I" :IY “(ka “Ya ’t‘l'mafl‘j Paw ' “ III» I‘ " I- II 3, l"|"§f."_._. and WI. 1:}. l.,\b(\“‘ In...” , A ‘ 'I ' um“. .i', I“? 1". ml. :3. 3:41.:"3 '. ' t m L}. I‘u I¥IMI (L5,. w “4’: it: h} (1'1”: I 9.‘Fv:},-3,C§LT~ .' .311... « l 1 ‘v I . III, ”II §§fiiix “>7." I , . I - f L‘ _ 51.1.2. ’ I l ‘1', ‘-.‘ ‘3' I, I': i:UIIIIJ fin"! W151. I :5 M ’3' . MC ' I: {:i \ "H‘y‘i Ilhf' .i|.:T- .I IN is I'llf‘zufdfl NJ, 21%;? 1? . .: | .l I}; [0 92‘5“); (. 4P #0 $1'C‘,‘2 dull“ ‘- {fi‘ug‘n 1:;§" ”I“; :‘E‘Vrgh; : I?” ’53:}; 1?".ng' ‘, : ,EI‘J‘MM «:1 k . 'l. w" P: 'd y. "‘04,: ‘4‘ l I: .E‘S j. I\\‘ \ '3K 5‘: YHAIV “H. ‘ V r? In?”- ‘ . ”'lr'h I- 's Ififlrniiy: |"I- (“In I [fif‘ '3"? .%~"I I.“ II u.'."._‘::! “I: l . l” ; ”III-III}? L U}: .~I ,3. -..,, m WP-gxmgig II I , 1" 1 _, :I I???» ~ III III; III ’i‘I II II I I .- ' :IIJ‘ I I 1 xi”; ~ r ,t n ', o ,, ..-3’ ,. - g :3fi3qgififi (5‘ “ii: ‘ 13‘ J " . D . "v .;. , .le'.” , Tfl'u ‘.\ ‘ " 3&3“: d‘m _.'..:l', 'k #M. N) ". I: 1"Ti‘ fag“ I If 'We :5 KW“ :‘:. III ' 1| I "l. 'I , It. "'5‘” ‘1‘! “W?! I' . " alrigh‘fl‘I‘WU ‘ -.' ' l ' II' 'I P: ' ’gfi’p ‘. @? '\f'1' ‘22:",ng fl}? 0,”. ‘. . ' 'fi‘g'lhu"ié’$1r;hi(fi-r fi’fl' Ir 0%} It" fiz... "- I I ' “ I III III I, p IIIIII'M M”: It. H 3M,” haw 'Ej'é’ 1“ |. gifza . '3: 1' a ’- - J .'-_‘ fiJ'W’_o - ‘ ‘ ”$21.; ‘I III . Lb; .II. N )‘II‘NHMIHQ'R' 13:" a: 'I I' I?“ ”L“ I“ ~I~.I..~:II.:I.II1‘,“II If I" I?!” bywan‘xk‘nw 1h", i'ST all“ II J ' "W (II:- MY ,1. In” _b . '— (“”f L " '.~'I‘i'I'l,'" ‘ll‘bll ‘ WI } i H ' I?! I I. ..' ”M I! “' 3'" » ‘. NJ 3% “ 4| n I ,3“?- 5” L.“ ‘ _ V'". HLO‘s). THESIS (I‘M? ’ 'i gnu-4; .I' .. to...“ " .3" '3’... - L I B R A R if" {I Fehigan Siam Univcwty 3:. a" } : l4"g§wo’ Fr ‘ This is to certify that the thesis entitled Economic Efficiency of Grain Production Systems for Traditional Agriculture In Southeastern Minas Gerais, Brazil presented by Carlos Antonio Moreira Leite has been accepted towards fulfillment of the requirements for Ph. D. ,1ng in Air. Economics ll ., , ( Major professor W7 l 0-7639 MSU LIBRARIES “—- RETURNING MATERIALS: ace in 00 rop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. ' Mrfirii $039 ECONOMIC EFFICIENCY OF GRAIN PRODUCTION SYSTEMS FOR TRADITIONAL AGRICULTURE IN SOUTHEASTERN MINAS GERAIS, BRAZIL By Carlos Antonio Moreira Leite A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Economics 1981 ABSTRACT ECONOMIC EFFICIENCY OF GRAIN PRODUCTION SYSTEMS FOR TRADITIONAL AGRICULTURE IN SOUTHEASTERN MINAS GERAIS, BRAZIL By Carlos Antonio Moreira Leite This study is an attempt to understand rural poverty in the Zona da Mata region of Minas Gerais State in Brazil. This is a depressed area of the state where traditional agriculture predomi- nates and socioeconomic problems are chronic. A comprehensive develop- mental program designed to meet the needs of poor farmers in the region has been implemented through a joint effort between local governmental institutions and the World Bank. Its objectives focus on revitaliz- ing the agricultural economy of the Zona da Mata and upgrading the welfare of its population. Many activities are found on different-sized farms as well as various opportunities for off-farm jobs, resulting in highly variable levels of farmer income. Grain production and consumption constitutes a relatively important source of food and income for farmers in the region. The specific objectives of this research were: (I) to develop a conceptual framework of income determination for Zona da Mata farm- ers. This provided a mechanism for identifying, among categories of farms, the differences in resource endowment, resource use and their return, management efficiency, farm output and its use, cash farm Carlos Antonio Moreira Leite income, off-farm income, and total family income; (2) to analyze the grain-production systems including the economic efficiency of input use and possible changes in resource allocation in order to improve farm incomes; and (3) to discuss the findings and implications of the research for future policy actions toward improving farm income of the study area. Data for the analysis came from a sample of 550 family farms divided into five categories of farmers: sharecroppers, and landowners with 0-10 hectares, lO-SO hectares, 50-l00 hectares, and lOO-ZOO hec- tares. Analysis of income determination revealed relatively low investment in capital. Family labor was the most intensive input used in the many activities performed in the region, but there was a tendency toward farm specialization in dairy and coffee production among farmers with more than 10 hectares. Among the many farm activi- ties of smaller farmers (sharecroppers and landowners with less than l0 hectares), grain production was the most common. These farmers had a high percentage of illiteracy, receiving less assistance from the agricultural extension service, and only a small percentage partici- pated in agricultural cooperatives. About 50 percent of the small farmers' family income was generated in off-farm jobs. The l0-50 hectares landowners' family income was about the average of the study area. Sharecroppers and O-lO hectares landowners had relatively lower incomes, whereas those farmers with more than 50 hectares had incomes considerably higher than the region average. Carlos Antonio Moreira Leite The economic analysis of the grain subsector suggested no differences in farmers' production functions between small and large farmers. Econometric analysis of the production systems of corn, beans, and rice suggested different production functions for different subregions of the study area. Analysis of economic efficiency of inputs use indicated that in the corn-beans production system there was no statistical evidence of misallocation of labor and modern inputs. On the other hand, land, labor, and modern inputs are misallocated in rice production and in sole-cropped production systems of corn and beans. Additionally, small farmers could achieve high farm income by producing with the optimum combination of inputs. Creation of off-farm jobs, enforcement of long-life contracts between landowners and landless, enlargement of social and agricultural services, cooperatives, and credit expansion for small farmers were among the recommendations derived from the empirical results, intended to raise farm income of the study area. ACKNOWLEDGMENTS I would like to express my appreciation to my major professor, Dr. Darrell Fienup, for his interest and guidance throughout my graduate program and for his assistance in the development of this study. Also, I wish to express my appreciation to Drs. Lester Manderscheid, Peter Schmidt, Vernon Sorenson, and John Ferris for their assistance as members of my guidance and thesis committee. I am grateful to the Brazilian Ministry of Education and to the Universidade Federal de Vicosa (UFV), for providing the opportu- nity to pursue advanced graduate studies. In particular, I am indebted to Professor Antonio Faqundes de Souza, Dr. Teotonio Dias Teixeira, and Dr. Antonio Lima Bandeira, chairman of UFV's Departa- mento de Economia Rural. Special appreciation is extended to my friends, Ismael de Mancilha, William Whelan, David Armstrong, and Dr. Manoel Dias, for their constant support. I would like to thank Mrs. Lucy Wells, who typed part of the first draft of this research, and Mrs. Sue Cooley, for her editorial assistance and for typing the dissertation. Finally, I am sincerely grateful to my family for their love and care. ii TABLE OF CONTENTS LIST OF TABLES ......................... LIST OF FIGURES ......................... INTRODUCTION ...................... Problem Statement ................... Objectives of the Study ................ Organization of the Study ............... GENERAL CHARACTERISTICS OF THE STUDY AREA AND THE SAMPLING PROCEDURE .................. Introduction ..................... The Structure of the Farm Sector of the State of Minas Gerais .................... General Characteristics of the Zona da Mata ...... The Farm Sector of the Zona da Mata .......... The Integrated Rural Development Program for the Zona da Mata Region of Minas Gerais State ...... Project Cost and Components ............. The Sample and the Data ................ The Sample ..................... The Data ...................... Summary ........................ CHARACTERIZATION OF FARM PRODUCTION SYSTEMS ....... Introduction ..................... Income Determination Conceptual Framework ....... Analysis of the Farm Family Income Determination Conceptual Framework ................ Resource Endowment ................. Resource Use and Productivity ............ Management Efficiency and Family Composition . . . Production, Consumption, and Marketable Surplus . . . Market Prices .................... Farm Income, Off-Farm Income, and Family Income . . . iii 050100 -‘ \J IV. ECONOMIC EFFICIENCY OF THE GRAIN SUBSECTOR ....... The Grain Cropping System in the Zona da Mata ..... Specification of Grain-Production Functions and Some Theoretical Considerations ........... Production-Function Specification .......... Production-Function Estimation Procedure ...... Selecting Subsamples for the Production-Function Analysis ...................... Estimations of the Grain-Production Functions ..... Elasticities of Production ............. Marginal Productivities ............... The Economic Efficiency of Input Use ........ Minimum Cost Combination of Inputs ......... Gains From Operating at the Least-Cost Combination Summary and Conclusions ............... : V. SUMMARY AND CONCLUSIONS ................. Introduction ..................... Summary of Problem, Objectives, and Methodology . . . . Summary of Findings .................. Implications and Policy Issues ............ Limitations and Suggestions for Further Research APPENDICES ........................... A. SHARECROPPERS DEFINITION ................ B. THE GINI INDEX OF INCOME CONCENTRATION OF THE ZONA DA MATA ........................ C. TESTS FOR HETEROSCEDASTICITY .............. D. FARMING SYSTEMS RESEARCH ................ BIBLIOGRAPHY .......................... iv Table 2.l. LIST OF TABLES Distribution of Farm Sizes and Income in Minas Gerais, 1974-l975 ................... Urban, Rural, and Total Population of Minas Gerais State and Zona da Mata, l950-l975 ........... Number of Farms by Size in the Zona da Mata, MG, l975 . . Sample Composition: PRODEMATA, Zona da Mata, Minas Gerais, 1976-1977 ................... Average Farm Size and Land Use by Class of Producers, Zona da Mata, MG, 1976-77 ............... Types of Labor Available Per Productive Unit, Man-Days, Zona da Mata, MG, 1976-77 ............... Total Adult-Man, Adult-Woman, and Child Labor in Man- Days Available by Class of Producers, Zona da Mata, MG, 1976-77 ...................... Capital Stock in Housing, Storage, and Animal Facilities, Zona da Mata, MG, Agricultural Sector, 1976-77 ........................ Capital Stock in Machinery and Equipment, Zona da Mata, MG, Agricultural Sector, l976-77 ........... Average Number and Value of Work Animals and Livestock in Production Units, Zona da Mata, MG, 1976-77 Capital and Labor Availability Per Hectare, Zona da Mata, MG, 1976-77 ................... Use of the Land in Zona da Mata, MG, for Selected Activities, by Group of Producers, l976-77 ...... Comparative Output Per Hectare of Selected Enterprises Among Different Groups of Producers, Zona da Mata, MG, 1976-77 ...................... Page 10 ll 16 24 33 35 36 38 4O 42 44 47 50 .10. .11. .12. .13. .14. .15. .16. .17. .18. .19. .20. .21. .22. .23. Man-Days of Labor Used Per Hectare for Selected Crops Among Different Producer Groups, Zona da Mata, MG, 1976-77 ........................ Output Per Man-Day for Selected Crops by Different Producer Groups, Zona da Mata, MG, l976-77 ...... Comparative Analysis of Different Forms of Capital/ Land Ratios Among the Different Groups of Producers, Zona da Mata, MG, l976-77 ......... Comparative Analysis of Different Forms of Capital/ Availability of Labor Ratios Among the Different Groups of Producers, Zona da Mata, MG, l976-77 Comparative Analysis of Different Forms of Capital] Labor Used in Agricultural Production Ratios Among the Different Groups of Producers, Zona da Mata, MG, 1976-77 ...................... Comparative Analysis of Different Forms of Available and Used Labor/Land Ratios Among the Different Groups of Producers in Zona da Mata, MG, l976-77 Use of Credit in the Agricultural Business by Selected Groups of Farmers, Zona da Mata, MG, l976-77 ..... Farmers' Participation in Cooperative Associations, Zona da Mata, MG, 1976-77 ............... Comparative Analysis of Selected Variables Related to Management Among the Different Classes of Farmers, Zona da Mata, MG, 1976-77 ............... Age Distribution of Farmers Within Each Category of Producers, Zona da Mata, MG, 1976-77 ......... Selected Educational Characteristics of the Farmers and Wives of Zona da Mata, MG, 1976-77 ........ Use of Formal Education by the Head of the Family, Zona da Mata, MG, 1976-77 ................. Average Family Composition of the Rural Zona da Mata, MG, 1976-77 ...................... Production, Consumption, Sales, and Marketable Surplus of Selected Products for the Entire Sample, Zona da Mata, MG, 1976-77 ................... vi Page 52 53 56 57 59 61 62 65 69 70 71 73 76 79 4.1. 4.2. 4.3. 4.4. Production, Consumption, Sales, and Marketable Surplus of Selected Products for Sharecroppers, Zona da Mata, MG, 1976-77 ................... Production, Consumption, Sales, and Marketable Surplus of Selected Products for 0-10 ha Landowners, Zona da Mata, MG, l976-77 ................. Production, Consumption, Sales, and Marketable Surplus of Selected Products for 10-50 ha Landowners, Zona da Mata, MG, 1976-77 ................. Production, Consumption, Sales, and Marketable Surplus of Selected Products for SO-lOO ha Landowners, Zona da Mata, MG, 1976-77 ................. Production, Consumption, Sales, and Marketable Surplus of Selected Products for lOO- 200 ha Landowners, Zona da Mata, MG, 1976- 77 ............... Analysis of Variance of Different Prices Received for Selected Products by Farmers, Zona da Mata, MG, 1976-77 ........................ Net Farm Income, Off-Farm Income, and Family Income of Farmers, Zona da Mata, MG, 1976-77 ......... Average Off-Farm Incomes of Five Groups of Producers, Zona da Mata, MG, 1976-77 ............... Percentage Contribution of the Revenue of Selected Products to Farmers' Gross Income, Zona da Mata, MG, 1976-77 ...................... Percentage of Farmers Growing Grains and Percentage of the Production Used on Farm by Farm Size Categories, Zona da Mata, MG, 1976-77 ............... Total Number of Grain Producers of the Zona da Mata and Number of Cases Selected for Grain-Production- Function Analysis ................... Production Functions of Grains for Small Farmers of the Zona da Mata, MG, 1976-77 . . . . . . . . . . . . . Production Functions of Grains for Large Farmers of the Zona da Mata, MG, 1976-77 ............. Production Functions of Grains for the Zona da Mata, MG, 1976-77 ...................... vii Page 81 83 85 86 88' 92 95 97 99 102 117 120 121 123 Input and Output Sample Means, Factor Marginal Productivities, and Marginal Value Productivities, Zona da Mata, MG, l976-77 ............... l27 Marginal Returns/Factor Cost Ratios and Tests for Differences of the Ratios From Unity of Selected Grains at the Geometric Means of Resources, Zona da Mata, MG, l976-77 .................. l3O Optimum Input Combination and Deviations of Actual Input Combination From the Optimum at the Geometric Mean of Production of Small Farm, Average Farm, and Large Farm Of Bean, Corn, Rice, and Corn-Bean Intercropped, Zona da Mata, MG, l976-77 ............... l32 Value of Output and Input Per Hectare of Selected Grains, Zona da Mata, MG, 1976-77 ............... 