an endeavor with other rural development programs. Cooperatives serve as a viable organizational mode to transfer technology for the production of nontraditional commodities, and private agents provide capital and technical assistance. Most importantly, the collaboration is based on mutually beneficial economic, not political, incentives.

PAGE 187

CHAPTER 8 SUMMARY AND RECOMMENDATIONS Colonial forces confined Honduran campesinos to share tenancy systems that precluded human capital investments and advances in agricultural technology and marketing. Unionization of banana workers and the revolution in Cuba precipitated domestic and international pressures for agrarian reform. Those reforms gave limited land access to groups of farmers who were encouraged, through support services and through denial of individual land titles, to work collectively. HARCs' poor performance record was attributed to the collective mode of production and to internal mismanagement. Design of the Analvsis The analysis begins at the base level of technology. A logit regression examines a range of factors that influence technology adoption. Stochastic frontier production functions are estimated for maize and beans. Debreu-Farrell technical efficiency, the ratio of observed output to optimal output for a given set of inputs, is calculated for each producer from the observed frontiers. A producer's response to prices are reflected in allocative efficiency, which measures the adjustment of input mixes relative to input prices. It is an estimate of costs incurred unnecessarily given the actual level of technical efficiency. Allocative efficiency is calculated from an analytically derived cost function. 175

PAGE 188

176 In conjunction with other data, profiles of efficient producers are drawn from technical and allocative efficiencies. The most important aspect of the HARC profile involves the efficiencies of collective parcels relative to those of individual parcels. Human and social capital factors are examined by comparing the group means of technical and allocative efficiency averages. A comparison of advanced and traditional technologies is made by contrasting advanced maize production with traditional bean production. Finally, the empirical results are interpreted within a broader institutional model that examines the viability of internal and external implicit HARC contracts. The model is premised on the Coasian notion that a firm's existence is not based fundamentally on technology, but on transaction cost minimization. Economizing action is thus reviewed in transaction components and governance mechanisms. The model is empirically applied to the notably successful cooperative Guanchfas, which was cited by many advocates of collectivization as a paradigm worth replicating. Summary of the Results Collective production systems emerged in this analysis as being more technically and allocatively efficient than individual systems. The collectivity variable in the maize regression is also positive and statistically significant for average OLS and frontier estimations. These results counter conventional notions that cooperatives fail because laborers and managers have incentives to shirk work responsibilities. They also conform to direct survey evidence that shirking on collective parcels is not observed by HARC members. That collectives are more efficient reveals an important policy option for transferring technology, improving agricultural production and integrating the rural poor.

PAGE 189

177 That collective parcels are more efficient to the institutional "black box" of the firm suggests that the internal dynamics of the cooperatives are sound. Mutual agreements of laborers to carry out work responsibilities is a necessary condition to achieve collective efficiency. Agreements between HARC managers and workers are also reliable, due to local scrutiny and the low return associated with managerial shirking at the local level. However, transactions with support agencies are unreliable due tc the disparate incentives of HARC members and agency personnel. Commitments with external agencies are not credible and are not braced by a credible threats. Implicit contracts with external entities do not hold because HARCs' administration and internal vigilance, unlike the Israeli Kibbutz upon which HARCs were modeled, lack the necessary human capital to oversee complicated transactions. The unreliability is reflected in the low allocative efficiencies for maize and in poor marketing services. HARCs' potential for success is also demonstrated by the cooperatives which contracted for support from the private sector, most notably Guanchfas. A prominent result that emerges both in technology adoption and technical and allocative efficiency is the significance of the proportion of output sold. The more commercialized farmers become, the more prone they are to correctly adopt new technologies ond the more efficiently they operate. The positive benefits of commercialization also suggest that HARC members respond to markets and that profit maximization is a fundamental cooperative goal. Maize allocative efficiencies are significantly lower than th)se of beans. Maize production relies on more advanced technology and on a broader range of inputs than bean production. Unavailability of inputs reduces the allocative efficiency of production in that producers cannot adjust input mixes to minimize production costs. The low allocative efficiencies for advanced maize production systems underscore the importance cf effective distribution

PAGE 190

178 systems, particularly as the Honduran economy attempts to improve its agricultural technology and efficiency. Extension methods in general bear significant influence on technology adoption and efficiency, as demonstrated by extension coefficients in statistical regressions and efficiency measures. Personal instruction, unaccompanied by teaching aids, appears most effective in reaching farmers regardless of their prior education, and may be a substitute for education at low levels of education. However, other forms of extension, including publication circulation, appear as complements to different types of extension at higher education levels. Complementarity of education with certain extension types renders greater flexibility in designing extension programs. For example, the per unit costs of reaching rural farmers are much lower for publication circulation than training and visit extension. Human capital variables yielded predictable results. Investments in health and education are shown to stimulate technology adoption and improve production efficiency. Experience, particularly with herbicides and insecticides, also enhances efficiency. Religion is the only social capital variable that appears to influence efficiency. Social capital is treated tangentially in this study, as the methodology does not lend itself to examining complex social fabrics. Nonetheless, social capital is gaining prominence in the economics literature and constitutes worthy future research with respect to cooperatives. Recommendations Policy makers in Central America are faced with the difficult task of balancing social and economic integration of the rural poor while simultaneously expanding export programs in response to structural adjustment pressures from international donors. They also face the challenge of supplying growing urban populations with nontraditional foods that provide the

PAGE 191

179 various nutrients of a healtiiy diet. The results which emerge from this study suggest that cooperatives are an important policy option which may assist them in meeting those objectives. Cooperatives provide the economies of size that permit the lumoy investments required by the nontraditional commodities in which Central America has a comparative advantage. Cooperatives also serve as a proficient means for transferring the advanced technologies needed to produce those commodities. The important policy question is: Can cooperatives serve as efficient organizational forms in terms of production? The evidence to date, as measured by tlnancial performance, is that they cannot. However, this study shows that cooperatives are organizational forms which enhance technical and allocative efficiency in production. Cooperative failure is not attributable to shirking, but to unreliable contracts with support agencies whose incentives are based more on politics than economics. The results of this study, and successful cooperatives such as Guanchfas that have received solid private sector support, bear that out. Moreover, HARC members appear amenable to profitable nontraditional ventures, as demonstrated by the positive influence of commercialization on performance. Future Research Two policy paths could be pursued to address problems that burden HARCs, the details of which are grounds for future research. The first involves the attraction of private capital and attendant management expertise, the second the restructuring of support agencies. The Potential Role Private Capital Policies designed to attract private capital investments to cooperatives would be beneficial in meeting the dual needs of rural development and achieving export grcwth. Private managers

PAGE 192

180 of capital funds have the incentives to carry out the daily myriad tasks involved in operating a competitive and complex enterprise. Agreements designed to generate investment returns would benefit both investors and cooperatives and would serve broader national development goals. Net social benefits of attracting private capital merit examination relative to those associated with government assistance programs that are intended to develop poor rural areas. Perhaps private capital would have to be attracted at some cost, but may be a more efficient way of achieving social goals than direct government assistance. The mechanisms and incentive schemes for drawing private capital thus merit investigation. The Potential Role of NGOs Restructuring support agencies to be more operationally transparent and to align agency incentives with HARC incentives would be beneficial. Perhaps non-governmental organizations (NGOs) could lend assistance in this endeavor. NGOs dedicate dependable but meager resources to directly attack underdevelopment and often accomplish more than government agencies. If NGOs could direct their scarce resources toward improving the functioning of government support, not in decision-making but in the auditing necessary to insure that those decisions are carried out, the pay-off would be much higher. NGOs assemble a great deal of expertise capable of conducting the types of simple audits required by HARCs. Transferring human resources from direct de\elopment assistance to auditing government development programs would have a multiplicative effect on NGO efforts. It would combine meager NGO funding with government funding to secure the fulfillment of government development goals. Future research could examine how NGOs could be employed to complement ongoing government programs.

PAGE 193

APPENDIX I DATA AND STUDY AREA Integrated Pest Management Program of Honduras CMIPH') The Integrated Pest Management Program (Spanish acronym MIPH) mounted a three year pest management training and extension program in 1985. Operating through the Pan-American Agricultural School (Spanish acronym EAP) in Zamorano, Honduras, MIPH provided four different training techniques to 27 randomly selected HARCs in the El Parafso and Olancho regions of Honduras. HARCs were randomly divided to receive different types of training with respect to agronomy and pest management. One group was set aside with no training to serve as a control group against which extension efforts could be measured. The training types were: 1 . Printed material only 2. Lectures only 3. Lectures and printed material 4. Lectures, printed material and electronic visual aids. Trained agronomists visited the groups on a regular basis to give lectures and/or supply printed information. MIPH focused on common problems faced by basic grain producers and suggested cost-effective means for overcoming them. The Sample MIPH selected a sample of HARCs between 1985 and 1988 to participate in the pest management program. The HARCs were chosen at random by .extension agents, then subsequently and randomly assigned to receive one of the above training techniques. 181

PAGE 194

182 Data on the groups analyzed in this study were gathered between November, 1987 and June, 1989. A two-month sondeo was conducted at the outset to determine the strategy for surveying in light of the proposed research objectives and the availability of logistical support. All the cooperatives were contacted during the sondeo to secure their agreement to record data for a one year production cycle. Fieldwork was carried out through the primera (the heavier of two rainy seasons which occurs from May to August) and postrera (September to December) production seasons of 1988 89. Corn is the principle crop in primera and red beans are grown in postrera. Crossectional data were gathered for one production season. Fortunately, both primera and postrera brought good to average rains during the data gathering period. MIPH randomly selected the HARCs from among basic grain cooperatives in the two regions. The types of training offered to the HARCs were also selected on a random basis. The data gathered comprise a sample of basic grain HARCs, but the entire population of HARCs and HARC members who participated in the MIPH program. Most of the HARC members had an individual corn plot and over half had an individual bean plot. Individual production data were collected by literate coop members or children of coop members trained in each cooperative to maintain investment records on all activities pertaining to individual land parcels. Gathering production data is a difficult task, made more challenging by the distrust that people naturally feel about revealing economic and financial records. Cooperatives provide an excellent network for collecting accurate data in that trust need not be established with every individual but just with someone trusted within the cooperative. Although the author met every HARC member in this study, personal relations were more closely and efficiently cemented

PAGE 195

183 with a trusted person within the HARC who could observe production practices and elicit the confidence necessary to obtain accurate records. Nonetheless, it would be imprudent to argue that distrust was totally allayed, but the establishment of trust and personal relations that began with many HARCs during my Peace Corps service significantly improved the accuracy of data that would be offered to enumerators unfamiliar to HARC members. It should be noted that collective yield data were the most challenging to gather, in that outstanding debts threatened the extraction of surplus revenues if such information were divulged to debtor institutions. The provision of yield data on individual production posed little if any risk. Input records were also readily available as most farmers enjoy talking about their individual production practices. Lending institutions maintained records of coop purchases for all collective and some individual parcels. Human capital and demographic data were gadiered over a period of three months in 1989. The survey was designed through interviews with HARC members and extension personnel from MIPH and government agencies. It was tested extensively in cooperatives that were not included in the sample but operate in the sample areas. Of the 27 HARCs in the sample, 19 had a fully collective maize parcel and 25 had production systems that were completed under individual responsibility. The distribution of coop size in terms of membership and land access varied. Table AI.l compares nationwide average with sample averages for member and landholdings per asentamiento and landholdings per socio. Table AI.2 provides a breakdown of asentamiento names, members and land access per group and per socio.

PAGE 196

184 Table AI.l. Asentamientos by name, membership and land access Group Members Collective Individual Total Per Soci Ideas en Marcha 1 A 14 1 AA 100 100.00 7.14 El Boqueron 12 50 24 74.00 6.17 Empalizada 16 62 15 77.00 4.81 El Benque 7 5 5.00 0.71 Los Bienvenidos 23 44.5 12.5 57.00 2.48 bl tsfuerzo 24 140 140.00 5.83 Los Peregrines 1 o 18 66 36 102.00 5.67 Esquilinchuche 11 16 16.00 1.45 San Nicolas 7 12 11.5 23.50 3.36 Los Almendros 14 57 23 80.00 5.71 La bsperanza 18 102 30 132.00 7.33 Santa Cruz 1 o Is 92 92.00 5.11 L.ayo tsianco 35 5 40.00 4.44 Z/Opiiotepe 11 75 75.00 6.82 uudyiuuroS o An 47 47.00 7.83 Ld v^oncepcion 10 1 O 38 168.00 4.42 San Juan de Linaca A O 15 47 62.00 1.29 La Puzunca 15 90 90.00 6.00 Tempiscapa 7 10 14 24.00 3.43 La Providencia 30 45 115 160.00 5.33 19 de Abril 23 29 23 52.00 2.26 El Coyolar 13 14 26 40.00 3.08 El Plomo 9 46 8.5 54.50 6.06 Los Dos Naranjos 14 110 70 180.00 12.86 Los Venecianos 8 28 4 32.00 4.00 La Libertad 14 20 40 60.00 4.29 Montanuelas 10 28 28.00 2.80

PAGE 197

185 Table AI.2. Comparison of sample averages with national averages Land per Asentamiento Socios per Asentamiento Land per Socio Nationwide 212.98 (154.35) 22.44 (8.8) 9.04 (2.76) Sample 90.52 (53.35) 16.04 (9.75) 8.27 (5.86) The Surveys Quantifiable production and demographic data were gathered in the regions of Olancho and El Parai'so, Honduras. General information was also gathered as a byproduct of the numerical data obtained for statistical analysis. These regions share similar terrain and cultural characteristics. Data were recorded on the entire gamut of cooperative operations. The human survey instrument (appendix) was designed to obtain information regarding the health and educational background of each HARC member, his living conditions and social affiliations with the cooperative. The purpose of collecting data regarding the personal characteristics of HARC members and facets of social interaction was to distinguish features that influence HARC efficiency and operations. Short to long-term endogenous factors were surveyed to facilitate the delineation of policies that improve HARC efficiency and performance. Besides short-term IPM training techniques, longer term components such as schooling, housing and health data were gathered to amplify the explanatory power of statistical modeling. Exogenous factors such as age, family size and types of social interaction were also included to render information regarding the