137 Participation of Sharecroppers in Contracts With Landowners of Selected Enterprises, Zona da Mata, MG, 1976-77 ...................... 159 Distribution of Farm Gross Income Among Five Classes of Producers, Zona da Mata, MG, l976-77 ........ 162 Simple Correlations Among the Variables of Estimated Function for Beans in Table 4.4 ............ 169 Simple Correlations Among the Variables of Estimated Function for Corn in Table 4.4 ............. 170 Simple Correlations Among the Variables of Estimated Function for Rice in Table 4.4 ............. 171 Simple Correlations Among the Variables of Estimated Function for Corn-Beans Combination in Table 4.4 . . . . 172 viii LIST OF FIGURES Figure Page 2.1. Location of the Zona da Mata in Brazil .......... 8 2.2. Regions of Zona da Mata ................. 23 3.1. Farm Family Income Determination Conceptual Framework . . 29 B.1. Illustration of a Lorenz Curve .............. 163 0.1. Schematic Representation of Some Determinants of the Farming System ..................... 176 ix CHAPTER I INTRODUCTION There is increasing concern about the poverty problem in Brazil. In attempting to determine the causes of the problem, sev- eral reasons have been found to explain why earnings of middle-income and upper-income groups have risen more rapidly than those of the poor. The capital-intensive type of development strategy adopted in Brazil after the Second World War has produced a greater concentration of income in contrast to other countries, such as Taiwan and Korea, which adopted policies that distributed more widely the benefits of moderni- zation.1 Fields used the absolute poverty measures in place of the usual relative inequality indices to deal with the problem of unequal 2 He concluded distribution of income in Brazil during the 19605. that the poor in Brazil clearly did share in a decade of economic development. Some poor were lifted out of poverty. However, for those left behind, even though their income grew in absolute terms, in relative terms it did not grow as rapidly. Fields commented that 1Hollins B. Chenery, "Poverty and Progress--Choices for the Developing World," Finance and Development 17 (June 1980): 12-16. 2Gary S. Fields, "Who Benefits From Economic Development? A Reexamination of Brazilian Growth in the 60's," The American Economic Review 64 (September 1977): 570-82. the very rich became richer than before in both absolute and rela- tive terms. The bulk of the poverty problem in Brazil seems to be con- centrated in the rural sector. Government efforts have been focused on understanding the critical conditioning factors of that problem and on implementing policies that would minimize them. An attempt in that direction was supported by the Brazilian government for a rather large research project (Development Alternatives for Low-Income Groups in Brazilian Agriculture) to be carried out jointly by six domestic institutions and one from abroad. Overall objectives of that project were (a) to gain an increased understanding of the rural poor and the environment in which they live and (b) to derive strategies whereby the income and welfare of this group could be improved. The regions included in that project were Canidé, state of Ceara; Vale do Ribeira, state of Sao Paulo; and Campo dos Vertentes and Zona da Mata, state 1 The majority of research that came out of that of Minas Gerais. project was macroeconomic in nature, and because of the very unequal distribution of land holdings, limited employment alternatives, and chronic concentrated income distribution, emphasis was placed on the Canidé region. Despite the insights into the macroeconomic aspects of rela- tive poverty in the agricultural sector of the Zona da Mata, very 1For a summary of the project and its recommendations, see Guilherme Leite da Silva Dias, "Pobreza ruraleno Brasil: Caracterizacfio do problema e recomendacOes de politica," Colecfio Analise e Pesquisa, vol. 16 (Brasilia: Comissao de Financiamento da Producfio, Ministério da Agricultura, Agosto 1979). little research has been conducted with the objective of understand- ing its microeconomic dimensions. The present study is concerned with the process of determining household income and the economic efficiency of the major subsector of the region--the grain sub- SECtOY‘. Problem Statement As is characteristic of Northeast Brazil, the Zona da Mata of the state of Minas Gerais is considered a depressed area. Among the many factors that contribute to that area's backward position relative to other regions of the state have been cited the lack of official developmental policies to promote agricultural research compatible with the resource endowment of the region;I lack of invest— ment in human capital, and lack of governmental support to the farmers to start farming again after coffee eradication, which took place between 1962 and 1966. Besides milk and coffee, which are produced mainly by large farmers, it is believed that, as an income generator, grains (corn, beans, and rice) are the second most important enterprise for farmers in the Zona da Mata. Even with the tendency of farmers in this region to produce cash crops, a large proportion of the rural 1Most of the research conducted in the neighborhood research institutes was intended for mechanized or capital-intensive farms, which is not the case of most of the farmers in the Zona da Mata. population of the Zona da Mata live in poverty1 and practice tradi- tional agriculture.2 A comprehensive developmental program intended to address the needs of poor farmers of the area has been implemented: Programa Integrado de Desenvolvimento da Zona da Mata, MG (PRODEMATA). This program is the result of a joint effort between domestic governmental institutions and the World Bank. An increase in the supply of credit is its main component. Indeed, one of the major hypotheses of the program is the positive correlation between farmers' income and use of modern production inputs, which is to be increased through increased use of credit. Because of the many activities developed on different-sized farms as well as different opportunities for off-farm jobs, it should be useful to study the process by which farm income is determined. As the grain subsector constitutes a relatively important source of food and income for Zona da Mata farmers, coupled with the fact that this subsector is a potential recipient of a large percentage of credit from the PRODEMATA, it is also important to study the economic efficiency of producing beans, rice, and corn. It is argued that the knowledge of these factors may constitute valuable inputs for 1Diagnéstico EconOmico da Zona da Mata de Minas Gerais, Universidade de Federal de Vicosa, 1968, Ch. 7. 2The term "traditional" or "subsistence" agriculture is used here as defined by Clifton R. Wharton, Jr., "Subsistence Agriculture: Concepts and Scope," in Subsistence Agriculture and Economic Develop- ment, ed. Clifton R. Wharton, Jr. (Chicago: Aldine Publishing Co., 1969) and refers to farmers who use mainly family labor in the produc— tion process, and, although some output may be sold when a surplus occurs, production is devoted primarily to on-farm consumption. development strategies in the region. For those responsible for agricultural policy implementation, that knowledge will be important in elaborating policies compatible with the real farm situation, whereas for research institutions such knowledge will be important in generating production techniques conforming with the characteris- tics of different groups of farmers. Objectives of the Study The overall objective of this study is to generate better knowledge about the process of income generation in the Zona da Mata farm sector. Particular emphasis is placed on the grain subsector because it is believed that this subsector plays a special role in the income generation of that sector. To fulfill this goal, the following specific objectives were set: 1. To develOp a conceptual framework of the income- determination process for the Zona da Mata farmers. This will provide a mechanism for identifying, among different categories of farms, the differences in resource endowment, resource uses and their return, management efficiency, farm output, on-farm consumption, marketable surplus and cash farm income, off-farm income, and total family income. This analysis will help identify typical combinations of enterprises for each class of farms. 2. To analyze the grain-production system in the study area and to verify differences in production among classes of farms. This analysis also should provide a basis for analyzing economic efficiency of resource use in grain production and possible resource realloca- tion in order to improve farm incomes. 3. To discuss the findings and implications of the research for future actions toward improving farm income in the study region. Organization of the Study This study is organized into five chapters. Chapter I con- tains a discussion of the problem setting, the importance of the study, and the objectives of the research. A brief characterization of the agricultural sector of the study area as well as the goals and spe- cific components of the PRODEMATA project are presented in Chapter II.. The sampling procedure used to generate the data for this study is also explained. Chapters III and IV deal with research results. Analysis of the income-determination conceptual framework is developed in Chapter III. The economic efficiency analysis of the grain subsec- tor is presented in Chapter IV. Finally, Chapter V contains a summary of the major findings of the study, their policy implications, and suggestions for further research. CHAPTER II GENERAL CHARACTERISTICS OF THE STUDY AREA AND THE SAMPLING PROCEDURE Introduction The main objective of this chapter is to present selected characteristics of the study area and of the PRODEMATA. The initial sections of this chapter deal with key structural characteristics of the farm sector of the state of Minas Gerais and some general geo- graphical and sociodemographic characteristics of the Zona da Mata. Some background information about the PRODEMATA project is presented later in the chapter, and its costs and component programs are exam- ined. Concluding the chapter, the sampling procedure and data used in this study are discussed. The Structure of the Farm Sector of the State of Minas Gerais This study was developed for the Zona da Mata, located in the eastern part of the state of Minas Gerais. (See Figure 2.1.) The agricultural sector of Minas Gerais is of great importance for the state itself and for the whole country. According to the agricul- tural census of 1975, the number of farms in the state was 454,465, and the total income of these farmers was about 15 billion cruzeiros (equivalent to US $4,000 per farm). Brazil State of Minas Gerais l———— Zona da Mata Figure 2.1.--Location of the Zona da Mata in Brazil. There are relatively few large farms in the state. About 28 percent of the farms consist of fewer than 10 hectares, and about 81 percent of the farms consist of fewer than 100 hectares. Farms larger than 100 hectares represent about 19 percent of the total, but they earn about 61 percent of the total income of the sector (Table 2.1). The Gini Index of Income Concentration, or the Gini Income Ratio, derived from the data presented in Table 2.1, is about .537.1 The census data show that farms with an area between 20 and 50 hectares are most common, accounting for about 30 percent of the total farms in the state. The farmers with the largest share of income owned farms of between 200 and 500 hectares and received 19.6 percent of the income generated in the agricultural sector in Minas Gerais in 1975. The percentage of the rural population of the state has decreased Since 1950 (Table 2.2). In 1975 it was estimated that about 40 percent of the state's population lived in rural areas. Since 1960, in fact, in absolute numbers, the rural population has decreased. The initial incentive for outmigration was the increased demand for labor during the construction of Brasilia, the new capital city, and lately the relatively more favorable wages in urban areas. 1To compute the Gini Index of Income Concentration or the Gini Income Ratio, the following formula was used: k Gini Ratio = 1 - 1&1 (fi+1 ' f1) (Y1 + Yi+11 (The variable definitions are presented in Appendix B.) A formal mathe- matical presentation of the Gini Ratio was made by Charles H. Riemen- schneider, "The Use of the Gini Ratio is Measuring Distributional Impacts" (M.S. research report, Michigan State University, 1976). See also James Morgan, "The Anatomy of Income Distribution," Review of Economics and Statistics 44 (1962): 270-83. 10 .AmumF .ocwmcma mu owmv mwmgmw mace: mo owngumaogm< omcmu .muPumwumumm m mwemgmomw on ogwmpvmmgm ouauwpmcH omumucam "muczom oo.oop mm.o mme.wm oo.oop No.o mm oooopm. mm.mm em.P ¢-.wo~ mm.mm mo.o new oooopuooom mo.mm mm.m www.mmw cm.mm mm.o mme.F ooomuooom wo.~m mm.“ mo_._np.~ Fm.mm mm.o mpe.m ooomtooo— mp.mw ne.- nem.mmu.~ ow.mm mm.p emo.m coopuoom oo.mm mm.mp Nom.mmo.m nw.mm «m.m ~o~.m~ oomuoom op.¢m mm.¢~ ooe.m~m.~ mm.om mm.m www.me oomnoop -.mm em.- Pop.moo.~ em.ow mm.mp mmo.~e copuom mp.o~ ~n.~p po~.Pmm.p m~.~o mm.- pme.¢o~ omuom _¢.mp mn.m NmN.me om.e¢ no.o_ meo.mu omnop on.“ om.m Nwo.mmm m~.m~ vw.N_ mmm.mm opum op.e oc.~ mpo.mo¢ mm.mp om.p_ cop.pm m-~ om.F mm.o m_m.mw mp.¢ mm.~ NP~._F NIP ~m.o mm.o www.mep mm.F mm.P mmo.~ fiw 33:3ch mfiumw 8.. ”JUNE”. ammuzmugma mrMumw web”... cmumpasauu< ucmugma mucemzosh umumpzszuu< ucwugma amass: “mm“ Amogvm~zgu coo.Pv msoocH mLmELmu .mNm—IQNmF ammeww manta F... 0:52;. Ucm mwNwm ELML. mo cowuznwLfimwQII._..N mpnmh. 11 Table 2.2.--Urban, rural, and total population of Minas Gerais State and Zona da Mata, 1950—1975 (in thousands). Minas Geraisa Zona da Mataa Year Rural Urban Rural Urban Total (%) (%) Total (%) (%) 1950 7,782.2 5,459.3 2,322.9 1,283.3 898.2 385.1 (70.2) (29.8) (70.0) (30.0) 1960 9,657.7 5,832.5 3,825.2 1,523.0 955.8 367.2 (60.4 (39.6) (62.8) (37.2) 1970 11,487.4 5,427.1 6,060.3 1,600.8 805.2 795.6 (47.2) (52.8) (50.3) (49.7) 1975b 12,550.6 5,199.9 7,350.7 1,623.7 789.1 834.6 (41.4) (58.6) (48.6) (51.4) Source: Fundacao Brasileira de Geografia e Estatistica (FIBGE), Anuario Estatistico do Brasil, 1955 and 1978 issues. aThe figures in parentheses are the percentages of rural and urban population. bEstimated by Fundacao Brasileira de Geografia e Estatistica (FIBGE), Anuario Estatistico do Brasil, 1978. General Characteristics of the Zona da Mata The Zona da Mata of Minas Gerais State covers an area of 36,012 km2 bordering on the states of Rio de Janeiro and Espirito Santo.1 The southern regions are rolling, becoming quite hilly toward the north, with areas of poor drainage in the valleys. It has been estimated that about 16 percent of the entire Zona da Mata 1Fundaca'o Instituto Brasileiro de Geografia e Estatistica, Anuario Estat1stico do Brasil (Rio de Janeiro, 1968). 12 is constituted of flat land, whereas 44 percent and 40 percent are rolling and hilly land, respective1y.1 The temperature of the region averages about 22°C. The mean annual rainfall is about 1,400 mm, with a dry period for six months from April to September. It was estimated that in 1975 the total population of the Zona da Mata was about 1.6 million (Table 2.2). The population den- sity was about 35 persons per kmz, about four times the average for Brazil. Estimates for 1975 indicated that about 49 percent of the Zona da Mata population was living in rural areas. That percentage has been decreasing since 1950, even though from 1950 to 1960, in absolute terms, that population increased. Rural per capita income for the area was estimated in 1974-75 to be about US $250 equivalent, which was about 25 percent of the 2 Considering the pov- per capita income of the country as a whole. erty level as one-third of national per capita income, the rural per capita income was below the national relative poverty level of US $340 equivalent. The region's social infrastructure is poor, and available health and education services are deficient.3 As observed in many 1Tacito Claudio Andrade Taveira, "Analise de Localizacao da Producao Agricola em Relacfio ao Mercado de Juiz de Fora--Minas Gerais" (M.S. thesis, Universidade Federal de Vicosa, 1976). 