PAGE 198

186 responsiveness of characteristic clusters and thus determine, apriori, which groups are more inclined to operate efficiently and are relatively more amenable to extension services. A test regarding integrated pest management techniques was £lso administered in the survey to examine the effectiveness of the varying training techniques tc which the cooperatives were exposed. Farmers were never asked directly if they adopted a given technology, as such questions are known to bias responses. Rather, they were asked open-ended questions free of suggestions about their production and pest management practices. Responses were scored as correct only if they conformed to the instructions given by MIPH. All individual data take the household parcel or parcels, not the household itself, as the unit for which all inputs and activities were measured. Financial resources simply did not permit measurement of all household activities. The argument could be made that competing household production activities will bear an influence on efficiency and adoption measurements. The relative degree of importance that the individual parcel plays in the household utility function could vary, rendering the coefficients vulnerable to bias and an overestimation of the disturbance variance. However, it is reasonable to assume that opportunities available to agrarian reform households are minimal. If they were not, they could not participate in the reform. The only activity that may compete with other activities is coffee harvesting. But this occurs primarily in the first three months of the year, when most basic grains have been harvested. Moreover, given the intensity of land disputes, the efficiency of land units, not household units, assumes primary importance. The number of outside economic endeavors is documentet! in the human resource survey and may, through the inclusion of dummy variables, account for some efficiency variation. Observations were made on cooperative operations throughout the data gathering period. It was necessary to attend dozens of cooperative meetings and maintain close contact with support agencies in order to obtain data and information. Close association with the cooperatives and

PAGE 199

187 cooperative agencies provided an excellent opportunity to observe how actual HARC performance deviated from desired HARC performance. Amiable assistance from MIPH and government agronomists was invaluable in corroborating data to account for discrepancies and ensure data reliability. Such observations will be useful in interpreting statistical results.

PAGE 200

APPENDIX II FIELD SURVEY ENCUESTA DE RECURSOS HUMANOS Grupo Encuestador: Fecha / / Hora : I INFORMACION GENERAL \. Nombre 2. Edad 4. Dirreccidn 5. Estado civil Casado Soltero Viudo 6. ^Donde Nacid? Jamastran El Sur Guayape Linaca Otro: 8. Sf no naci(5 en esta zona: iQuQ afio lleg(3 aquf a esta zona? 9. Ano que se uni(3 el grupo 10. Cargo en el grupo desde 11. Cargos que ha tenido: Presidente: aiios Vice Presidente: aiios Ttisorero: aiios Secrataria: aiios 188

PAGE 201

189 12. (,Antes de hacerse socio del grupo, que hacia usted para ganar la vida? . Asdariado Alquilaba tierra Otro: 13. (,Cree usted que ser miembro del grupo aumenta(A), baja (B) o no cambia (N) lo siguiente?: Seguridad Consumo de comida Sueldo Dignidad 14. Numero de miembros de la casa: RelaNombre ci(3n Edad Educ Pep Nombre 1 2 3 4 5 6 10 11 12 * HO = hijo, HNO = hermano II. PRODUCCION 1. ^.Cuantos anos tiene de trabajar con Relacidn Edad Educ Pep J L semillas mejoradas en bolsa? herbicidas? veneno de plagas? abono? ;Relcalque mucho que las preguntas 2-24 referien solamente a la arcela individual por la cual el archivero llevo el control! ;Ubique al asociado al:a mentalmente en una manera conversacioal y por cada pregunta refieranse a esa parcela! 2. Por cada uno de los siguientes insumos preguntele al asociado si utilizd el producto o no en su parcela individual Si contesta "no" escriba "no" a la derecha del colon, y si contesta "sf" ponga el numero de aiios tiene de trabajar con tal producto a la derecha del colon

PAGE 202

190 3. (,Utiliz(5 otros insumos que no estan mencionado arriba? Sf No Si contesta "si." ,;,CuaIes son? : 9 : 9 : > 4. ^Como prepard la tierra de su parcela individual? Tractor Bueyes Azaddn Barreta 5. (.Cuantos pases de arado hizo en primera en su parcela individual?. 6. (.Cuantos pases de rastra hizo en primera en su parcela individual?. 7. iQue tipo de arado utiliz6 en su parcela individual? Disco Vertedera Row plow De palo 8. iCon que surque<3 la labranza? Bueyes? Maquina? No surqued 9. ^En su milpa individual que distancia tiene entre: surcos? posturas? 10. i,Cuantas semillas puso por postura (por ej. 2:3 O 3:3) en su parcela individual en: maiz? : friiol? : 1 1 . En su parcela individual sembrd frijoles en Primera Postrera 12. Para usted: ^Cuantas barras cuadradas tiene una tarea? iQUE NO LEA A LOS ASOCIADOS LAS RESPUESTAS DE PREGUNTAS SOBRE LAS PLAGAS #13 #22 (O SEA JUEGOS DE RESPUESTAS QUE SON TRAZADAS!) DEJE LA PREGUNTA ABIERTA PARA QUE CONTESTE EN SU PROPIA MANERA. 13. Sigamos en la misma parcela individual. /Antes de encontrar algunas Gallinas Ciegas este ano, hizo algo usted para prevenirlas ? No hizo nada Prepard bien el suelo Aplicd veneno Otro:

PAGE 203

191 i,La Gallina ciega infestd esa parcela este ano? Sf No Si contesta "No" pasen a #14 ^Hizo muestreo? Sf No ^Cuantas tareas de a _ se le infestd? ^,Como la combatid usted? No hizo nada Prepard bien el suelo Aplicd veneno Otro: ^.Logrd controlaria? Sf No 14. Sigamos en la misma parcela individual. (,Antes de encontrar algunos Cogolleros este ano. hizo algo usted para prevenirlos? No hizo nada Lluvia los matd Aplicd veneno. Otro: ^El Cogollero infestd esa parcela este ano? Sf No Si contesta "No" pasen a #15 ^Hizo muestreo? Sf No (.Cuantas tareas de a _ se le infestd? i,Como lo atacd usted? No hizo nada Lluvia lo matd Aplicd veneno Otro: i,Logrd controlarlo? Si No 15. Sigamos en la misma parcela individual. ( Antes de encontrar algunas langostas este ano, hizo algo usted para prevenirlas? No hizo nada Cruzd la milpa Controld zacate Aplicd veneno aro: iha Langosta (Gusano Medidior)infestd esa parcela este ano? Sf No Si contesta "No" pasen a #16 (,Hizo muestreo? Sf No ^.Cuantas tareas de a _ se le infestd? ^Como la atacd usted? No hizo nada Cruzd la milpa Controld zacate Aplicd veneno. Otro: (.Logrd controlaria? Sf No