2The World Bank, Brazil--Staff Project Report of the Inte- grated Rural Development Project in the State of Minas Gerais, Report No. 1291 Br. (Washington: World Bankjil976). 31bid., p. 14. 13 underdeveloped regions in Brazil, the health status of the inhabi- tants of the Zona da Mata is characterized by high mortality and morbidity rates caused by communicable diseases; high infant mortality rate, caused mainly by infectious diseases; and serious incidence of malnutrition as a basic or associated cause of child mortality. It has also been observed that schistosomiasis is a widespread problem; in some localities, a high percentage of the population is infected. State investment in education in the rural Zona da Mata is relatively low. It has been estimated that two—thirds of the education is financed by the municipios (municipalities) and one-third by the state government. Besides the serious lack of facilities, the inade- quacy of the teaching services has also been observed. It has been estimated that a high percentage of the school staff members do not have the basic qualifications required by law, and the curricula are generally not compatible with the students' future needs. The limi- tations of educational opportunity seem to be reflected in the low educational attainment of the labor force. The population census of 1970 revealed that 60 percent of the agricultural workers of that region had not had any formal education.1 The Farm Sector of the Zona da Mata Colonization of the Zona da Mata was similar to that in many other areas in Brazil. The gold race and the desire to make a fortune from precious stones brought explorers from the Brazilian coast to the 1Fund CEO Instituto Brasileiro de Geografia e Estatistica, Anuario Estat1stico do Brasil, 1978. 14 southeastern part of Minas Gerais during the seventeenth century. At that time, agriculture was either nonexistent or, in some areas, was conducted on a small scale largely at subsistence level. Mining profits were sufficient to pay for importing food and other necessi- ties from other regions of the country. About 1830, coffee plantations were introduced to the Zona da Mata. Cattle had also been introduced to the region. Mining pro- ductivity was declining. Because of the relatively good price of coffee, coffee growing became an incentive to farming in the Zona da Mata, and coffee became one of the most important products of the state's economy. Small-scale industries grew up in the Zona da Mata to supply the market demand generated by the coffee economy.1 Until the first quarter of the twentieth century, the Zona da Mata occupied an important position among the other regions of the state. Since then, however, it has become relatively less important for many reasons: 1. Lack of development of new agricultural technology. Because land and labor were abundant, little attention was given to these factors in public policy. 2. The land is hilly and the soil is poor. These facts imply a need for proper land management, which increases production costs and decreases the competitive market position of the Zona da Mata. 1Universidade Federal de Vicosa, DiagnOstico EconOmico da Zona da Mata (Vicosa: Imprensa Universitaria, 1971). 15 3. Limited basic education, health, and extension services in rural areas. These factors decrease labor productivity and con- sequently increase unit production costs. 4. Stagnation of the industrial sector followed by lack of dynamic business practices and obsolescence of the industrial sector.1 After the program to eradicate coffee plantations from 1962 to 1966, which had as one objective to raise coffee prices, the econ- omic degeneration of the Zona da Mata was accelerated because of the elimination of its most important source of income. It is worth noting that since the nineteenth century, besides coffee and sugar cane, cattle ranching was a subsidiary activity. The production of milk and meat in some regions became as important as coffee because of poor returns from producing coffee at low prices. With the elimi- nation of coffee as a crop, the plan was to emphasize cattle raising. However, the introduction of more adapted and specialized animals changed the production cost structure. Animals with better genetic characteristics, capital investment and equipment, and specialized labor were the most important components. High production costs limited the substitution of cattle for coffee on small farms as well as many medium-sized farms. Only the large farmers were able to survive in this long economic crisis. The number of farms in the Zona da Mata was about 67,474 in 1975 (Table 2.3). About 76 percent of these farms were smaller than 1Universidade Federal de Vicosa, DER, Programa Integrado de Desenvolvimento da Zona da Mata--MG, Primeiro Relat6rio Anual de Avaliacao (Vicosa: Imprensa Universitaria, Marco 1979). 16 50 hectares, and 33 percent were smaller than 10 hectares. The World Bank estimates that there are also about 24,000 Sharecroppers1 engaged primarily in the production of subsistence food crops. The problem of land ownership has been indicated as being associated with the major economic problems of the Zona da Mata. Silva computed the distribution of land concentration (Gini Ratio) as being equal to .68, indicating concentration of land ownership in the region.2 Table 2.3.--Number of farms by size in the Zona da Mata, MG, 1975. Class of Total Number Percentage Cumulative Farm (ha) of Farmers of Farms Percentage of Farms 0- 10 22,171 32.9 32.9 10- 50 28,962 42.9 75.8 50-100 8,627 12.8 88.6 TOO-200 4,996 7.4 96.0 200 and higher 2,718 4.0 100.0 Total 67,474 100.0 -- Source: Fundacao Instituto Brasileiro de Geografia e Estatistica, Censo Agropecuario de Minas Gerais (Rio de Janeiro, 1975). The relative abundance of labor and the relatively low yields of most crops are distinguishing characteristics of tradi- tional agriculture in the Zona da Mata. Action toward bringing about 1World Bank, op. cit., p. 1. 2Carlos Arthur 8. da Silva, "Factors Affecting Enterprise Choice: An Analysis of Traditional Food Production in Southeastern Minas Gerais, Brazil" (Ph.D. dissertation, Michigan State University, 1981). 17 changes in the study area have been implemented. Several state organizations have combined their efforts with those of the World Bank to implement a comprehensive development program (PRODEMATA), which is described in some detail in the next section. The Integrated Rural Development Program for the Zona da Mata Region of Minas Gerais State The Integrated Rural Development Program for the Zona da Mata Region of Minas Gerais (PRODEMATA) has financial participation from the World Bank. The main domestic agencies involved in this program are the State Secretariat of Planning (SEPLAN), which is responsible for its implementation; the State Rural Development Agency (RURALMINAS), which is the overall coordinator; and the Departmento de Economia Rural (DER) of the Universidade Federal de Vicosa, which is respon- sible for evaluation of the project through time. The objectives of the PRODEMATA focus on revitalizing the agricultural economy of the Zona da Mata and upgrading the welfare of its population. These objectives are to be accomplished by: 1. Raising the income levels of the poor families by provid- ing credit and technical services, 2. Increasing agricultural production by introducing techni- cal innovations and expanding the variety of crop and livestock enterprises, and 3. Generally improving the quality of life by expanding and upgrading social services such as education, health, and sanitation. However, increased agricultural production would be the backbone of the project. 18 Project Cost and Components1 The cost of the project is estimated at US $139 million. About 60 percent of the project's total cost is allocated to the agricultural credit component. The beneficiaries of these services are primarily Sharecroppers and small farmers with fewer than 100 hectares of rainfed land. Productive credit covers normal crop, livestock, and forestry production costs, including consumption credit for the smaller farmers and Sharecroppers and hired labor and other inputs. Loansfbr investments are mainly for the establishment of sugar cane and fruit trees, pasture establishment and improvement, reforestation, the purchase of breeding stock, fencing, farm equip- ment, livestock handling facilities, and rural electrification. A special line of credit is available to large-sized farmers (100 to 200 hectares), especially for land reclamation and reforestation purposes. The rural electrification component, with a total cost of about US $6.2 million, is intended to provide electricity for home, farm, and agro-industrial use in parts of the project area where elec- tricity is not available. Under the land-reclamation component, it is estimated that .about 8,000 hectares of individually owned, poorly drained valley-bottom land will be reclaimed by means of appropriate irrigation, drainage, land leveling, and flood control. RURALMINAS is the agency in charge of this project component, and it is estimated that the final 1For more details concerning the project cost and components, see World Bank, op. cit., pp. 17-32. 19 payment by farmers for this land improvement will be about 70 percent of the market price for similar services. The production support services component comprises mainly agricultural research, technical assistance, and extension and coop- erative services. A relatively small program of applied agricultural research and demonstration is carried out under the project. About US $1.2 million is the total investment in such a service. This share of the total project cost is intended to finance the cost of field trials, staff salaries, operating costs, and the purchase of equipment, including vehicles. Technical assistance and extension services, which have been directed primarily at larger farmers of the area, are to be enlarged to serve smaller farmers and Sharecroppers. About 150 new extension agents have been recruited and are being trained with assistance pro- vided by the Universidade Federal de Vicosa and EPAMIG. It is esti- mated that a total of US $10.3 million will be used to finance the cost of new equipment, including vehicles and other operational costs for enlargment of technical assistance services. The plan concerning cooperatives in the Zona da Mata is at least to double the present 31 cooperatives with 1,600 members. The goal is to establish a new regional office of State Superintendency of Cooperatives (SUDECOPE) to (a) promote cooperation and encourage the formation of cooperatives among small farmers, (b) provide technical assistance to existing cooperatives; (c) organize training courses for cooperative managers and administrators; and (d) offer courses to cooperative members, especially on marketing and production. 20 Finally, the social services component includes investments in basic aspects of health and education. The main objectives of the health component would be to provide health posts distributed throughout the project area, to offer low-cost health services, and to emphasize preventive medicine and promotional activities. Further investments would be made in sanitation programs, with the objective of diminishing morbidity and mortality from diseases caused by poor sanitary conditions. In addition, a vaccination program would be implemented to diminish the incidences of communicable diseases in the area. Also, health authorities would develop a program to investigate schistosomiasis foci, especially in the varzeas (valley lowlands), of the project area. A nutrition program would also be carried out with the objective of preventing and diminishing caloric and protein malnutrition. Priorities of this program were established in the following order: pregnant women, lactating mothers, children under three years of age, and three- and four-year-old children who are undernourished. With regard to investment in education, the strategies of the project are to expand the role of the rural school by making it a multipurpose center for the provision of adult education services; to expand and improve rural primary education; to disseminate, by means of school and out-of—school education activities, basic knowledge of agricultural technology, farm management, rural organization, and family education; and to develop better education-management capabili- ties by means of technical-assistance programs. The Sample 21 The Sample and the Data The Departamento de Economia Rural, Universidade Federal de Vicosa, selected the sample during July 1977. The sample comprised 851 family farms selected at random from the files of the National Institute of Colonization and Agrarian Reform (INCRA). Landowners and Sharecroppers constituted the sample. The sampling procedure was as follows: 1. Regionalization of the Zona da Mata according to the administrative division of the Extension Service Agency of Minas Gerais (EMATER), including the regions of Muriaé, Vicosa, and Juiz de Fora. Identification of the municipalities in each of the three regions. Identification of the villages and their populations in each municipality to be covered by EMATER-PRODEMATA. Selection of four municipalities in each region, two of them selected at random and the other two having the fol- lowing characteristics: a. high population and large number of farms receiving assistance, b. typical or traditional producers of products of interest to PRODEMATA (such as tobacco or sugar cane), not included in other municipalities. Distribution of the sample members according to the number of properties in a selected municipality. 22 6. Random selection of 700 landowners within the selected municipalities, using the INCRA files for the year 1970. These farm owners were to be interviewed during the course of PRODEMATA. 7. Identification of sharecroppers,1 as indicated by the selected landowners, because there are no records of this class of producers. The number of Sharecroppers selected by region was 50. For each municipality selected at 2 random, at least 12 Sharecroppers were selected. Municipalities selected by region were the following: Region of: Juiz de Fora Hflliéé. Viggsa 1. Alto Rio Doce 1. Carangola l. Ervalia 2. Juiz de Fora 2. Leopoldina 2. Ponte Nova 3. Santos Dumont 3. Manhuacu 3. Raul Soares 4. $50 Joao Nepomuceno 4. Muriaé 4. Uba The sample distribution by regions and categories of farmers is presented in Table 2.4. See Figure 2.2 for geographical locations of the regions considered. The three regions of the Zona da Mata (Juiz de Fora, Muriaé, and Vicosa) are relatively homogeneous in terms of agronomic char- acteristics; however, they differ in other aspects. Juiz de Fora 1Sharecroppers and landless producers of tenants who, under contract, use landowners' land in exhcnage for payment in kind, in cash, and/or in production factor costs. (See Appendix A for more details.) 2Universidade Federal de Vicosa, DER, op. cit. 23 .®#ME GU QCON $0 waves: mcowmmm--.m.m beamed O \ ‘ Q h a s o a O O a \ C ‘ Q Q ’ n \\ o o \\ or ; § \ 0"! s s o- . § § _ b . ~ \ \ ’ \ s .