PAGE 204

192 16. Sigamos en la misma parcela individual. (;,Antes de encontrar algunas Babosas este aiio, hizo algo usted para prevenirlas ? ;Cuando? . . . . No hizo nada Cebo Controld malezas Basura trampa Matanza nocturna Prepard terreno Quema rapido Otro: En primera En postrera iLa Babosa infestd esa parcela este aiio?Si No Si contesta "No" pasen a #17 ^Hizo muestreo? Sf No ^Cuantas tareas de a _ se le infestd? (.Corno la atacd usted? No hizo nada Cebo Controld malezas Basura trampa Matanza nocturna Quema rapido Otro: ^.Logrd controlarla? Si No 17. Sigamos en la misma parcela individual. ( Antes de encontrar algunas tortugillas este aiio, hizo algo usted para prevenirlas? No hizo nada Prepard bien el suelo Insecticida Otro: La Tortugilla infestd esa parcela este aiio? SI No Si contesta "No" pasen a #18 ^Hizo muestreo? SI No i,Cuantas tareas de a _ se le infestd? (,Como la atacd usted? No hizo nada Insecticida Otro: t,Logrd controlarla? Si No

PAGE 205

193 18. Sigamos en la misma parcela individual. ^ Antes de encontrar algunas empoascas este aiio, hizo algo usted para prevenirlas ? No hizo nada Lluvia la maX6 Insecticida Otro: ;,La Empoasca infestd esa parcela este ano? Sf No Si contesta "No" pasen a #19 ^Hizo muestreo? Sf No ^Cuantas tareas de a _ se le infest(5? ^Como la atac6 usted? No hizo nada Lluvia la mat(3 Insecticida Otro: ^,Logr6 controlarla? Sf No 19. Sigamos en la misma parcela individual. (,Antes de encontrar algunas picudos de la vaina del frijol este aiio, hizo algo usted para prevenirlos? No hizo nada Controld malezas Destruyd rastrojos Control quimico Otro: iEl picudo de la vaina del frijol infestd esa parcela este afio? Sf No Si contesta "No" pasen a #20 iWizo muestreo? Sf No ^Cuantas tareas de a _ se le infestd? ^Como lo atac6 usted? No hizo nada Controld malezas Distruyd rastrojos Control quimico Otro: (.Cuando lo hizo? t,Logrd controlarlo? Sf No 20. iOtro plaga? ^.Cuantas tareas de a _ se le infestd? ^Como lo controls?

PAGE 206

194 iOtro plaga? (.Cuantas tareas de a _ se le infest(3? ^Como lo control^? 21. Sigamos en la misma parcela individual. ^El zacate infestd esa parcela en el tiempo critico del desarroUo del mafz Sf No 22. Sigamos en la misma parcela individual. ^EI zacate infestd esa parcela en el tiempo critico del desarrollo del frijol Sf No 23. Sigamos en la misma parcela individual. i,La hoja ancha (monte o hierba) infest(5 esa parcela en el tiempo critico del desarrollo del mafz Sf No 24. Sigamos en la misma parcela individual. t,La hoja ancha (monte o hierba) infesto esa parcela en el tiempo critico del desarrollo del frijol Sf No 25. (.Que opina usted? (,Cual m^todo de produccidn pretlere en la cooperativa? Colectivo Individual Los dos III. EDUCACION DEL SOCIO 1. Anos cumplidos en la escuela primaria secundaria nocturna 2. ^Participd en la campaiia de alfabetizacidn? Sf No 3. ^Sabe leer usted (como periodico)? Sf No 4. ^Sabe escribir usted? Sf No Si contesta "sf" a #3 o #4: i,Donde lo aprendi(3? 5. i,Sabe su madre leer? Sf No i,Sabe su padre leer? Sf No 6. Cursos de capacitacidn que ha recibido en cuanto a: cooperativismo. agricultura

PAGE 207

195 8. ^,Usted ya sabe mucho de la agricultura, de donde ha aprendido lo mas? INA Parientes Trabajando en finques grandes RR.NN. Experiencia Buenos agricultores MIPH Socios Otro: 9. i,De donde ha aprendido mas sobre el manejo de plagas? INA Parientes Trabajando en finques grandes RR.NN. Experiencia Buenos agricultores MIPH Socios Otro: IV. AFILIACIONES SOCIALES 1. (,Tiene usted parientes quienes son socios del grupo? Sf No Numero de: hermanos cunados primos suegros hijos yernos padres tios sobrinos 2. (.Pertenece usted a alguna religion? Sf No si contesta "no" pasen a #3 ^.Cual religion? Catdlico Evangel ico Otro 3. (,Participa usted en actividades deportivas con otros socios del grupo? Sf No si contesta "no" pasen a #4 i,Cuales depones? Futbdl Dados Naipes Billar Otros 4. ^Participa usted en actividades musicales con otros socios del grupo? Sf No si contesta "no" pasen a #5 5. ^Participa usted en actividades religiosas con otros socios del grupo? Sf No i,Cual religion? Catdlico Evangel ico Otro_

PAGE 208

196 6. t,Tiene hijos que participan en actividades con hijos de los otros socios del grupo? Nada Deportes Religion Escuela Otro: 7. (,Puede identificar usted el miembro del grupo quien tiene la mayor influencia (no necesariamente un oficial)? Si No Si contesta "sf:" ^.Quien es? V. MEDIDAS DE SALUD 1 i,De donde consigue el agua de la casa? Rio Pozo Llave Quebrada Otro: 2 iSt hierve el agua que se bebe en su casa? Siempre A veces Solo bebidas calientes 3 (.Que tipo de piso tiene su casa? Tierra Cemento Madera Otro 4 iQue tipo de pared tiene su casa? Adobe Madera Bloque Ladrillo 5 (.Que tipo de techo tiene su casa? Texas Asbestas Zinc Zacate 6 (.Tiene huerto de hortalizas? Sf No i.Cuales son? papa chile repollo cebolla rabano tomate yuca remolacha otro: pepino zanahoria otro:

PAGE 209

7 ^Tiene usted arboles de frutas? i,Cuales son? guineos aguacate mango naranja toronja maranon otro: otro: otro: 8 (,Como se guarda maiz en su casa? 9 (,Como se guarda frijoles en su casa? 10 ^Como se guarda arroz en su casa? 1 1 i,Cuantos dlas de trabajo perdid usted este ano por enfermedades? (Esto incluye enfermedades de usted o de un pariente o amigo a quien usted tuvo que cuidar) 13 ^,Cuantos animal es tiene? toros , vacas , bueyes , cerdos , burros , polios , caballos , pavos , perros , mulas 14 ^Ademas de la siembra tiene otros ingresos?