\ I'l-O‘ . r \ 0 \\ O Q s \ \ \ ‘0 \ a 5 fl ~ \ 6 ago; an Nwzw ' Sou; 24 region, for instance, is considered the major milk-producer region of the Zona da Mata, supplying the local market and the market of Rio de Janeiro city. The region of Muriaé is also a relatively important one in dairy products in the Zona da Mata. However, coffee seems to be its more important enterprise. The region of Vicosa is character- ized by a diversified agriculture. Traditional crops such as grains and sugar cane are among the main agricultural activities. The region of Juiz de Fora is characterized as the industrial center of the Zona da Mata. In the other two regions there are no major industries, and access of rural areas to demand centers is their peculiar character- istic. Table 2.4.--Sample composition: PRODEMATA, Zona da Mata, Minas Gerais, l976-1977. Categories Reg1ons , Zona da Mata Juiz de Fora Muriae Vicosa Sharecroppers 52 52 49 153 Landowners: 0- 10 ha 64 94 74 232 10- 50 ha 62 165 89 316 50-100 ha 24 50 24 98 100-200 ha 18 25 9 52 Total 220 386 245 851 Source: Universidade Federal de Vicosa, DER, Programa Integrado de Desenvolvimento da Zona da Mata--MG, Primeiro Relat6rio Anual de Avaliacao—(Vicosa: Imprensa Universitaria, Marco 1979). 25 Following the orientation of the Universidade Federal de Vicosa, for the purposes of this research the sample size was reduced to 550 observations, including only those respondents whose answers were consistent throughout the questionnaire. The new sample was distributed among the categories of farmers as follows: Sharecroppers 129 observations Landowners: O- 10 ha 123 observations 10- 50 ha 220 observations 50-100 ha 59 observations 100-200 ha' 19 observations The Data Basically, the questionnaire used to obtain the data sought information related to (a) agricultural production, (b) education, (c) health, (d) nutrition, and (e) sanitation. A general overview of the information obtained from the survey is presented below: 1. Agricultural Production (agricultural year 1976-77) A. Stock of production inputs 1. Land with its use 2. Capital: buildings, machinery, and equipment 3. Labor: hired labor and family labor 8. Agricultural production 1. Area used for production, the total production, and participation of Sharecroppers in the productive process 2. Percentage of production paid in kind by the share- croppers to the landowners for each product II. III. IV. VI. 26 3. Home consumption, quantity sold, and price received by the producer for each output C. Technology and input used in each enterprise 1. Quantity of labor used 2. Seeds and fertilizer used 3. Pesticides, machinery, and animals used 4. Amount spent on animal care and feed 5. Sharecroppers' share of all expenses D. Use of credit in each enterprise Wages received outside the farm A. Farm labor 8. Nonfarm labor Education A. Family composition 8. Formal and informal education C. Use of formal education received 0. Cooperatives and farmer involvement E. Social groups health A. Medical attention to the family 8. Medical service availability and use Nutrition: Description of the family's daily consumption Sanitation A. General household care with sanitation problems B. Conditions of the water used at home C. Stagnant water and waste disposal 27 Summary This chapter presented general characteristics of the Zona da Mata regibn of the State of Minas Gerais and an overview of the inte- grated rural development program (PRODEMATA) that has been implemented in the region. The objectives of this project are to improve living standards and incomes of small farmers and Sharecroppers in the study area through (a) increasing farm production by expanding the area under cultivation and raising yields and (b) expanding and improving social services to farmers and the general rural population. The major component of the project is provision of agricultural credit, which accounts for about 60 percent of the project's cost. Other major components are provision of supportive agricultural services; provision of social services, including investments in sanitation, health, and education; and rural electrification and land reclama- tion. In evaluating the project, the Universidade Federal de Vicosa surveyed about 851 family farms; this sample was later reduced to 550 farms. This reduced sample was used throughout the present study. CHAPTER III CHARACTERIZATION OF FARM PRODUCTION SYSTEMS Introduction Overall characteristics of the Zona da Mata farmers are pre- sented in this chapter. The sample has been divided into five cate- gories of farmers including: sharecroppers; and landowners with 0-10 hectares, 10-50 hectares, 50-100 hectares, and 100-200 hectares. A descriptive analysis of these categories of farmers will be made in terms of the conceptual framework presented in Figure 3.1. The objec- tive of that framework is to characterize farmers in terms of their farm size and use of land, labor, and capital in the production pro- cess. Additionally, an attempt is made to identify production systems within different farm categories which include off-farm activities of the family members in order to better understand how total farm family income is determined in the study area. A description of the conceptual framework is presented below. Income Determination Conceptual Framework Norman defined a farming system as a complex interaction of several interdependent components.1 Among these components is the 1David W. Norman, "The Farming System Approach: Relevancy for the Small Farmer," Michigan State University Rural Development Paper No. 5 (East Lansing: Michigan State University, 1980). See Appendix D for further details. 28 29 Resource Endowment Management Efficiency Resource Use and ' Productivity 1 [ Production 1 l I} Government Price Policy Marketed . Family Surplus [ Consumption l Composition gfigtgt Cash Farm Nonfarm Income Income Demand Family Income Figure 3.1.--Farm family income determination conceptual framework. 30 rural household acting as both a production and a consumption unit. If income growth and development are to be achieved, the multiple uses of farm and family resources must be understood, and the production-consumption-sales relationship must be related to the agricultural market and off-farm employment opportunities. Farm income and farm profitability can be dealt with in a similar manner for the same time period. As Brown suggested, farm 1 In the income can be viewed as a whole or as an isolated enterprise. first approach, returns to farm labor, capital, and management invested in a given time period, say one year, are studied. In the second approach, the profitability of each phase can be measured at various stages, such as gross output, net output, gross margin, and profits. Basic to the framework used in this research is that the cash farm income of a small-scale farm in the Zona da Mata will depend on the farm's marketed surplus and the compensation to family labor off the farm. Production is directly affected by (a) resource endowment, (b) resource use, and (c) management efficiency of the farmer, in addition to other exogenous variables, such as weather, that cannot be controlled by the farmer. Resource endowment includes availability of land, labor, and capital. It is hypothesized in this conceptual framework that land supply is limited, capital investment is small, and labor is usually limited to the family supply. Operating capital may be closely related 1Maxwell Brown, "Farm Budgets--From Farm Income Analysis to Agricultural Project Analysis," World Bank Occasional Paper No. 29 (Baltimore: Johns Hopkins University Press, 1979). 31 to farm income if improved seeds, fertilizer, and hired labor are used efficiently. Apart from credit availability, since operating capital is a function of savings, it can be concluded that the net farm income of a previous period is a critical factor determining current income level. Concerning the resource use variable, which is closely related to farm income, some empirical researchers have shown inefficient resource allocation by small-scale farmers in Brazil. In two regions of Minas Gerais State--Zona da Mata and Campo das Vertentes--Garcia found that small-scale farmers were not efficient in labor allocation.1 Assuming that farmers aim to maximize profit, it was found that the conditions for maximization were not satisfied. The farmers employed excessive labor. Consequently, the required equality of marginal product of labor and labor wage was not attained. Graber, studying the explanatory factors of farm production of small farms in Vale do Ribeira, S50 Paulo State, and Canidé, Cearé State,2 and Teixeira, analyzing the resource efficiency of small farmers in Canidé,3 concluded that there was a poor allocation of acquired pro- duction factors among these producers. 1JOEO Carlos Garcia, "Analise de Alocacao de Recursos por Proprietarios e Parceiros em Areas de Agricultura de Subsisténcia" (M.SS thesis, Universidade Federal de Vicosa, Imprensa Universitaria, 1975 . 2Kenneth L. Graber, "Factors Explaining Farm Production and Family Earnings of Small Farmers in Brazil" (Ph.D. dissertation, Purdue University, 1976). 3Teot6nio Dias Teixeira, ”Resource Efficiency and the Market for Family Labor: Small Farms in the Sertéo of Northeast Brazil" (Ph.D. dissertation, Purdue University, 1976). 32 The management ability of the farmer is critical to the level of income. Many factors determine a farmer's ability to make a cor- rect decision. Management decisions are generally grouped as fol- lows: (a) what to produce, (b) how much to produce, (c) the kinds and amounts of resources to use, (d) the technology to use, (e) when and where to sell and buy, and (f) how to finance the operation.1 The final component to be discussed in this conceptual frame- work is farm family income. The cash income generated from the farm refers to farm sales. Another component of family income in addition to sales is the income generated by working off the farm, which might be an important income source for the smaller producers of the study region. Analysis of the Farm Family Income Determination Conceptual Framework Resource Endowment Land. For the whole study area, the average farm size was about 27 hectares. The average land holding by Sharecroppers was about 6.28 hectares, approximately the same as the average size of the 0-10 hectares farm category, which was 6.44 hectares. The coef- ficient of variation of farm size for the whole study region was 126.29 percent; it was 196.0 percent for the Sharecroppers. As observed in Table 3.1, the coefficient of variation of land ownership among the categories of farmers analyzed was lower the higher farm size; however, it increased for the 100-200 hectares landowners. 1J. H. Herbst, Farm Management--Principles, Budgets, Plans, 4th ed. (Champaign, 111.: Stipes Publishing Co., 1976). 33 .mgwuauoga do asogm some >5 mm: vcm_ do mmmpcmogma men mmmmcpcmema cw macaw?» echo .zmw>gmpcw on» do we?“ mg» um wmgmpuwu ucm_ mo mucwx ages on» we mace asp mcwsssm an cmusnsou mm: cmozuoga any 25 upm; amen quop anew .>m>L:m mpasem ”muezom Lee e: 2%. 2%. Ame Lee is”... mm.mm Ne.mep AMWWW AMW.V Awmmw Ampomw Awwmmw_ AmmmMW a; oom-oop o~.mm Na.ae Ammww Awmmw Ammwm Awamwv Amwmww AwmeW a; oo_-Om Ne.ee em.- Amway Amwwv Ammmw Amwowv AMPWWW Awwwmv a; om-o_ 2.3 2.... Lame Ame a: 2m: Laws Exams .2; i. e Lee Lee 5.”. is 9%“ : Ls ages to eoeeaeee> mmepu ea mmeeepoz Ameeepdazv am: aces mmwwu we ucowuweemou ucmh mo mmmcm>< m.~nuonm_ .oz .mumz mu acoN .mgmuzcoga Co mmmpu An mm: ccmp use mNmm Exam mmmgm>oam Lo .mgmma mp op Ponce Lo Lune: wmoga com m.o can meow» em new m_ we mean esp :mozumn mu~zem com o.p op Peace mm umcwewv mp ecu pee: cowuuauoga m cw Lonmp do mgzmmmstupca m we accuses m>czm mpasmm "mugzom New mm N m mm cop N a Na Now mp Po NNa ape: wwawmwm mmN._ N_e N_ Na Nmm Nae e_ mo FNe mme ep eN Nam a; CON-oo_ NNo._ NF. N N mo_ mom F NN mam com mp me omm a; oop-om Mme Ne N o we mm N N FN New NN Nm Nee a; om-oN Nee op o o o_ mm _ e NN ONe NP Na mpm a; o_-o awe N o o N -- -- -- -- NNa ON as PNa demandedeeaem .5”; a... .fi ”Mum“ am“, E: is. ”um“ mm”. 5e .3 in a”? egg... Peach Loam; twee: mgmnaogumcmsm pcmcmELma Loam; xpwsmm mmmpu .Nuuoxm— .wz .mumz on mcoN m.mamuucme .uwcz m>wpuznoga Lea wpampwm>m Loamp mo mmaxhus.m.m mpamh 36 .mceueeeee we mmepe peep ee xpwpweepwe>e Leeep _euep esp Le>e gene, ee eewx wee ee memeeeeeeee wee memeepeecee em mecemwe echo .am>gem eFQEem Neeceem see ANN.mV Am.epv AN.eev aaaee>< NN eN NNm ape: we eeON mmN._ RNN.NV ANN.NV Ace.va ea NN_ eNm., a; eeN-ee_ _Ne._ Awe._e Nee.ee AN._NV we me New a; ee_-em Nee Amm.me Aee.ev Ae.NNV NN am New a; em-e_ wee Ape.mv fie.eNv Am.va a, em mmm a; ep-e Ame.ee Ape.mv Am.eev ewe om me e mme mceeeeeeecesm apaepee>< eaeepeee ease: peace ea: e_=e< meedaeoea maeoucez we mmepo .aeee "sate d_aapea>< Lane; to wave: Peace .NNIeNmP .uz .epez we eeeN .meeeeeeee we mmepe me epeepee>e maneuees cw ceeep epwce ece .eeeezuwpeee .eeEIepeee peueb11.m.m mFee» 37 force of sharecroppers, 0-10 hectares landowners, or 10-50 hectares landowners. Among all classes of producers, Sharecroppers had the most child labor available, which accounted for about 4 percent of the total labor. Producers with 50-100 hectares had the lowest share of child labor--about 2 percent. Capital. The analysis of capital is considered through three major categories: (1) capital in the form of all buildings including animal facilities, (2) capital in terms of machinery and equipment, and (3) capital in the form of work animals and livestock. The average investment in all buildings on Zona da Mata farms was about Cr$30,000, which included the proprietor's and employees' houses, storage facilities, barns, poultry and hog houses, and other small buildings (Table 3.4). AS expected, the larger the farm, the higher the investment. Total investment ranges from Cr$4,377 on sharecropper units to Cr$258,547 on farms with 100-200 hectares. Over two-thirds of the capital invested in buildings is for the proprietor and employees housing. The largest amount is invested in the proprietor's house, but the absolute value of housing increases substantially for both the owner and employees as farm size increases. Housing, while small in total value, represents over 80 percent of building investment for Sharecroppers and for farms with 0-10 hectares. This reflects the extremely low level of investment in buildings used for farm production purposes. Investment in grain storage is the second most important building item on most farms--ranging from 7.1 to 11.5 percent of total 38 .eseweu ecu cw eeumww weuweee we aeemepee esp cw “sesame>ew Ecew we memeueeeeee wee memegueeeee cw meeemww mew e .epe .ewe .meegm eewgeee .memeeem eeeweew mmewepwee eegpoe .ze>eem eweEem neeeeem emN.em Am.ev Ae.Nv Am.pv Am.ev A_.epv Ne.m_v Am.mmv ameee>< we_.m moo.p wwm omm.m omm.m omw.m emm.o~ eeez ee eeeN Nem.emN Ae.e_v AN._V A_._v AN._PV Ae.Nv AN.PNV Ae.eev mpo.m~ pmm.m Nem.~ oom.m~ omm.mp omo.mm Nem.mwp e; oomnoow Nee._N Am.mv Am.NV AN._V AN.m_v AN.F_V AN.N_V Ae.mev mom.e mom.~ omm ope.m woo.w mw~.~w www.mm e; oowuom mem._e Ae.NV Ne.mv A_.Nv AN.NV Am.PPV AN.m_e Aa.mmv mmm.m ome._ com mo~.m mow.e ewe.m owe.NN e; om-o_ emm.N_ Am.mv AN.PV Am.v Ae._v A_.ev Ae.ev Ae.ewv mme mom mm mmm wee.P mom mpm.e_ e; o_-o NNm.e Am.mv Am.mv Am.Fv Aa.v A_.NV Am.MNv aAe.eev mm, mew em mm owm omo._ emo.~ meeeeeeeeeeem mMHWflmemcflw eegpo memee: memee: mceem emeeeum eeaepeEm -emumwee mceeeeeee e me: see see agweo ewes . Feuew P w mewmee: we mmewu .Ameeweeeee wmm_ ewv swuonmw .Leeeem Peezeweeweme .uz .euez ee eeeN .mewewpweew weswee eee .emeeepm .mewmee; cw xeepm weuweeuuu.e.m eweew 39 building capital. The absolute amount increases with size of farm and reflects the fact that nearly all farmers produce some grains. Buildings for cattle are Significant only on farms larger than 10 hectares and become especially important for the two largest farm size groups. Most of this investment is for the dairy enter- prise, which exists mainly on larger farms that have enough pasture area to support a dairy herd. Dairy barns account for 13.2 percent of total building investment on farms with 50-100 hectares. Poultry and hog houses are low-investment items for all farm size categories-- a maximum of 2.1 percent for poultry and 3.6 for hogs. The figures shown in Table 3.4 reflect in relative and abso- lute terms the extensive nature of livestock production in the Zona da Mata. Very little capital is invested in buildings and equipment to support the livestock enterprise. This reflects the fact that live- stock productivity is low and is carried out on a traditional basis. The capital stock in machinery and equipment averaged about Cr$1l,500 (Table 3.5). Sharecroppers had the lowest level of invest- ment in this form of capital, and the total value of capital in machinery and equipment increased from the lowest class of landowner to the highest. The sharecroppers' largest investment was in tools and utensils such as hoes, axes, and hand-saws; for the other cate- gories of farmers, motorcars (including tractors, pick-ups, and other automobiles) constituted the major investments except for the large farmers, who invested more in machinery and equipment that was less common for other categories of farmers due to their coffee and large- scale dairy operations. 40 .uewseweew eee xewewcees cw eewEwmw>ew Pepe“ wee ew>e esewew we» ew ewemwp wewweee wee cw eeprmw>ew Egew we mwmeueweewe we» wee mwmwcwcweee cw mweemww wcwe .>w>e:m wweEem ”wweeem _em.__ AN.NmV Ae.Pev Am.va AN.¢V Am.mv weeew>< mmw.m www.e moe.w Pme wmp.w ewe: ee eeeN eem.mN Am.eee Ae.va AP.NV Ae.mv AN.mPV www.mm Pmo.m~ oom.m Pem.~ mwe.ow e; oomnoow e—F.mm Ao.omv Am.mmv Ao.mv A¢.¢v AN.©V epo.e mom.wp mwm.~ mwN._ wow.w e; oowuom FFN.__ Re.emv Ae.mmv Am.e_v Am.av AN.ev wm~.e mew.m mmo.~ mew mum e; omuow Pmm.m Am.epv AN.mmv AN.mNV Am.Nv Am.ev ewm mmm.w mmm om pom e; 0.10 amm.P aN.e_v Am.eNv A_.NNV Am.wv Am.mmv mNN nee emm MN New mewaaoeeweeem wewsepecm a . pewEQweem a meeu Ppee wwwmewwa a AMMMhmmwueflw xcwewguez uwmmwz weswc< mzewe .pewseweeu mHMemmMflm Pepe» segue eee mpeeu .mpeew .Ameeww~ewe wwmw :wv wwuewmw .eeuwwm weeeppzeweme .oz .euez ee eeeN .pewEeweew eee zewewcuee cw xweum Pepweeouu.m.m w—eew 41 The analysis of the percentage of capital invested in each category of capital presented in Table 3.5 shows that investment in plows, tools, equipment and utensils was about 15 percent of the total investment. This percentage was relatively low if we consider the relatively abundant labor in the study area. Work animals and livestock. The stock of work animals and livestock at the end of the 1977 agricultural year indicates that the Zona da Mata investment in work animals and livestock was about Cr$3l,000 per farm (Table 3.6). As expected, the value of this capi- tal increased with the size of the production unit. Indeed, a very large difference of investment in these classes of animals was observed among the many categories of producers. From one farm cate- gory to the next, investment in livestock more than doubled. Sharecroppers and owners with 0-10 hectares had less invest- ment in all kinds of animals than the average of the study area. This observation holds for both the value of the investment and the number of animals of each kind. Farmers with the highest percentage of land in pasture in the Zona da Mata also had the highest farm investment in cattle. The average number of cows per production unit was about seven, with higher concentrations among larger farmers. Relating data from Table 3.6 with land use, it was concluded that about 35 cows per farm with 100-200 hectares and about 20 cows per farm with 50-100 hectares were associated with pasture areas of about 103 hectares and 47 hectares, respectively. Investment in animals for work was second in importance in value terms. ‘42 .zee wee ew ewwmw— mpeewee we aeemwuew we» cw muewswmw>ew .meELew we eeeem we wmeuewugwe wee mwmwguewgee :w mwgemww wnwe .xw>cem w_esem "wweeem mpeswee eewuweeece eee mmepu ewweeeew eme.em mNe.eN_ aea.ae e_e.em eem.m eeN.N see: we w=.e> .eeaw A..V Ape.e A_e.e AN.V Ae.v Am.Ne ea e. eN Ne .N ea a=_e> m_es_ee ewumweee ewgwo Ae.me Am._e Am.me Ae.ev A_.__V Ae.eNv amm._ mPe.N .eN.N Nee.N .me Nme e=_e> om.m MN.e me.m .e.m NN.F ee._ Leasez mow; Am._v Ae.v Am.v AN.NV Aa.ev Ae.__v .em N_e._ NeN NeN men NNN a=_a> Ne.NN .N.ee ap.Nm mm.em ea.e. _e.e_ meewa we amassz Acupeee Ae.ev Ae.me Am.ee Am.Ne Ae.mv A_.ee eNe.N _MN.N mme.m wee.N N.m eNN a=.a> _e.e ma.eN .e.e_ eN.e MN. mm. tease: mw> pcu Ae.ae A_.N_v Ne.ee Am.ee “a.Ne AN.m. Nee.m ee_.aN mNN.N mNa.N Ne. we a=_e> NN._ Nm.e ee.m ee._ e_. on. Lease: :mxo tam mp—am AN.eee AN.eee Am.NNV Ae.eev .N.Nee Ne.mmV NNN.eN aeN.e__ .e_.pe ae_.NN mmN.N wee a=_a> Na.e Ne.mm ee.a. ee.N mN._ me. Leasez mIOu AN.e_v Ae.ee AN.ev Ae._.e Ne.mNe eA_.e.e mN_.N eeN.__ eea.e epe.e MNN._ ema e=_e> em._ NN.e Ne.N NN._ ee. NN. Lease: aweswee see: awe: ae aeeN a: eeN-ee_ e; ee.-em e; em-e_ a; e_-e meaaaeedaeeem m_ae_e< we meeww NNN_ e_e NN-eNm_ .ez .eeez ee eeoN .me_e= eowedaeeea ew Neeemw>_F eee .Ameeww~=cu m—eawee ewe: we wepe> eee eweE=e wmeew><-u.c.m wweew 43 The investment in cows was the highest among investment in animals for all producer categories. Figures in Table 3.4 indicate that for landowners with more than 10 hectares the percentage of investment in dairy barns was the highest relative to investment in all animal facilities. About 25 percent of investment in animals of Sharecroppers was in swine, and the figures in Table 3.4 also indicate that these producers' highest investment in animal facilities was in hog houses. The percentage of poultry investment over the total investment in animals decreased with farm size; however, this tendency was not observed for poultry facilities. The 0-10 hectares landowners had the lowest percentage investment in poultry houses, and the 10-50 hectares landowners were those with highest investment. Despite sharecroppers' and 0-10 hectares landowners' rela- tively low investment in poultry and swine, it was recognized that such animals can represent an important source of financial income to these producers. Final comments on resource endowment. Table 3.7 shows availa- bility of capital and labor per hectare. The average value of invest- ment in buildings and animal facilities per hectare for Zona da Mata was about Cr$l,500. However, this amount did not express the pro- ductive capital directly and, when the values of proprietors' and employees' houses were excluded from the calculations, the average investment in productive buildings and animal facilities decreased to about Cr$480. The relatively high investment in buildings by land- owners with 0-10 hectares decreased from Cr$3,100 to Cr$480 with the 44 .eewwe>w:ew mxeeuees cw eweemewzw .mweswee wwumeee ewcee eee .mmwe .xewweee .cwxe eee wwwee .mw>wee .mweewee xeez we wewe> mweewwcHe .ucwseweew eee Acwewguee ewswe ece .meew Lewes .meee Zeeéeewee use 3:3 .33.. 375:3: eee .wewseweew..mwe3 we wewe> mweewweww .mwmee; wwzeweew eee aweuwweeeee we wewe> meweewexw .Aev me wEeme .mmewewwee ewspe eee .mwwwwmmwe .mwmee; aeuweee .mewee .mwmeesweez .mwmee; wwaeweew eee aeeuwweeeee we mwewe> mweeweeHe .>w>e:m wweEem "weweem pm. m~.o mm.o w—.w w¢.~ mm.m we:\weeew ewwsu uh. xuwwwnewwe>< . . . . . . @m£\LOnm_. om N on _. DO _. 0m N MN OF mm N C0203!“ Psfim .._.0 NA“. P _. wan _. wm>< . . . . . . wee\eeee_ ee.PNN.P ee._MN.F ee.NPN.P ee._mm.P ee.mem ee.eee xemee\xeoemw>ww eee mpeewee 3 cw pewEumw>ew we wewe> . . . . . . we;\wewseweew e;\mwwuwwwwew oo.wwe oo.mmm oo.wmm oo.mmm oo.wwe oo.mmw weswee eee mmewewwee w>ww -eeeeee cw newswmw>ew we wewe> oo.m~m.w oo.www._ oo.m~o._ oo.m~o._ oo.www.m oo.omw ee;\mwwwwwwwew weewee eee mmew uewwee ew uewEwmw>ew we wewe> e; e: e; e; mcweeeeu ewe: ee eeeN oomnoow oownom omuow ow-o tweeem Low wmegw>< mmepu quseowa .NN-eNm_ .ez .eeez we eeON .weeedwe Lea wew_wee_we>e Loee_ eee weewaee--.w.m wwaew 45 new calculations, and similar decreases also occurred with other classes of producers. The average value of machinery and equipment per hectare was about Cr$460. The sharecroppers, 10-50 hectares and 50-100 hectares landowners' investment in machinery and equipment was below the average of the study area: the 0-10 hectares and 100-200 hectares landowners were those farmers with more investment in such a form of capital per hectare of farm land. The average investment in work animals and livestock per hec- tare also varied among the many classes of producers. The sharecrop- pers and 0-10 hectares landowners had investment in animals below the study area average. The 50-100 hectares and 100-200 hectares land- owners' investment in such a form of capital was about the average for the whole region, and the 10-50 hectares landowners were above that average. The situation of labor availability seemed to be reversed. Those who owned more land, as expected, used less labor per hectare since they invested more in capital in the form of machinery and equipment. It was observed that the Sharecroppers and the farmers who owned 0-10 hectares and 10-50 hectares had labor availability per hectare above the average of the study area in all forms of labor: adult men, adult women, and children. Resource Use and Productivity A description of resource use and productivity of the Zona da Mata farmers is presented below. The use of land for selected 46 activities is shown, and the output per hectare of important crops is compared among groups of producers. The use of labor and output per man-day of selected crops were also compared among farmer groups, and analysis of variance of ratios involving land, capital, and labor are analyzed. This section concludes with an analysis of farmers' use of credit and their participation in cooperative associations. Land use. Approximately 60 percent of the land of the region is used for pasture (Table 3.8). Coffee, another nonsubsistence activity, occupied a considerable part of the region. Further, the amount of land allocated to these two uses increased with the farm size. Sharecroppers used 10 percent of their land in coffee planta- tions. For landowners, the percentage of farm land allocated to coffee plantations decreased from 11.6 percent for 0-10 hectares land- owners to 5 percent on 100-200 hectare units. Because of the topography of the study area, pasture area increased with farm size. Sharecroppers used the least land for pasture. About 19 percent of their land was in pasture, and this increased to about 75 percent for the 100-200 hectares landowners. The other three crops to which the Zona da Mata producers allocated most of their land were corn, beans, and rice. About 9 percent of the whole study area was used for corn production, and about 6 percent was used to produce beans and the same percentage to produce rice. About 40 percent of the cropped area was used to pro- duce corn, and about 26 percent and 23 percent were used to produce beans and rice, respectively. These crops were of great importance 47 Table 3.8.--Use of the land in Zona da Mata, MG, for selected activities, by group of producers, 19"6-77.a Group of Producers Zona da ACthltles Share- 0-10 10-50 50-100 100-200 Mata croppers ha ha ha ha Average ------------------------- Hectares-------------------------- Corn 2.03 1.01 2.70 3.76 5.08 2.36 (35.1w,C (17.5) (9.8) (5.4) (3.7) (9.4) Beans 1.47 .59 1.67 2.73 4.08 1.57 (25.3) (10.4) (6.5) (3.9) (2.4) (6.3) Rice 1.20 .56 1.42 3.02 2.83 1.39 (20.8) (9.8) (5.5) (4.4) (2.1) (5.6) Sugar cane .05 .16 .39 1.24 .79 .37 (.9) (2.8) (1.6) (1.8) (.6) (1.5) Tobacco .16 .05 .18 .14 -- .13 (2.9) (.8) (.7) (.2) -- (.6) Fruits .01 .03 .08 .03 .11 .05 (.2) (.6) (.3) (.05) (.1) (.2) Vegetables .03 .04 .05 .05 -- .04 (.6) (.7) (.2) (.09) -- (.2) Coffee .58 .66 2.09 3.71 6.92 1.76 (10.0) (11.6) (8.2) (5.3) (5.1) (7.0) Pasture 1.14 2.09 14.20 47.11 103.41 15.03 (19.2) (36.3) (47.8) (67.9) (74.9) (59.9) Source: Sample survey. aBecause of intercropping in the study area, these activity areas may exceed 100 percent of the land held by the producer group. b The figures in parentheses are the percentage of the total farm area allocated to specific activities per group of producers. cTo compute the percentage of the cell, the declared farm area was used instead of a summation of the isolated activities. 48 for the smaller producers and sharecroppers. Sharecroppers allo- cated 35 percent of their land to corn production, more than any other use. This percentage decreased as the farm size increased; the 0-10 hectares landowners allocated about 17 percent of their land to corn production, and this decreased to 4 percent among the 100-200 hectares landowners group. This pattern of land allocation is similar for beans and rice. Sharecroppers allocated more land to these crops: about 25 percent of their total land was in bean production, and about 21 percent in rice. This percentage decreased for each size group of farms up to 100-200 hectares. The largest landowners used only 3 percent of their land for beans and 2 percent for rice. ' Tobacco production showed a similar pattern. The average area allocated to this crop in the Zona da Mata was less than one hectare per farm. The Sharecroppers allocated about 3 percent of their land to this crop, and this percentage decreased as the farm size increased. Landowners with 100-200 hectares did not report any tobacco. The percentage of the total farm area devoted to sugar cane was greatest among the 0-10 hectares landowners: about 3 percent of their farm area was allocated to this crop. Producers with 50-100 hectares of land had an average of 1.2 hectares of sugar cane, which represented nearly 2 percent of their land area. Land used to grow vegetables and fruits in the Zona da Mata was less than 1 percent of the whole area. About 1 percent of the area held by the 0-10 hectares landowners was allocated to each of 49 these activities. Again, percentage of land used for these activi- ties decreased with farm size. Landowners in the 100-200 hectares category did not produce vegetables. Table 3.9 shows the means and significance level (ANOVA) for test of the hypothesis of equal output per hectare for selected farm activities among the producer categories. A general observation of the figures presented in this table is that outputs per hectare of all agricultural activities were not statistically different at the conventional 5 percent level among the different classes. This result seems to suggest that the technologies used by all farmers were not substantially different. However, this result should be interpreted with care because it deals only with output per hectare of land. It should be noted that corn production per hectare was statistically different at the 18 percent level, and other crops considered as subsistence crops, such as beans and rice, were statis- tically different at the 46 and 74 percent levels, respectively, among the different producer categories. The landowners with 100-200 hectares were those who achieved the highest yields of the tradi- tional crops (corn, beans, and rice). Following this category of farmers, Sharecroppers had the highest yield of these traditional crops in the sense of return on production per unit of land. Share- croppers attained the highest yield per hectare of sugar cane and the highest value per hectare of vegetables. Production of tobacco and coffee per hectare was highest among the large farmers. The production of coffee per hectare was highest among landowners with 50 .wuw .mweeewee .weeuuww .wmeeeee .ewweem .mwewesew eweewwefiw .wuw .me>e=m .mwweeewewe .meeewe>e eweeweew em—e ww wweem wwesm e eeezze.mpe=eeee ewes we» we meeeeee eee meweew meewww eweewueHe weewew; ewe meewwNeeu Nwmw cw wewe> :ewpweeeee "wwweepwmw> weewww; ewe meewwNeeu wwmw cw we—e> eewuueeeee ”meweew weewww;\mx om "emeeeeeee wwwweu weewuw;\mx mp "emeeeeee= eeweeew weewuw;\meee "weew eemem weeeuw;\mmee m: cm ”wow: weeuww;\mmee ex ea ”meewm weeeewe\mmee a: on "eeeu nmzewwew me eweemews we: weeeww; ewe weaveoe .»w>e:m wwesem "weeeem wmo.ew pm. -1 oom.w Ne~.