PAGE 210

198 VI. MATRICES DE COLORES: Vista %, + Hora: 1_ 2_ 3_ 4_ 5_ 6_ 7_ 8_ 9_ 10_ 11_ 12 Ab B Hora: Altura cm Peso libras Brazo cm Pierna cm

PAGE 211

REFERENCES Abler, D.G. and J.S. Shortle (1995). "Technology as an agricultural pollution control policy." American Journal of Agricultural Economics, 77(1) pp. 20-32. Abrahamsen, M.A. (1976). Cooperative Business Enterprise, New York: McGraw-Hill. Aigner, D.J. and S.F. Chu (1968). "Estimating the industry production function." American Economic Review, 58, pp. 826-39. Aigner, D.J., C.A.K. Lovell and P.J. Schmidt (1977). "Formulation and Estimation of Stochastic Frontier Production Function Models." Journal of Econometrics, 6, pp. 2137. Alchian, A. A. and H. Demsetz (1972). "Information Costs, and Economic Organization." American Economic Review, 62(5), pp. 777-95. Amemiya, T. (1981). "Qualitative Response Models: A Survey" Journal of Economic Literature, 19, pp. 1483-1536. Asian Productivity Organization (1991). Agricultural Cooperatives in Asia and the Pacific, Report of APO Multi-Country Study Mission, Tokyo. Banker, R.D., A. Charnes and W.W. Cooper (1984). "Some Models for Estimating Technical and Scale Inefficiencies in Data Envelopment Analysis." Management Science, 30(9), pp. 1078-1092. Banker, R.D. R.F. Conrad and R.P. Strauss (1986). "A Comparative Application of DEA and Translog Methods: An Illustrative Study of Hospital Production " Management Science, 32. pp. 30-44. Bardhan, P. (1989). "Alternative Approaches to the Theory of Institutions in Economic Development." in The Economic Theory of Agrarian Institutions, Ed. P. Bardhan, Oxford. England. Clarendon Press, pp. 3-17. Barham, B.L., and M. Childress (1992). "Membership Desertion as an Adjustment Process on Honduran Agrarian Reform Enterprises." Economic Development and Cultural Change, 40(3), pp.587-613. Barnadas, J.M. (1986). "The Catholic Church in Colonial Spanish America." The Cambridge History of Latin America ed. Leslie Bethell, Cambridge: Cambridge University Press. 199

PAGE 212

200 Barnard, C.I. (1938). The Functions of the Executive Cambridge, Mass., Harvard University Press. Barro, R.J., and Sala-i-Martin. X. (1992), 'Public finance in models of economic growth', Review of Economic Studies 59, pp. 645-61. Barry, T. and K. Norsworthy (1990). Honduras, A country Guide, Albuquerque, New Mexico: The Interhemispheric Education Resource Center. Bartlett, W., J. Cable, S. Estrin. D.C. Jones, and S.C. Smith (1992). "Labor-managed cooperatives and private firms in North Central Italy: an empirical comparison." Industrial and Labor Relations Review, 46(1) pp. 103-18. Battese, G.E.. and T.J. Coelli (1992). "Frontier Production Functions, Technical Efficiency and Panel Data: With Application to Paddy Farmers in India." J. Productivity Anal, 3 pp. 153-69. Battese, G.E., T.J. Coelli, and T.C. Colby (1989). "Estimation of Frontier Production Functions and the Efficiencies of Indian Farms Using Panel Data from ICRISAT's Village Level Studies." Journal of Quantitative Economics, 5(2), pp. 327-48. Battese, G. and G. Corra (1977). "Estimation of a Production Frontier Model: With Application to the Pastoral Zone of Eastern Australia." Australian Journal of Agricultural Economics , 21(3), pp. 169-79. Bauer, P.W. (1990). "Recent Developments in the Econometric Estimation of Frontiers." Journal of Econometrics , 46, pp. 39-56. Becker, G.S. (1962). "Investment in Human Capital: A Theoretical Analysis." Journal of Political Economy, 70, pp. 9-49. Becker, G. and N. Tomes (1984). Human Capital and the Rise and Fall of Families Economics Research Center/NORC Discussion Paper. (1976). "Child Endowments and the Quantity and Quality of Children." The Journal of Political Economy, 84 pp. s 143-62. Becker, Gary S. and B.R. Chiswick, (1966). "Education and the Distribution of Earnings." American Economic Review, May. Behrman, J.R., B.L. Wolf and D.M. Blau (1985). "Human Capital and Earnings Distribution in a Developing Country: The Case of Prerevolutionary Nicaraj^ua." Economic Development and Cultural Change, 34, pp. 1-29 Behrman, J. and B. Wolfe (1984). "More Evidence on Nutrition Demand: Income Seems Overrated and Women's Schooling Underemphasized." Journal of Development Economics, 14(1-2), pp. 105-28.

PAGE 213

201 B.R. Albert and W.R. Cline (1979). Agrarian Structure and Productivity in Developing Countries, Baltimore: The Johns Hopkins Press. Birkhaeuser, D., R.E. Evenson and G. Feder (1991). "The Economic Impact of Agricultural Extension: A Review." Economic Development and Cultural Change, 39(3) pp. 60750. Blaug, M. (1976). "The Empirical Status of Human Capital Theory: A Sightly Jaundiced View." Journal of Economic Literature, 14, pp. 827-55. Bliss, C. and N. Stern (1978). "Productivity, Wages and Nutrition." Journal of Development Economics, 5, pp. 331-98 Bonin, L (1987). Economics of Cooperation and the Labor Managed Economy, Fundamentals of Pure and Applied Economics. Zurich: Harwood Economic Publishers. Bravo-Ureta, B.E., and L. Rieger. (1991). "Dairy Farm Efficiency Measurement Using Stochastic Frontiers and Neoclassical Duality." American Journal of Agricultural Economics. 73, pp. 421-28. Bravo-Ureta, B.E., and R. Evenson (1991). "Efficiency in Agricultural Production: The Case of Peasant Farmers in Eastern Paraguay." Agricultural Economics, 10, pp. 27-37. Brading, D.A. (1986). "Bourbon Spain and its American Empire." The Cambridge History of Latin America ed. Leslie Bethell, Cambridge: Cambridge University Press. Brenner, R. and N. Kiefer (1981). "The Economics of Diaspora: Discrimination and Occupational Structure." Economic Development and Cultural Change,29i3), pp. 51734. Bromley, Daniel, (1989). Economic Interests and Institutions, New York, New York: Basil Blackwell. Bueso, J. A. (1987). El Subdesarollo Hondureflo, Tegucigalpa, Hondures: Editorial Universitaria. Burrows, T.M. (1983). "Pesticide Demand and Integrated Pest Management: A Limited Dependent Variable Analysis." American Journal of Agricultural Economics, 65, pp. 806-810. CEDOH (Centro de Documentaci(5n de Honduras) (1992). Puntos de Vista: Temas Agrarios, Tegucigalpa: Linea Grdfica S. de R.L.. Chandler, A.D. (1990). Scale and Scope : the Dynamics of Industrial Capitalism Cambridge, Mass.: Belknap Press.

PAGE 214

202 Charnes, A,. W.W. Cooper and E. Rhoades (1978). "Measuring the Efficiency of Decision Making Units, European Journal of Operational Research, 2, pp. 429-444. Charnes, A,. W.W. Cooper and E. Rhoades (1981). "Evaluating Program and Managerial Efficiency: An Application to Data Envelopment Analysis to Program FollowThrough." Management Science, 27, pp. 668-697. Chiswick, B.R. (1983). "The Earnings and Human Capital of American Jews." The Journal of Human Resources. 18(3), pp. 313-36. Coase, R.H. (1937). "The Nature of the Firm." Economica, 4, pp. 386-405. Coase, R.H. (1991). "The Institutional Structure of Production." Nobel Lecture, December 9. Cobia, D. (1989). Cooperatives in Agriculture. Englewood Cliffs, NJ: Prentice Hall. Comes, R. (1992). Duality and Modern Economics, New York and Melbourne: Cambridge University Press. Davis, L.E. and D. North, (1971). Institutional Change and American Economic Growth, Cambridge, England: Cambridge University Press de Janvry A. (1981). The Agrarian Question and Reformism in Latin America, Baltimore: The Johns Hopkins University Press. Domenich, T.A. and D. McFadden (1975). Urban Travel Demand-A Behavioral Analysis, Amsterdam: North-Holland Publishing Co. Dorner, P. and D. Kanel (1971). "The Economic Case for Land Reform." In Land Reform in Latin America: Issues and Cases, ed. by Peter Dorner, pp. 41-56, Land Economics Monograph Series, no. 3 Madison: Published by Land Economics for the Land Tenure Center at the University of Wisconsin. Dublin, L. and Lotka, A. (1930). The Monetary Value of Man, New York: Ronald Press. Europa Yearbook (1991). London: Europa Publications. Fare, R., S. Grosskopf. and C.A.K. Lovell (1985). The Measurement of Efficiency of Production, Boston: Kluwer-Nijhoff Publishing. Fare. R.S., S. Grosskopf and C.A.K. Lovell, 1987, "Nonparametric Disposability Tests." Journal of Economics 47(1), pp. 77-85. Farrel, M. (1957). "The Measurement of Production Efficiency." Journal of the Royal Statistical Society (120) 253-81.

PAGE 215

203 Friedman, M.S.. and S. Kuznets (1945). Income from Professional Practice, New York: National Bureau of Economic Research. Furubotn, E.G., and S. Pejovich. (1970). "Property Rights and the Behavior of the Firm in a Socialist State: The Example of Yugoslavia." Zeitschrift fur Nationalokonomie, 30, pp. 431-54. (1972). "Property Rights and Economic Theory: A Survey of Recent Literature." Journal of Economic Literature, 4, pp. 1137-162. Furubotn, E.G., and R. Richter. (1991). "The New Institutional Economics: New Views on Antitrust." Journal of Institutional and Theoretical Economics, 147, pp. 1-6. Fersund, F.R., C.A. Knox Lovell and P. Schmidt (1980). "A Survey of Frontier Production Functions and of their relationship to Efficiency Measurement. " Journal of Econometrics, 13(1) pp. 5-25. Frank, A.G. (1979). Mexican Agriculture, 1521-1630: Transformation of the Mode of Production, Cambridge: Cambridge University Press. Garcfa, M., R. Norton, M. Ponce and R. van Haeften (1988). Agricultural Development Policies in Honduras: A Consumption Perspective, Washington, D.C.: United States Department of Agriculture. GREAN (Global Research on the Environmental and Agricultural Nexui) (1996). Taskforce on Research Innovations for Productivity and Sustainability, Gainesville, University of Florida. Greene, W. (1980). Econometric Analysis, New York: Macmillan Publ shing Company. Griliches, Z. (1964). "Research Expenditures, Education, and the Aggregate Agricultural Production Function." American Economic Review, 54, pp. 961-74. Guevara-Escudero, J. (1983). Nineteenth Century Honduras: A Regional Approach to the Economic History of Central America, Dissertation, New York; New York University. Harper, J.K., M.E. Rister, J.W. Mjelde, B.M. Drees and M.O. Way (1990). "Factors Influencing the Adoption of Insect Management Technology." American-Journal-of-Agricultural-Economics, 72(4), pp. 997-1005. Hayes, M.D. (1988). "The U.S. and Latin America: A Lost Decade?" Foreign Affairs, 68, pp. 180-199. Helliwell, J.F. and R.D. Putnam (1995). "Economic growth and social capital in Italy." Eastern Economic Journal 21(3) pp. 295-307.

PAGE 216

204 Huffman, W.E. (1977). "Allocative Efficiency: The Role of Human Capital." Quarterly Journal of Economics and Statistics, 91, pp. 59-79. Huffman, W.E. (1974). "Decision Making; The Role of Education." American Journal of Agricultural Economics, 54, pp. 86-97. Immink, M.D., F. Viteri and R. Helms (1982). "Energy Intake Over the Life Cycle and Human Capital Formation in Guatemalan Sugar Cane Cutters." Economic Development and Cultural Change, 30(2), pp. 351-72. Institute Hondureiio de Desarroiio Rural, 1980, 84 Meses de Reforma Agraria del Gobierno de las Fuerzas Armadas de Honduras, Tegucigalpa. IN A (Institute Nacional Agrario) (1985). Resumen Bdsico de Los Grupos Campesinos Beneficiarios de la Reforma Agraria, Tegucigalpa, Honduras, Departamento de Planificaci(3n, Seccidn de Estadfstica e Informacidn. Jamison, D.T. and L.J.Lau (1982). Farmer Education and Farmer Efficiency, Baltimore: The Johns Hopkins University Press. Jamison, D.T. and M.E. Lockheed (1987). "Participation in Schooling: Determinants and learning Outcomes in Nepal." Economic Development and Cultural Change, 35(2), pp. 279-306. Jamison, D.T. (1986). "Child Malnutrition and School Performance in China." Journal of Development Economics, 20(2), pp. 299-309. Jensen, M. C. and W. H. Meckling (1979). "Rights and Production Functions: An Application to Labor-Managed Firms and Codetermination. " Journal of Business, 4, pp. 469-506. Jondrow, J., C.A.K. Lovell, I.S. Materov, and P. Schmidt. (1982). "On the Estimation of Technical Inefficiency in Stochastic Frontier Production Function Model." Journal of Econometrics 19, pp. 233-38. Kalirajan, K.P. (1990). "On Measuring Economic Efficiency, Journal of Applied Econometrics, 5(1), pp. 75-85. Kalirajan, K.P. (1984). "Farm Specific Technical Efficiencies and Development Policies." Journal of Economic Studies. 11, pp.313. Kalirajan, K.P. and J.C. Finn (1983). "The Measurement of Farm-Specific Technical Efficiency." Pakistan Journal of Applied Economics, 2, pp. 167-80. Kalirajan, K.P. and R.T. Shand (1985). "Types of Education and Agricultural Productivity: A Quantitative Analysis of Tamil Nadu Rice Farming." The Journal of Development Studies, 21(2), pp. 232-43.