~w www.mw www.mw emwweeewmw> pew.w mm. omw mwo.e Now.o mm_._w oom.m empweew m.mw we. o.m~ m.~w m.mw m.ow m.ww wwwweu m.ow wm. I- m.e~ m.ew w.~w w.mw ewweeew o.em mm. —.NN w.m~ o.wm m.mm N.mm weeu eemem N.o~ ew. N.~N o.ww w.w~ m.mw m.wm wwwm N. we. m.m w.e m.e m.m ~.m meewm m.~m mp. m.e~ m.m~ m._N m.mw m.e~ eeeu ww>w4 1 1 1 1 meweeeew wumwuww .wwemwm e; com cop e; cow om e; om op e; ow o tweeem mwmweeeweem <>oz< meew: uwuw25uwe< .wwuewmw .w: .epe: ee eeeN .mewweeeee we meeeem pewewwwwe meese mwmweeewwew ewewwwwm we eweepuw; ewe peevee w>wueeeeeeuuu.m.m wweew 51 100-200 hectares, and tobacco production per hectare was highest among the 50-100 hectares farmers. The highest return in cruzeiros per hectare of fruits was among the 0-10 hectares farmers; the large farmers had the lowest return on this activity. Labor use. The average number of days of labor used per hectare for selected crops is presented in Table 3.10. From the ANOVA results, it is observed that use of labor in rice, bean, and corn production was statistically different at the 1 percent level among the five classes of producers. Labor used for production of other crops, such as coffee, sugar cane, fruits, and vegetables, was not statistically different at the 5 percent level among the various categories of producers. Farmers with 100-200 hectares of land tended to use less labor on subsistence crops such as rice, beans, and corn. Overall, the enterprises to which farmers allocated the most labor per hectare were fruits and vegetables. Table 3.11 presents the means and significance levels (ANOVA) for the comparisons of output per man-day for different enterprises by different categories of producers. It was observed that produc- tion of rice per man-day was not statistically different among farmer groups at the 5 percent level. Concerning the production of beans and corn per man-day, the analysis indicated that the 100-200 hec- tares group had the highest returns. The O-lO hectares group had the lowest production of beans and corn per man-day. The return to labor on other crops was not statistically different at the 5 percent level of significance. 52 .>w>e=m wweEem ”wweeem Mme MM. -- Me. we, NNN MPN Meeeeewew> em N_. ee -- ee ee_ eN Meezee ee ee. NM eN Me em em weee eemam Mm eN. eM MM me me Me wwwwoe MM Mee. MN eN eN we MM eeoe MN wee. Ne eN NN mM FN Meewe em Nee. em eM pm we em edee eew: .wwnmum a; eeN-ee_ 8; eee-em e; eM-ee 8; ee-e Mwwmmmme wwWwwnwwwm eeewewe <>ez¢ Meme: deewEeeee< Eeee .eeueeew .o: .eee: ee eeeN .meeeem eweeeeee eewewwwwe meese meeew ewpwwwwm eew weeeew: ewe ewme eeeew we maeeuee:1-.o_.m wweew 53 meewwweew wwe— cw wewe> eewwweeeee "wwweeewmw> meeww~eew weep cw wepe> eeweeeeeee ”mwweee me Me Me eM meow "weew eemem "emeeeeeee eweeeew "emeeeeee= wwwweu mmee me on "eeeu mmee ex on "meewm ween ex om ”wwwm ”mzewwew Me wewz ewme weepee we meweze .xw>eem wweEem "wweeem Me. mM. -- mNe we. Me me_ Mweeeewew> emN em. mNM eee MeN Mee eeN Meeaee mm. me. em._ NM.~ me. am. Me. weed eeezm mN. em. -- eN. eM. Me. eN. oeaeeae em. «N. we. Ne. em. ea. Ne. wwwwoe Ne. ee. NN.F ee.e No. Me. em. eeoe _M. ee. Ne._ NN. NM. Ne. MN. MeewM em. e_. ee. we. Nm. ee. we. weee eew: _w>ee 8; eeN-ee_ e; eee-em e; em-e_ a; ew-e WWMMMNMO weeweww .wwemwm Mwaeeewwem <>ez< Meme: deewEeeee< .wmuommw .w: .epe: ee eceN .meeeem ewweeeee uewewwwwe we meeew ewwewwwm eew exeeueee ewe weeueouu.ww.m wweew 54 The figures summarized in Table 3.11 seem to indicate that producers with more land were using technologies that resulted in more return to labor used in the production process, such as machinery and other equipment. This inference is supported by the total invest- ment in machinery and equipment of this class of farmers, presented in Table 3.5. Comparisons between the figures presented in Tables 3.10 and 3.11 seem to imply that in bean and corn production O-lO hectares landowners are using relatively more labor than other categories of farmers, or possibly other production factors have been used by the other categories of producers in substitution for labor. Capital use. The analysis of the use of capital was broken down into four classes of capital: 1. Farm Assets (K1). This class of capital included the value of all machinery, equipment, and work animals as reported by the farmers at the time of the study. 2. Livestock (K2). Livestock capital included the value of all cattle, poultry, and swine reported by farmers in the survey. 3. Operating Expenses (K3). This form of capital was com- puted as the value of all inputs used in crop production, such as seeds, chemical fertilizers, lime, pesticides, and the value of inputs used in livestock production, such as medicines and salt. 4. Permanent Structures (K4). This category of capital included the value of permanent structures such as barns, storage facilities, and other permanent structures on the farm. Two other aggregated classes of capital are also considered. They are defined as Total Capital 1 (KA), which is the summation of K]'+ K2 + K3 + K4. Total Capital 2 (KB) included KA plus the value of the proprietor's and employees' houses. 55 The analysis performed in this section is an extension of the analysis presented in the resource endowment section. Capital/land ratio and capital/labor ratio used in agricultural production units in the Zona da Mata are of primary interest. The analysis of these ratios is expected to bring realistic insights of how capital resources are related to land and labor. Analysis of variance is used for com- parisons of the ratios for different farm classes, and at the end of this section some conclusions are drawn about the results of that analysis. Table 3.12 presents the capital/land ratios by the categories of producers in the study area. With the exception of the operating expenses/land ratio (K3/L), all other sharecroppers' ratios were the lowest ones among the groups of farmers. Consistently for all kinds of land (total farm land L, crop land L], and pasture land L2) con- sidered in the ratios, farm assets/land ratios were not statistically different at the 5 percent level among the categories of producers. This result indicates that investments in machinery, equipment, and work animals per hectare of land do not differ statistically at the 5 percent level among farmer groups. Regarding other figures pre- sented in Table 3.12, we may conclude that there is not a consistent correlation between ratio sizes and farm sizes. Tables 3.13 and 3.14 are analyzed together. Table 3.13 pre- sents different forms of capital/availability of labor ratios, and Table 3.14 presents similar ratios, but labor actually used in the production unit is used instead of labor available to the producer. (Available family labor, available Sharecroppers labor, and available 56 .»w>eem wweEem "woeeem eeM.e ee. eNM Mme MMM.e NeM.e MM ;e\ Mee.N Ne. NeN.N e_e.M Nee.N mem.e e_M :\: e NeM we. MMM MMM FNM NNN MeM e\ e eeew\mweeueeeum newceeewe «MN ee. MNN eeN Mee wee._ Ne NT\M¥ eee Me. Fee._ Mee _Me._ eee _ae :\M M eMM ee. eeN ewe MeN NNe NMM \ e eeew\mwmcwexww ecwueeweo eM_.N MM. eew.e wee._ e_e.N NeN.N NNe NN: Nee.M we. wee.e eee.e MMF.N .Ne.N Mee :\N¥ N .eM ee. Mee.e eme.e eeN.e eme eew e\ eea_NxdoeMa>we MeM._ me. New eMM NeM._ e__.N NMN Ne\e¥ NMM.N we. eeM.M Mme.e NeN.M NPM.P eew ;\J p MNe eM. Nee Fem eMe mew eMe e\ e 2.0 _.\mu.wmmw ELM“— Mee.m eo. Mee.M eNM.e Pee.m MNM.e_ eee.e Neeme eee.Me Ne. MNN.PN MeM.NN MMN.MN eeN.Me eme.N MeM.M ee. eMe.M .MN.N NMN.M NMe.e MMM.N fleeex UCMPNN vawnmu PMHOH mMe.M _e. e_M.M ema.M eeM.M eNM.N Me_._ T\« ewe.M ee. eee.ee MNm.e_ MMM.Me MeN.M eMM.e :\f < MNN.N Fe. NMM.N ewe.N Nem.N NNe.N eew.e e\ e eee_\_ Peewemw Meeee ww>ww e; e; e; e; meweeeew eew: .eeemem eeN-eee eee-eM eM-ee e_-e -eeeeM Mwmmwewefiwawewmumwv weewcwo <>oz¢ Meew: wwewscuwe< . e e 4 w . u .eeuowmw .o: .eue: ee eeeN .mewoeeeee we meeeem uewewwwwe wee meese mewpee eeew\—epweew we mseew wewewwwwe we mwmxweee w>wueeeeEeuun.~w.m wweew 57 MM, NM. MMM MM MM MN NM NZNN MN MN. MM MM MM PM -- FEm. MM ee. MNN MN NM __ M 3N; MM Me. MM PM MN M M 3N; seem; we MNMMMMMMMMMMNMNMMm< ELMM NMM PM. NMM MMN NMM NNM MMN gzN NNM eo. MMM NMM MNM MMF -- 3N MNN ee. MM.._ eNM eNN NNM MN §N MN_ MO. eMM NNN MMM NM MN zNMg Noam; we MMMMMMMMMM>e4 AxmeuemeNMegmeeguv ”mwum .wwawM eeN-ee. eeM-oM eM-eM ew-o -MLMMM MowMMN MOMMA wo MUMMMN_MNPMMM M M <>ez< McMMz UMMMEMNML< .unummmp ocean Mewpmg Leeew we zawwwemwwe>ereuweee we magew uemgewwwe we Mwmxwece m>wpeweeseuuu.mw.m eweew .mz .muez me meeN .mgmeeeege we Meeegm pcmemwwwe ecu 58 .am>gzm mpasmm "mugzom ~_F mm. cup mm_ ¢o_ em om m3\wg mm co. «mp mm mm om .. P3\¢¥ mm oo. omp mm Fe m_ m 3\ g e pm co. me pm mm o_ v 3\ g Loam; mo xpmrwnm_wm>n \mmgauusgum ucmcmegmm on FF. we om mm mm NP m3\M¥ Pm me. me mm on mp -- _3\mg mp co. mu mm op m m 3\ x m m oo. op up a m m 3\ x Loam; $0 zumpwampwm><\xuopmm>w4 m; a; m; m; mgmaaogu Pm>m4 - Ammuucms\mogwm~:guv com: . u n . -mgm Fmgmcmw $wcmwm com cop cop om om op op o cm mowumm Loam; mo xuopm\~mppamu <>oz< mammz ovumsspwg< .umacwpcoo--.mp.m mpamh 59 .zm>gzm mpasmm "mugzom me Po. oo_ mm mg mg m 3=\¢¥ mg=g_:uwgm< cw umm: Loamg\mmgzguzgum pcmcmEgmm m_ oo. mu m. o_ m_ F_ 3=\mx mgzupsuwgm< cw ammo gonmg\mmmcmaxm mcwumgwao so co. o_¢ Fm, co. mm 0F 2=\~¥ mgsppzuwgm< cw can: Loam;\xuopmm>w4 we oo. mm_ cm we mm NF 3=\F¥ mg=¢_:uvgm< cw umma Loam4\mpmmm< Exam mmm co. mfim w~¢ Nam qmm om z=\m¥ mgappzuwgm< cw yum: goam4\~ _mgwgmu Page? Fmp co. mg“ mmm com N¢_ Rm 3=\<¥ mgappsuwgm< cw umma goam4\_ Fmgwgmu FQpOH Fw>m4 on m; m; m; mgmgqogu cam: .ypcmpm oom-oop oop-om omuo_ op-o -mgmsm Axmu-cme\mogwm~zguv ngmcmu <>oz< meow: Uwume;uwg< ovumm cum: gonm4\_opwamu .Nnumnmp .oz .mpmz av mcoN .mgwuanoga $o mnaogm pcmgm$$wv as“ macaw mowgmg cowguzuoga pmgzppzuwgmm cw umm: LonmP\Pmpwnmu we maze; acmgmmwmu we mwmxpmcm m>wpmgmasoouu.¢p.m mpnmp 60 hired labor are represented by w], "2’ and N3, respectively, and UN is labor actually used on the farm.) Labor availability is very well known by producers and is one of the variables that influence farmers' decisions on investment of capital in the production process. In most of the cases analyzed in Tables 3.13 and 3.14, the ratios increase with farm size. Statistical results presented in Table 3.14 indicated that all ratios of capital/utilized labor were statistically different among farm classes at the 5 percent level. Laborlland ratios. The figures in Table 3.15 dealing with labor/land ratios complement the analysis of capital presented above. In this table all sources of available labor as well as labor actually used on the farm are considered. The general observation of the figures in this table is a tendency for the ratios to decrease with increasing farm size. Comparing the figures of this table with those presented in Table 3.l2, we may conclude that there is not a straight- forward trend of land, capital, and labor ratios with increasing farm size. This fact may suggest that as farm size increases, different enterprise mixes are emphasized (which requires different resource combinations), or resources of land, capital, and labor might not be homogeneous in the study area. However, there seems to be a trend for larger producers to substitute capital for labor when labor is not available. Further analysis will clarify this point. Use of credit and cooperative associations. The use of credit and cooperative services is analyzed in this section. Table 3.16 shows the average per farm and per group of farmers as well as the percentage of farmers who used technical assistance along with credit. 61 .xm>gam mPaEmm ”mocaom Rm oo. NP mp we «PP PP m4\3= FR pm. mm mm ow em we _4\3: om co. m m up mm um 4\3: ncm4\mg=ppzuvgm< cw mum: Lonm4 cc co. m «P me mm, -- ~:\ mm mm. mm Fm mm nm N_ ;\ mp oo. N m o_ we om 4\m3 vcm4\mpanwm>< gonmw vmng we co. up mp mm app .. ;4\ mm mm. me mm mm em -- :\ mp co. m pp m_ um -- ;\ ucm4 \mpampwm>< Loam4 mgmamogomgmsm o4, oo. N_ mp ow «RN KN N4\? mvp oo. Pu um um, mmm wm_ :4\3 m“ 00. m m um opp mop 4\3 vcm4\anm_wm>< goam4 x—wEmm mmp oo. mm um mm Pam um .4\3 um, Po. mFP “op Nm_ emu wmp 4\: an co. m_ op mm mpp ~m_ 4\2 flucm4 \mpnmpwm>< L83 pouch a; a; a; m; mgmaaogu pm>m4 . n u l - Augmwom;\mxmuucw24 pumwumw .ww=m_m com cop cop om om op op o wgmgm momuwm u:m4\gonm4 <>oz< meow: owuw55pwg< .Nuuoua— .oz .mumz mu mcoN cw mcwuzuogg mo mazogm acmgmwemu mcu ocean mowpmg ucmp\gonm_ now: new m_ampwm>m mo magoe pcmgwmwwu mo mwm»_m:m m>wgmgmasoouu.m_.m mpnmp 62 .moe_m~:gu Rump cw mmgzmwum .>m>g:m mpasmm "muezom pwumgu gum: mucmpmmmmm mm mm om mm 2 :8 .253 um> .8qu on: mgmzoggoa we mmmpcmuxmm mm mm we em o_ meaegae co mmm_u can mgmzogeon we mmmpcmogma NF em mop we m_ memzocgoa mo L.3532 omm.upp ovo.ee Nvm.wm omn.~_ mFm.N muwumgu mo pczosm mmmgm>< a; oom-oop a; oopuom m; omuop m; o_-o mgmaaogumgmgm mEmpH mgmsgmm eo aaogw mu m:o~ .mnummmp .92 .mng .mLmEme mo mazogm umuum_mm an mmmcwmsn FmgzpF=owgmm ms» cw “wumgu mo mm2--.m_.m wpnmh 63 The amount of credit is shown in the table in aggregated form, including credit from official institutional sources as well as credit from noninstitutional sources, such as from other farmers or from private businessmen. The figures in this table seem to confirm the descriptions given in the introductory sections of this study concerning access to credit by wealthy farmers. Sharecroppers used less credit than the other groups. An average of about Cr$2,800 was contracted per farm, and only about 10 percent of the Sharecroppers used this ser- vice. The percentage of farmers using credit increased substantially as the size of farm increased. About 63 percent of the 100-200 hectares landowners used credit, which was close to nine times more credit than landowners with 0-10 hectares. Concerning the use of technical assistance together with credit, which is a precondition for institutional farm loan, the figures in Table 3.16 seem to suggest that this mode of credit was positively associated with farm size. The reason for this may be related to the preference of extension service to work with larger farmers. About 80 percent of the largest producer group received technical assistance, in contrast with about 30 percent of the 0-10 hectares landowners and 12 percent of the sharecroppers. Concern about the tendency of the extension service to work mainly with larger farmers in Brazil has been pointed out by Souza.1 1Antonio Fagundes de Souza, "Pesquisa, Assistencia Técnica e Extensao Rural," A Homern e o Campo, Fundacao Milton Campos (Brasilia: Senado Federal-Centro Grafico, 1976). 