PAGE 217

205 Khaldi, N. (1975). "Education and Allocative Efficiency in U.S. Agriculture" American Journal of Agricultural Economics, 57(4), pp. 650-657. Koopmans, T.C. (1951). Activity Analysis of Production and Allocation , Cowles Commission for Research in Economics Monograph No. 13, New York, John Wiley and Sons. Kopp, R.J., and W.E. Diewert (1982). "The Decomposition of Frontier Cost Function Deviations into Measures of Technical and Allocative Efficiency," Journal of Econometrics, 19, pp. 319-331. Kopp, R.J. and V.K. Smith (1980). "Frontier Production Function Estimates for Steam Electric Generation: a Comparative Analysis." Southern Economic Journal, 47, pp. 1049-1059. Kurian, G.T. (1987). Encyclopedia of the Third World, New York: Facts on File Inc. Lele, U. (1981). "Co-operatives and the Poor: A Comparative Perspective" World-Development 9, pp. 55-72. Lele, U. (1974). "Role of Credit and Marketing in Agricultural Development." in Nurul Islam (ed.) Agricultural Policy in Developing Countries, London: Macmillan. Liebenstein, H. (1957). Economic Backwardness and Economic Growth, New York: Wiley. MacLeod, M.J. (1983). "Ethnic Relations and Indian Society in the Province of Guatemala, ca. 1620-ca. 1800" in Spaniards and Indians in Southeastern Mesoamerica : Essays on History of Ethnic Relations edited by Murdo J. MacLeod and Robert Wasserstrom. Lincoln: University of Nebraska Press. Macneil, I.R. (1974). "The Many Futures of Contracts." Southern California Law Review, 47 pp. 691-816. Maddala, G.S. (1983). Limited Dependent and Qualitative Variables in Econometrics, Englewood, N.J. Cambridge University Press. Marshall, A. (1920). Principles of Economics, London: Macmillan. Martin, M.J. and T.G. Taylor (1995). "Evaluation of a Multimedia Extension Program in Honduras." Economic Development and Cultural Change 43, pp. 821-834. Martin^z-Peldez, S. (1975). La patria del Criollo, San Jos6: Editorial Lniversitaria Centroamericana. Mccommon, C.N. Rueschhoff. L.T. and J. Wilkowski (1985). Guanchias Limitada: a Case Study of an Agrarian Reform Cooperative and its Long-term Relationship with a Multinational Firm in Honduras, Washington, D.C.: U.S. Agency for International Development.

PAGE 218

206 Meeusen, W. and J. van den Broeck (1977). "Efficiency estimation from CobbDouglas Production Functions with Composed Error." International Economic Review, 18, pp. 435-444. Meister, R.T. (1980). "Pest Management CoopsThe Small Grower's Answer to the IPM Puzzle?" Agricultural Consultant and Fieldman, 36, pp. 8-11. Meng, R. and J. Sentance (1984). "Religion and the Determination of Earnings: Further Results." Canadian Journal of Economics, 17(3), pp. 481-88. Milon, W. (1987). "The Science and Art of Efficiency Analysis: The Role of Other Performance Criteria," in Economic Efficiency in Agricultural and Food Marketing, Kilmer and Armbruster eds. . Ames, Iowa: Iowa State University Press for the Farm Foundation and the Institute of Food and Agricultural Sciences of the University of Florida, pp. 67-87. Mincer, J. (1989). Job Training: Costs, Returns, and Wage Profiles National Bureau of Economic Research Working Paper: 3208. Mincer, J. (1962). "On the Job Training: Costs, Returns and Some Implications." Journal of Political Economy, 70, pp. 50-79. (1958). "Investment in Human Capital and Personal Income Distribution." The Journal of Political Economy, 60, pp. 281-302 Morris, J. A. (1984). Honduras : Caudillo Politics and Military Rulers, Boulder: Westview Press. North, D. (1992). "Institutions and economic theory" American Economist, 36(1), pp. 3-6. (1991). "Institutions." Journal of Economic Perspectives, 5, pp. 97-112 Pinheiro, A. (1992). An Econometric Analysis of Farm Level Efficiency of Small Farms in the Dominican Republic, M.S. Thesis, University of Connecticut. Pudasaini, S.P. (1983). "The Effects of Education in Agriculture: Evidence fi-om Nepal." American Journal of Agricultural Economics, 65(3), pp. 509-15. Monthly Review, December (1985). p. 25. Morley, S.G. and G. Brainerd (1983). The Ancient Maya, 4th ed. / revised by Robert J. Sharer, Stanford, Calif.: Stanford University Press. Moock, P.R. and J. Leslie (1986). "Childhood Malnutrition and Schooling in the Terai Region of Nepal." Journal of Development Economics, 20(1), pp. 33-52.

PAGE 219

207 Miiller, J. (1974). "On Sources of Measured Technical Efficiency: The Impact of Information." American Journal of Agricultural Economics, 55, pp. 730-38. Newson, Linda A. (1986). The Cost of Conquest : Indian Decline in Honduras Under Spanish Rule, Boulder: Westview Press. Nguyen, D.T. and M.L. Martinez-Saldivar (1979). "The Effects of Land Reform on Agricultural Production, Employment and Income Distribution: A Statistical Study of Mexican States, 1959-69" Economic Journal, 89, pp. 624-35. Palanigounder, D. (1989). Human Capital and Adoption of Innovations in Agricultural Production: Indian Evidence Yale Economic Growth Center Discussion Paper: 577. Paus. Eva, ed. (1988). Struggle against dependence: Nontraditional export growth in Central America and the Caribbean. Series in Political Economy and Economic Development in Latin America, Boulder, Colo, and London: Westview Press. Piesse, J., C. Thirtle and J. Turk (1996). "Efficiency and Ownership in Slovene Dairying: A Comparison of Econometric and Programming Techniques." Journal of Comparative Economics, 22, pp. 1-22. Posas, M. (1981). Conflictos Agrariosy Organizacion Campesina: Sobre los Origenes de las Primeras Organizaciones Campesinas en Honduras, Tegucigalpa, Honduras: Editorial Universitaria. Posas, M. (1987). El Subdesarrollo Hondureflo, Tegucigalpa, Hondura;.: Editorial Universitaria. Prychitko, D. and J. Vanek (1996). Producer Cooperatives and Labor Managed Systems, England; Edgar Elgar Publishing Limited. Putnam, R. D. (1993). Making Democracy Work: Civic Traditions in Modern Italy, Princeton: Princeton University Press. Putterman, L. (1986). Peasants, Collectives and Choice: Economic Theory and Tanzania's Villages, Greenwich, Conn: JAI Press. (1989). "Agricultural Producer Co-operatives," in The Economic Theory of Agrarian Institutions, Ed. Pranab Bardhan, Oxford, England: Clarendon Press, pp. 319-39. Quiiiones, E. and M. Argueta (1978). Historiade Honduras, Tegucigalpa, Honduras: Escuela Superior del Profesorado "Francisco Morazan." Romer, P. (1990). "Endogenous Technological Change' Journal of PoMcal Economy, 98, pp. S71-102.