64 The extent of participation in farm cooperatives is presented in Table 3.17. Sharecroppers did not use cooperatives at all, and only 6 percent of the 0—10 hectares landowners were active members. The percentage of participation jumped to about 30 percent, 50 per- cent, and 60 percent for the 10-50 hectares, 50-100 hectares, and 100-200 hectares landowners, respectively. Farmers were asked about the kinds of services provided by the cooperatives with which they were associated. The figures in Table 3.17 seemed to imply that as farm size increased, farmers were more interested in “marketing cooperatives" than input-supply coopera- tives. Only the group of farmers who owned 0-10 hectares of land had a high percentage of their members associated with cooperatives that had both supply and marketing services. However, it is not known whether all of these kinds of cooperatives were available to all the farmers in the study sample. Summary and conclusions. As expected, family labor was the most important source of labor for the smaller producers of the Zona da Mata. About 75 percent of the total labor available for these producers was family labor. The presence of permanent sharecroppers and hired labor becomes more frequent as farm size increases. Perma- nent sharecroppers are an important source of labor for the larger landowners. The supply of family labor apparently is not adequate to meet the farm demand. Investment in all forms of capital also increased with farm size. The difference in resource ratios reflects the differences in the quantities of land ownership, capital, and labor held by different 65 .mceELew megeuem; oomueow mp eee .mLeELew meeeeee; eepuem mm .mgesgew megeuee; omuep mFN .mgesgew megepee; opue NNF .meeeeeeeegesm wmp we eemeQEee me: ePQEem eew "euez .»e>gem e_e2em "eegeem m.em m.me m.mo m.em me>weeemeeee new: eeeeweemme uez . . . . me>weegeeeee _ mm o ow N FF F e mcwumxgee ece aweeemumpeeeH e.pm ¢.NN m.ep e.F mm>weazeaoou ecwemxemz m.m m.o~ w.e o me>wpegeeeee aweeem1meeeeH me>waeeeeeeu we eewx Lee mLeELem empeweemm< we emeaemeeee mgeeeeeee we mme_e Lee mm mm _m m mgesgew empeweemme we emeuceegee NF om mo N mgesgew emueweemme we Leesez me>wpegeeeeu eew: eewueweemm< e; oo~-oop e; oepuom e; omuop e; opno mLeELem we geese mseuH .wwuewmp .ez .euez ee eeeN .meewpeweemme e>wpegmeeee cw eewpeewewegee .meeELeuuu.ww.m eweew 66 groups of farmers. However, it is questionable if the share-leasing institutional arrangements between permanent sharecroppers and land- owners led to the optimum sharing of costs and returns of farm opera- tions and, consequently, these arrangements implied barriers for augmenting resource ownership (capital and land). It is also ques- tionable, on the other hand, if relatively abundant family labor and permanent sharecroppers' labor compete with increasing use of capital services. Another factor that may be responsible for the differences in resource ratios is that rental arrangements by farmers may be a means of gaining control over greater quantities of land and capital. The use of credit by sharecroppers was very low relative to other classes of producers. We may conclude that besides the credit limitations other factors such as education and share-leasing insti- tutional arrangements may be a limiting factor for expanding capital use of this class of producers. In turn, it seems that relatively greater capital/land ratios observed among landowners can be explained by the fact that even hav- ing permanent sharecroppers on their farms, they may be led to rein- force returns in contracts with sharecroppers that involve use of landowners' capital. 0n the other hand, in enterprises in which share- croppers do not participate, the landowners have no intrafirm disso- ciation of costs and returns and, consequently, they would tend to increase capital-resources use, which is facilitated by the availa- bility of credit. The relatively high labor/land ratios of sharecroppers imply that their production systems are not restricted on labor availability. 67 For this category of producers, the marginal productivity of labor can be expected to be low because low capital/labor and high labor/land ratios were observed. Returns to land and capital can be expected to be relatively higher than returns to labor. However, this is a pre- mature conclusion because resource productivities also depend on rela- tive values of elasticities of production. Overall, the ratios capital/labor, capital/land, and labor/land seem to imply low levels of technology employed by the small farmers of the Zona da Mata. The use of land analyzed in this section implies two marked uses of land in the study area: (a) production of grains and (b) use in pasture. Production of grains (corn, beans, and rice) seemed to be of relatively greater importance for sharecroppers and smaller pro- ducers. Land used for pasture, on the other hand, increased with farm size. It should be pointed out that pasture land refers to natural pasture which is generally part of the farm that cannot be used to produce grains and other crops due to the hilly topography and low soil fertility. Management Efficiency and Family Composition Management efficiency. In the literature on agricultural development, attempts have been made to use proxies for management efficiency to better explain production variation.1 In this section, 1See, for instance, Martin Upton, "The Influence of Manage- ment on Farm Production on a Sample of Nigerian Farms," Farm Economist, 1970, pp. 526-36; J. Bessell, "Measurement of Human Factor in Farm Management," International Journal of Agrarian Affairs, July 1969, 68 four proxies for management efficiency are analyzed. They are: (a) farmer age, (b) farmer education, (c) number of days the farmer worked off the farm, and (d) commercialization index. The age of the farmer as a proxy for management efficiency has to do with acquiring experience throughout the years in the busi- ness; education to reflect the investment in human capital; number of days the farmer worked off the farm to indicate contact with the outside world or different society; and commercialization index to reflect attempt to produce surplus for the market. For the whole sample, the arithmetic mean of the farmers' age was about 50 years (Table 3.18). The ANOVA conducted on this variable suggested that ages of farmers in different categories of farms were not different at the 5 percent level of significance. The age distribution of the managers within each category of producers is presented in Table 3.19. About 33 percent of all share- croppers were included in the age class between 40-50 years old. The ages of the producers who owned O-lO hectares, lO-SO hectares, and 50-100 hectares of land were concentrated in the category 50-60 years old. The larger farmers (100-200 hectares landowners) were the oldest group of producers, and their ages were concentrated in the 60-70 years range. A general description of the education of family heads and their wives is presented in Table 3.20. In general, the percentage of illiteracy decreased with an increase in farm size. Thus, the highest Supplement, pp. 37-44; and Andrew B. Tench, Socio-economic Factors Influencing Agricultural Output (Saarercken, SSIP, 1975). 69 .ze>g=m ewesem "eugeem Aemeeeeeseev xeeew Lepemm Pegeuweewgme use mm no. em NF mm _m m eewmpee seew esp wwe eexgez when we geesez genome wegep—euwgme em oo. o mm oe me me_ cw use Egew we wwe eexeez mane we Leesez em 8. em 8 s em 3 3.523 2% cam: Pa>a4 a; oo~-ee_ a; oo_-em a; em-o_ a; o_-o memaaoeemeaem .223 $3.5 $335, <>oz< meme: ewpe52ewc< .wwuewmp .wz .eeez ee eeeN .mgeseew we memmewe pcmeewwwe en» meese peesemeees ep empepeg mepeewee> eeueewem we mwmxpeee e>wuegeeeeuu-.mp.m epeew 70 .mmepe eme ecu cw agemeeee Leeeeege use we memepeeegee ewe memegeeeeee :w megemww egwe .xw>gem ewesem "eugeem AePV hope Ame 4me Ame mm m e w, e m ta>o new ow 4emv ANNV Aepv 4mmv Ame mm m mp mm mm A em-oe APNV AFme Acme 4m~e 4e_e we, e e_ ee em em oe-om Ampe Acme heme AONV Anne New m N_ mm mm me om-o¢ 4_Ne 4m_v Amwv 4¢_e AQNV eew a m mm n_ em e¢-om 4Ne Awe Ame Ampv am -- _ m e A_ om-e~ Ame Awe a4_v NP -- -- e m P em cage mma4 memeeege mwmuauoga m; OONIOOF m: oopuom a; omlop a: oplo Iwgmsm we Leesez memmepu em< meeeeeege we mewgemepeu .Ruowmwdz .93: ee eceN .mgeeeeege we .Cemeeee seem 55.5 msmsgew we :ewueewwumwe mm<--.mw.m e32. 71 .mewpmweeueeeece eeueewem new; me>wz eee meme; AFwsew we emepeeegeee .mgeseew we eeegm Lee geeseze .>e>eem eweEem ”eugeem ANee Ava 4_NV Ame 4e_v eeeeeeeee _eEeew we m mm we Pp mp wees ge meme» Lee; ANee 4_ev Amee Amee Aeev eeeeeeeee _eeeew we e eN me me me eeeee eeew eeee eee4 4F.V ANNV AeNV Amev 4_ev N m, mN mm mm eeeeeewppe me>w3 4em4 4_mv ANNV 4__v Aeev eeweeeeee peseew we __ mp me up op egos Le meme» seem ANee Amev ANev Amev Aemv eeweeeeee _eseew we e Nm mm, eN ee meeee Leew eeee eme4 AeV ANV 4e_v AmNe Amee e e em Fm mm eeeeeeepee eeeex seesee ee eeN-ee. ee eep-em ee em-e_ ee ep-e meeeeeeeeeeem mgeeeeege we mewgemeueu "we Fe>e4 —e:ewpee:em .wmuonmp , .wz .epez we eeeN we me>w3 eew mLeELew esp we mewumweepeegece Fecewpeeeee empeewemuu.om.m eweew 72 percentage of illiteracy was among the sharecropper heads of families-- about 43 percent. All of the 100-200 hectares landowners group were literate. The figures in Table 3.20 suggest that it is very common to have heads of families and their wives with fewer than four years of formal education. The 100-200 hectares landowners had the highest percentage of heads of families and their wives with more than four years of formal education. The use of formal education was tested in the survey. The percentages of their affirmative or negative responses to selected questions by the five groups of producers are presented in Table 3.21. The figures presented hithe table do not total 100 because of other alternatives available to the farmers in answering these items in the survey. A variable that may be related to management efficiency--the use of material from the official agricultural extension service-- received a negative response from about 90 percent of the sharecrop- pers. However, it was observed that the use of these materials increased with the size of the farm operation, and about 37 percent of the 100-200 hectares landowners made use of those materials. Close to 50 percent of the sharecroppers did not know how to add or multiply. Again, the knowledge of such operations was rela- tively more widespread among those farmers who owned more land. Among the other mathematical operations farmers were asked about, the compu- tation of percentages and interest was the most difficult. About 70 percent of the sharecroppers did not know how to compute percentages 73 .mweeeeewe we awemmpee use we emepeeewee eueeweew mewemww egwe .xe>w=m ewesem "eeweem PP om oe em pm ez mm mw mm mm mm me> memepeeewee epe_ee_ee emu FF mp mm mm on ez mm mm em me am me> umewewew eweweepee emu mo cw Pm me mm ez mm mm mm om oe me> awewuwes ceu o o— up mm me ez eew om mm mm wm me> eew emu mm Pm mm mm om ez mm mm «P we — me> Fewwmaes eew>wem cewmeeuxe Fewepweeweme _ewewwwe mew: e; oomuoop e; oopuom e; omlow e; opuo meeeeewemgegm mgeeeeewe we mewwemeueu eewpeeeeu we em: e.ww-ewew .ez .eeez ee eeeN .ewwEew age we eeee age we eeweeeeee weeeew we ewe--._N.m eweew 74 or interest. These figures were better among the landowners; only about 10 percent of the Too-200 hectares landowners did not know how to perform these calculations. The number of days farmers worked off of the farm was computed in two parts. The first was computed as the number of days the farm- ers worked in the agricultural sector; the second was the number of days they worked out of the agricultural sector. The number of days the farmers worked off of their own farms but in the agricultural sector was statistically different at the 5 percent level among groups of farmers. Sharecroppers, as expected, worked off their land more than did the other groups of producers. The average number of days other groups worked off of their farms decreased with the size of farm (see Table 3.18). The second measure--the number of days the producers worked off the farm and outside the agricultural sector--presented a differ- ent result. The arithmetic means of this variable were not statis- tically different at the 5 percent level among the categories of producers. This may be an indication of widespread opportunities for off-farm jobs. The commercialization index, which was computed by taking the percentage of the total production that was marketed is presented in Table 3.18. Using ANOVA, the arithmetic means of this index were found to be significantly different at the 5 percent level among the classes of producers. Those who had the highest commercialization index were the farmers who owned the most land. It is important to 75 notice that the sharecroppers had a higher commercialization index than did the 0-10 hectares landowners.‘ Family composition. The composition of the family living on the farm has an effect on farm composition, family labor supply, and off—farm income (see Figure 3.1). The average family in the Zona da Mata was composed of 5.71 persons (including other people living with the family). The share- croppers had the most sons and daughters, followed by farmers owning SO-lOO hectares and those with 100-200 hectares (see Table 3.22). These results were expected because a larger percentage of share- croppers are in the lower age brackets; consequently, more sons and daughters should be living with them. For larger farmers, more people could be living at home because more on-farm work opportunities are available. The categories of producers that had more nonfamily members living at home were the 50-100 hectares landowners and the 10-50 hectares landowners, in that order. Overall, the 50-100 hectares landowners had the largest families, with 6.31 persons. The sharecroppers had the next largest families, with an average of 6.11 people living together. The smallest families were found among the 0-10 hectares landowners, who averaged about 4.93 persons per family. The availability of family labor can be inferred from the figures presented in Table 3.22, assuming that family members could 1This result has also been reported by Garcia, 1975, op. cit. 76 .»e>w:m ewesem ”eeweem wweeeeewe we mewwemeueo me.m wm.e we.m mm.e ww.e Stew wee mcemwee we L.eeesz . . . . . ewwsew eew: eew>ww eweeee mm me Nm 3. _.N Lap—Ho we Lmnszc _.muo._. me. eN. ew. ww. NP. newesew PN. mw. wN. we. me. mewez newwsew eew eew; mew>wp epeeee emcee . . . . . Nwwsew eew: eew>ww eeeeeeeee Ne._ ew._ ee.w we.w ew._ eeeeeeeee we geese: weeew e e we. e e meme» ee we>e PP. ow. co. me. we. meme» me-wm mm. mm. mm. me. me. meme» emuew Ne. mm. em. em. Nm.w meme» mw wees: "weepsmeee um._ o_.~ em.w em.w e_.N meem we geese: Feuew o e we. we. 0 meme» we ge>o ew. mw. me. me. we. meme» me-wm ew. em. mw. ee. em. meme» em-ew we. m_._ Fw.w em. me.w meme» mp geeee ”mcom e; eeN-eew e; eew-em e; em-e_ e; ew-e wweeeeweeweem eewuwmeeEeu apwsew .wwumwmw .oz .epez we eceN ween; esp we cewewmeesee a_wsew mmewm>w:m eweEem "eeweem AeV em Ne. ew.meN.e we. we.ewm ee.mwe.e eeeeeeee 4wwz Awmv eew we. em.ee_.mm me. mN.wee._ ee._eN.ee 4wwz mwev eNN we. em.ewe me. mm.eee.w em.wee.w eeeweee va eew we. me.eee.N ee. we.e_e.N mw.wew.m eewzm Ava em. ew. ew.mwe.e wN. ee.eem.N ew._eN.Nw eweeee Ame eN ee. ee.Nme.e ee. me.mmm Nm.mee.e eeweeeeeee Ame Ne em. wm.ewe.w ee. wm.eeN._ ee.emw.N eeweew 4mmv New me. Ne._w we. ew.e e_.ww eewwee Amee ewe em. ew.NN we. me.em me.wm eeee Aeev Nee we. em.m em. mw.m ew.e eeeee weev men ee. Nw.eN em. ee.mw ew.ee eewe :ewueewewe mweeeeewe mewewem cewpeeeege e we .ez epmmwmnuwz epeepexwez eewue53meeu :ewpe23mceu :ewueeeewe empeeeewe meewpeweeo .wwuewmw .ez .euez we eeeN .mweEem ewwuee we» sew mpeeeewe empeewem we mewegem epeeuexwee eew .mewem .eewueaemeee .eewuezeeweuu.mm.m eweew 80 A high percentage of the production of other products such as poultry, swine, and fruits was consumed on the farm. The percentagescfl’farmers who raised these products were about 45 percent, 25 percent, and 9 percent, respectively. Analysis of the marketable surplus for cattle indicated a high percentage of production that was or could be marketed. Nearly 80 percent of the cattle constituted marketable surplus; however, only 25 percent of the farmers produced beef cattle. Milk production also had the characteristics of a commercial enterprise. About 30 percent of the entire sample produced milk, and the percentage of marketable surplus from total production was about 97 percent. Over 90 percent of coffee and milk production was market- able surplus; the percentage of the sample involved in these activi- ties was about 35 percent and 31 percent, respectively. Even though only about 5 percent of the entire sample grew vegetables, this activity had commercial characteristics. Only about 4 percent of total vegetable production was used for farm consumption. Table 3.24 shows the production, consumption, and marketable surplus of 11 selected products grown by the sharecroppers. The products raised by the most producers were corn, beans, and rice, which were grown by about 90 percent, 80 percent, and 60 percent of the sharecroppers, respectively. The percentage of these products consumed on the farm was about 90 percent of