PAGE 220

208 Rook, S.P. and G.A. Carlson (1985). "Participation Pest Management Groups" American Journal of Agricultural Economics, 67(3), pp. 563-66. Roy, K. and C. Clark, eds. (1994). Technological change and rural development in poor countries: Neglected issues, Oxford and New York: Oxford University Press. Ruhl, M. (1989). "The Honduran Agrarian Reform Under Suazo Cdrdova, 1982-85; An Assessment, Inter-American Economic Affairs, 39, pp. 63-80. Schaible, G.D., C.S. Kim, and N.K. Whittlesey (1991). "Water Conservation Potential from Irrigation Technology Transitions in the Pacific Northwest. " Western Journal of Agricultural Economics, 16 pp. 194-206. Schmid, A. (1972). "Analytical Institutional Economics." American Journal of Agricultural Economics. 54. pp. 893-901. Schmidt, P. (1976). "On the Statistical Estimation of Parametric Frontier Production Functions." Review of Economics and Statistics, 58, pp. 238-39. Schotter, A. (1981). The Economic Theory of Social Institutions, New York: Cambridge University Press. Schultz, T.W. (1975). "The Value of the Ability to Deal with Disequilibria." Journal of Economic Literature, 13(3), pp. 827-46. (1964). Transforming Traditional Agriculture, New Haven: Yale University Press, 1964. (1963). The Economic Value of Education, New York, New York: Columbia University Press. Seale, J.M. (1990). "Estimating Stochastic Frontier Systems with Unbalanced Panel Data: The Case of Floor Tile Manufactories in Egypt. " Journal of Applied Econometrics, 5 pp 59-74. Seiford, L. and R.M. Thrall (1990). "Recent Developments in DEA" Journal of Econometrics, 46, pp. 7-38. Seligsan, M. and E. Nessman, (1989). Land Titling in Honduras: An Im.pact Study in the Comayagua Region, unpublished manuscript. Shapiro, K.H. and J. Muller (1977). "Sources of Technical Efficiency: The Roles of Modernization and Information." Economic Development and Cultural Change 25(2) pp. 293-310. Sheffield, J.R., (1975). Retention of Literacy and Basic Skills, (Paper prepared for Education Department of the World Bank.

PAGE 221

209 Shephard, R. W. (1970). Tlieory of Cost and Production Functions, Princeton: Princeton University Press. (1953). Cost and Production Functions, Princeton: Princeton University Press. SITRAUNAH (Sindicato de Trabajadores de la Universidad Nacional Autonoma de Honduras) (1979). Conjura Antisindical Contra Honduras, Tegucigalpa, D.C., Honduras. Smith, A. (1892). An Inquiry into the Nature and Causes of the Wealth of Nations, London: Routledge. Spence, L.H. (1993). "Rethinking the Social Role of Public Housing" Housing Policy Debate 4, pp. 355-68. Stokes, W.S. (1947). "The Land Laws of Honduras." Agricultural History , 21, pp. 148-54. Stringer, R. (1984). An Analysis of Credit Use in the Honduran Agrarian Reform Sector, University of Wisconsin, Ph.D. Thesis. (1989). "Honduras: Toward Conflict and Agrarian Reform" in Searching for Agrarian Reform in Latin America. Ed. William C. Thiesenhusen, BostO'i: Unwin Hyman Inc pp. 358-383.. Taylor, T.G., H.E. Drummond and A.T. Gomes (1986). "Agricultural Credit Programs and Production Efficiency: An Analysis of Traditional Farming in Southeastern Minas Gerais, Brazil." American Journal of Agricultural Economics, 58, pp. 110-19. Toman, M. A. (1994). "Resources for the Future" Land-Economics, 70, pp. 399-413. Tomes, N. (1985). "Human Capital and Culture: Analysis of Variation m Labor Market Performance." AEA Papers and Proceedings, May. Torquemada, Juan de, ca. 1557-1664, Monarquia Indiana, por fray Juan de Torquemada. 3. ed. Mexico, D. F.. S. Chavez Hayhoe. 1943. USAID (U.S. Agency for International Development) (1982). AID Policy Paper: Private Enterprise Development, Washington, D.C.. (1978). Honduran Agricultural Assessment Drafts, Washington, D.C. . (USITC) (United States International Trade Commission) (1992). Potential Effects of a North American Free Trade Agreement on Apparel Investment in CBERA Countries USITC Publication 2541. Varian, H.L., (1984). "The Nonparametric Approach to Production Aailysis" Ecnometrica 52 pp. 579-97.

PAGE 222

210 Vaughan, R. (1989). Education, Training and Labor Markets: Summary and Policy Implications of Recent Research by Jacob Mincer, Conference Paper Series, The Institute on Education and the Economy, Teachers College, Columbia University, New York. Veblen, T. (1899). The Tlieory of the Leisure Class; an Economic Study in the Evolution of Institutions. New York: The Macmillan company: London, Macmillan & co. Villanueva, B. (1968). Institutional Innovations and Economic Development. Honduras: A Case Study, University of Wisconsin, Ph.D. Thesis Walsh, J.R. (1935). "Capital Concept Applied to Man," Quarterly Journal of Economics, 49, pp. 255-85. Ward, B. (1958). "The Firm in Illyria." American Economic Review, 48, pp. 566-89. Williamson, O. (1991). "Comparative Economic Organization: The Analysis of Discreet Structural Alternatives." Administrative Science Quarterly, 36, pp. 269-296. (1985). The Economic Institutions of Capitalism: Firms, Markets and Relational Contracting, New York, New York: Free Press. Williamson, O. and S. Masten (1995). Transaction Cost Economics, England: Edgar Elgar Publishing Limited. World Bank (1993). World Development Report, Washington, DC. Wu, C. (1977). "Education in Farm Production: The Case of Taiwan." American Journal of Agricultural Economics, 59(4), pp. 699-709. Yankelvich, P. (1988). Honduras, Mexico, D.F.: Editorial Patria, S.Af.. de C.V. Zepeda, L. (1990). "Predicting Bovine Somatotropin Use by California Dairy Producers." Western Journal of Agricultural Economics, 15 pp. 55-62.

PAGE 223

BIOGRAPHICAL SKETCH Michael Martin was born and raised in Des Moines, Iowa. He graduated from the University of Iowa in August, 1977, with a bachelor's degree in economics. He worked as a coop advisor with the Peace Corps in El Salvador and Honduras from 1979 to 1982. He enrolled in the Agricultural Economics Department at the University of Wyoming in 1983. In 1985 he conducted tleldwork and research in Somalia to fulfill the requirements for a Master of Science degree. He enrolled in the Food and Resource Economics Department at the University of Florida in 1985. In 1988, he returned to Honduras to gather data for his Ph.D. thesis from the cooperatives with which he had worked as a Peace Corps volunteer. His doctorate was awarded in December, 1996. 211

PAGE 224

I certify that I have read this study and that in ray opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. timothy G. Taylor, Chair Professor of Food and Resource Economics I certify that I have read diis study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. UmaLele, Co-Chair Graduate Research Professor Professor of Food and Resource Economics I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Chris O. Andrew Professor of Food and Resource Economics I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ^lies L. Scale, Jr. Professor of Food and Resource Economics

PAGE 225

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. rdi 'of History This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. f\ December, 1996 Dean, College of Agriculture Dean, Graduate School i it