the bean production, 50 percent of the corn production, and 30 percent of the rice pro- duction. These producers' consumption of beans is above the average of 81 -ewe e43 wwesgew we memepceewee ewe memegpc .eeueewecw eeeeewe use eeeee egee cw memesee eew .memee amp u erm ewesem e .mewwe~=we hump cw eewemeee ewe: meeeeeee weepe ecu eee .mmee agnom cw eewemeee we: eewwee .mmee mxnom cw emeemeee ewe: cwee eee meeee .mmee axiom cw eewemeee we: eewme .ae>w:m ewesem "moweem ANV N me. ee.eee.N me. ee.me. ee.me_.N eeeeeeee 4wwz ANV m ee. ee.mme.ew ee. ee.emw ee.eew.ew 4wwz Meme we ee. ee.eee em. ee.mNe ee.mee.w weeweee Ne. eN em. em.wwe ew. ee.eeN.N ee.wwN.m eewzm ANV N ee._ ee.eem.e -- -- ee.eem.e eweeee Mme e we. ee.eNe.w_ Ne. ee.eww ee.eee._w eeweeeeee> NV m wN. em.mee ew. ee.eew.m em.mew.e eeweew AFNV wN Ne. ee.wm ee. NN.m NN.ee eewwee wee mww Ne. Ne.mN we. ew.NN me.we eeee wew Few mw. me. we. em.e wm.w eeeem Ame we Nw. Nw.NN eN. me.ew ee.em eewe cewueeeewe wweeeeewe . mewewem eewueeeeee e we .ez ewmwwmuuwz eweepexwez eewuesemceu :ewpesemeeu eewpeeeeee emueeeewe meewpeweeo Lew mpeeeege eeueewem we mewewem eweeu .nmimmmp aw: .mumz 2U GCON emeQQOLUwLwcm execs eee .mewem .cewpesemeee .eewueeeeweuu.em.m eweew 82 the Zona da Mata region, and the consumption of corn and rice is below the average (see Table 3.23). The percentages of farm consumption of the total production of fruits (79 percent), swine (70 percent), and poultry (56 percent) were relatively high; however, the percentage of sharecroppers who grew these products was not high. Only about 2 percent of the share- croppers grew fruits, and about 36 percent and 2 percent, respectively, grew poultry and swine. Only 2 percent of sharecroppers produced cattle, milk, and milk products, and these products were produced mainly for sale. Coffee was also produced primarily for commercial purposes; only 10 percent of its production was consumed at home. About 20 percent of the sharecroppers grew coffee. The figures presented in Table 3.25 show that corn, rice, and beans were the crops grown by the highest percentage of farmers in the O-lO hectares landowners group. More than 50 percent of the farmers in this category grew rice and beans, and about 75 percent of them grew corn. About 50 percent of the rice production was consumed on the farm, as were 75 percent of the bean and the corn production. Other products with a high percentage of production allocated to home consumption were swine and poultry. Fruits, vegetables, milk, and milk products were consumed on the farm at a rate close to 10 per- cent of production; only about 10 percent of the O-lO hectares group raised those products. Such products as coffee, fruits, vegetables, milk, and milk products, despite being grown by the smallest percentage 83 -eee en: mLeELew we memepeeewee ewe memespcewee cw wweee=e eew .memee mmp u erm eweEem .eeueeweew peeeewe esp eeeee e .mewwe~:we swap cw eewzmees ewe: mueeeewe wegpe es» ece .mmee axiom cw eegemeee we: eewwee .mmee exuoe cw eeeemeee ewe: ewee eee meeee .mmee mxuom cw eewemees we: eewme .ae>;=m eFeEem "eeweem ARV m mm. ww.mmm.m No. mm.mpm mo.mmm.~ muueeewe xpwz Appv mp om. Po.mmm.w op. —m.mom Nw.mmm.m xpwz Awev Pm we. mm.mw¢ mm. eo.¢ow mm.omp.p Asppeee Amwv mm mm. no.mmw.p Po. Fm.mow.m me.mm¢.¢ ecwzm Awpv mp oo.— mm.mmm.w 1. u: wm.mmw.w eppueu on w mm. mm.mmw.w mo. oo.-m m~.ewe.m me—eepeme> Amv ow mm. om.ame.m NP. om.nw¢ oo.wom.m muwewu ANNV R 8. 8.8 2. m: :eN eewwee $3 8 mN. eee R. e3: Ne.eN See Mmmw em mm. mm. mm. Nw.m mw.m mceem em em me. oe.m Pm. we.m wo.mw eewm eewueeeewe eweeeeewe mewewem eewpeeeege D $0 .02 w~mwwmunwz wFQMwang :OwHQESmCOU :OmHQEDmCOU :OmHUSUOLQ mmH03UOLQ meewueeeeo .mmuonmp .wz .epez we eceN .mgeczeecew e: o—uo Lew mpeeeewe eepeewem we mewewem epeepexwee ece .me—em .cewue53mcee .eewpeeeeweuu.mm.m eweew 84 of farmers in the 0-10 hectares group, had characteristics of commer- cial activities. The production, consumption, and marketable surplus of the farmers who owned 10-50 hectares are presented in Table 3.26. Corn, beans, and rice were the most commonly grown crops among the farmers in this group. About 80 percent of the producers grew corn, and about 70 percent grew beans and rice. The percentage of rice con- sumed on the farm was about 40 percent; about 60 percent and 70 per- cent of bean and corn production, respectively, were consumed on the farm. The activities that required relatively more land and capital, such as coffee and milk production, were more common in this category of producers than for sharecroppers and producers with 0-10 hectares. Besides poultry, corn, beans, and rice, all other products had commer- cial characteristics. However, the percentage of farmers who grew these products was not so high, as can be observed in Table 3.26. The figures presented in Table 3.27 led to the conclusion that corn, beans, and rice were the most popular crops among those farmers who owned SO-lOO hectares of land. About 85 percent of these farmers grew corn, and the farm consumption of this product was about 60 per- cent of production. About 80 percent of this category of farmers also grew rice and beans, and the farm consumption of each of these products accounted for approximately 30 percent of production. Other activities that were common among these farmers were the production of cattle, coffee, and milk. About 60 and 70 percent of these farmers produced cattle and milk, respectively, and about 50 percent grew coffee. The 85 -ewe eg: eweewew we memeaeeewee ewe memecueewee cw eweesee eew .eemee emu u ewwm eeeEem .eeueeweew peeeewe ecu eeeee e .mewwe~=we wwmw cw eewemeee ewe: meeeeewe wegpe en» eee .mmee mxlom cw eewemees me: eewwee .mmee m4-ee cw eewemeee ewe: ewee eee meeee .mmee exuom cw eewemeee we: eewme .ee>w:m ewesem "eeweem wewv we Ne. ee.eee.e we. ee.eee ee.NeN.e eeeeeewe wwwz weev we ee. we.ee_.NN ee. ee.eee e_.ee_.eN xwwz weee we em. ee.wew ee. ee.eeN._ we.eee.w eweweee weNe Ne ee. ee.eee.N em. ee.wee.N ee.wee.e eewzm weme ww ew. ee._ew.e wN. ee.eew.w ee.wee.e eweeee wee ew we. ee.ewe.e ee. ee.eNe ee.eee.e eeweeeeee> wewv we em. ee.eeN.w we. eN.eee.w ew.eNe.N eewewe weev me we. we._w ee. ee.e we.ew eewwee wNeV we. me. we.ew we. Ne.ee ee.ee ewee wwwv wew em. Ne.e we. eN.m we.e eeeee weev eew ee. ee.eN ee. ee.ew wN.em eewe eewueeeewe wweueeewe IIIIIMfiIII. meeewem :ewueeeewe e we .ez ewmmweuuwz eeeepexwez :ewueEemeeu eewuee=meeu eewpeeeewe emaeeeewe mcewueweeo .wmummmp .wz .epez we eceN .mweczeecep e; omnop wew mueeeewe eepee—em we mewewem eweeeexwee eee .me—em .eewuesemcee .cewpeeeeweuu.om.m eeeew 86 -ewe es: eweswew we memepceewee ewe memeepcewee cw ewee53e eew .eeueeweew peeeewe egg eeeee .memee mm u eewm ewesem e .mewwewewe wwme cw eewemees ewe: maeeeewe wezue es» eee .mmee mxuom cw eewemeee we: eewwee .mmee mxuoe cw eewemeee ewe: ewee eee mceee .mmee exuom cw eewemeee me: eewme .xe>w=m eweEem "eeweem wewv e we. ee.eee.e ew. ee.eNe ee.eee.e eeeeeewe xwwz weee we we. we.ewe.ee me. ee.wee.w ee.eee.we xwwz wwee eN eN. we.eee Nw. ww.eNe.w ee.eww.N wweweee wwev eN ee. ww.Nee.w ee. ee.wwe.e ee.eee.ew eewze weev em Ne. eN.e_e.ew ew. ee.eee.e eN.ewe.e_ eweeee Mme N ee. ee.ewe.w Ne. ee.eew ee.eee.w eeweeeeeee ewe e e_. ee.eeN ee. ee.eeN.N ee.eee.N eeweww wwee eN ee. ee.ee we. ee.e ee._e eewwee weev em ee. ee._e we. ee.ee .ee.we ewee weev we ee. ee.e em. eN.e ee.ew eeeee wewv me ew. ew.we wN. ee.NN Ne.ee eewe mweeeeewe cewueeeewe e . wepewem eewpeeeewe we ez eemmwmuuwz eweeuexwez eewue53meeu eewpeE=mceu eewpeeeewe empeeeewe meewueweeo .wwuwwmp .wz .epez ee eceN .mwe::ee:ep e; oopuom wew mueeeewe eeueeeem we mewewem eeeeuexwee eee .me—em .eewpe23meee .cewueeeewenu.wm.m eeeew 87 on-farm consumption of these products accounted for only 18 percent, 3 percent, and 7 percent, respectively, of total cattle, milk, and coffee production. Fruits and poultry were produced mainly for on-farm consump- tion. Vegetables were delivered directly to the market; however, only 3 percent of the farmers who owned SO-lOO hectares of land grew vege- tables. As expected, the farmers who owned more land, such as the farmers with 100-200 hectares, were more trade oriented. The percent- age of production that was marketable surplus of rice, beans, coffee, cattle, swine, milk, and milk products was more than 50 percent of the total farm production for each item. About 80 percent of the farmers in this group grew rice; about 10 percent were involved in fruit production, 70 percent in cattle production, 40 percent in swine production, 80 percent in milk production, and 16 percent in the pro- duction of milk products. Corn was grown by about 80 percent of the farmers in this group of producers; about 60 percent of the produc- tion was allocated to farm consumption. (See Table 3.28.) Much of the corn is fed to livestock. All vegetables produced were consumed on the farm; however, only 5 percent of the farmers in this group grew vegetables. Poultry production was also mainly allocated to farm consumption; about 40 percent of the farmers raised this product. Summary and conclusions. In light of the analysis presented in this section, one can conclude that production of corn, beans, and rice is the most common enterprise in the Zona da Mata. Corn was .eepeewecw ueeeewe es» eeeee -ewe es: wweewew we meaeueeewee ewe memegueewee cw ewee53: eew .memee me n erm eeeEeme .mewweeewe wwmw ew eewemeee ewe: mpeeeewe weepe ecu eee .mmee owlem cw eeweeeee me: eewwee .mmee exuee cw eewemeee ewe: ewee eee meeee .mmee axiom cw eewemees me: eewme .Ae>w:m eeeEem "eeweem 88 wewv e ee. ee.eew.e we. ee.eNe ee.eew.e eeeeeewe wwwz wewv me me. ee.eee.ee Ne. ee.eew._ ee.eee.ee eew: wwev w ee. eN._mm ew. ee.emN._ ww.eew.w eweweee Mwev w ew” Nw.eNe.e eN. e..wee._ ee.NNe.e eewze eev ew ee ee.ewe.ew Nw. ee.eee.N ee.eee.ew eweeee wee _ --. -- ee.w ee.eee ee.eee eeweeeeeee wwwv N ww ee.eNe eN. ee.eeN ee.ewe eeweww wNev e ee. ee.wee Ne. ee.w ee.eee eewwee Mewv ew me. ew._e we. ee.wew ee.eew ewee eew m. ew. ee.ww eN. ee.e ee.NN eeeee wee N_ ew. ee.eww eN. ew.ee ee.New eewe seepeeeewa mweeeeewe . meeewem eewueeeewe e we .ez ewmwwmuuwz eweeuewwez cewuesemeeu eewuesemeeu :eweeeeewe emueeeewe meewueweeo .wwuewme .ez .epez ee eeeN .mwee3eeee_ e; oomuoow wew meeeeewe eeeeeeem we meeewem eweepexwes eee .mewem .cewpesemeeo .eewpeeeeweuu.wm.m eeeew 89 grown by about 80 percent of the sampled farmers, and about 70 percent and 65 percent of those farmers produced beans and rice, respectively. As farm size increased it was observed that production of grains per farm also tended to increase, and also this was the tendency of these products' marketable surplus. By verifying the percentage of farmers who grew each of the analyzed enterprises, we can conclude that sharecroppers are more concerned in producing subsistence crops such as corn, beans, and rice. There is a tendency, however, in raising poultry, and to participate in coffee and milk-production contracts in which their labor force represents important input for landowners' production process. A relatively large percentage of O-lO hectares landowners, besides grain production, raise poultry, coffee, and swine. Among the landowner groups, this group of farmers had a smaller proportion of the production of grains as marketable surplus. Their production of coffee, cattle, milk, and milk products, which tend to be cash enterprises in the Zona da Mata, had more commercial characteristics than subsistence ones for these producers. The lO-50 hectares landowners tended to produce and commer- cialize coffee, cattle, milk, and milk products, besides grains. Production of milk, cattle, coffee, and grains seemed to be the most important enterprises for the 50-100 hectares landowners. Following these enterprises, swine production seems to be also of relevance for this group of farmers. Finally, for the group of lOO-ZOO hectares landowners, besides grains, emphasis seemed to be on coffee, milk, and cattle 90 production. These farmers had 76 percent of their rice and beans production as marketable surplus. Because relative emphasis is also put on the production of swine and poultry in addition to milk pro- duction, a relatively small percentage of corn was marketed. As expected, a small proportion of coffee, cattle, milk products, and milk were consumed on the farm. Swine and poultry production seemed to be subsistence activities for the smaller pro- ducers. The proportion of consumption of rice, corn, and beans decreased with the size of farms among landowners. The sharecroppers consumed the highest proportion of beans produced on their land and were the group of producers who had highest corn marketable surplus/ production ratio among all categories of farmers. As farm size increased, it was observed the inclusion of grain production among the many farm enterprises raised in the Zona da Mata. The economic importance of grains for each farm group will be analyzed in the next chapter. Market Prices Market prices are determined by supply and demand, and by government actions through official price policies. Prices used in this study are those reported by farmers surveyed. The price data obtained from the survey did not permit an analysis of prices farmers received throughout the survey year. It is assumed that reported prices correspond to the average of all sales farmers made throughout the year. 91 A statistical analysis of selected product prices is presented in Table 3.29. The importance of this analysis is associated with the development of the next section, which deals with farm income. A The ANOVA conducted on the arithmetic means of prices received by the five groups of farmers suggested that at the 5 percent level of significance, the prices of rice, beans, corn, coffee, cattle, poultry, and milk were not statistically different among the groups. The only product for which price was not statistically equal at the 5 percent level of significance among the many producer groups was swine. The explanation is that, except for swine, the producers of these products have similar storage facilities and sell their products in the market at similar periods of time and with similar quality. For products like milk, whose price is controlled by the government, possible dif- ferences reflect transportation costs. A study conducted by Paniago et al.1 showed that prices of hogs are higher from April through August. Assuming that sharecrop- pers and landowners with O-lO hectares and 10-50 hectares raise hogs for their own subsistence and sell only the surplus, they market their products at times of higher prices. In contrast with these producer classes, the 50-100 hectares and lOO-ZOO hectares landowners who raise swine for commercial purposes may not be able to avoid lower seasonal prices. An alternative explanation for the different prices received by groups of farmers is associated with the product marketed. Smaller 1Euter Paniago et al., Estudos sobre uma Regifio Agricola: Zona da Mata de Minas Gerais (II)(Rio de Janeiro: IPEA/INPES, 1973). 92 .weeeee .m.: _ weeee mewwewewe m.~e eweueewxewee< \mewweeswe wwmp :w eewemees me: eew:m we eewwe eew use we eewwe ecu .eeewee we eee; we m4 me ”wepwewmewweeewe wwme cw eewemeee me: xwws meeewee we eeecwmewweeewe wwmp cw eewemees ewe: :wueeee ece ewuuee we meewwe esp mmee axiomwmewweeewe wwme cw eewemees me: eewwee we eewwe ecu ewee eee meeee we meewwe esp ”nee ownemwmewweeewe wwme ”men mxuomwmewweeewe wwme cw eewemeee ewe: cw eewemeee me: eeww we eewwe egwe .ee>w:m eeesem "eeweem ww.N ee. ee.N ee.N ew.N ee.N wN.e eew: ee.eN ew. ee.eN ee.NN ee.eN ew.eN we._N eweweee ee.eNN mee. ee.Ne_ ee.N_N eN.eNN ee.eNN ee.weN eewzm ee.eew.w ew. ee.ewN.w ew.wew.w ee.ee_.w we.Nee.w ee.ewe eweeee ee.Nee ee. eN.wee ee.eee ee.eee ee.ewN eN.eee eewwee ee.ee mw. ee.ee ee.ew ee.ee ee.ew ew.we ewee ww.eee ew. ee.wee ee.eee ee.eee ee.Nwe ew.eee eeeee Ne.ee: we. ee.eww ee.wew wN.ee_ ee.ee_ ew.eew eewe we>e4 eeez . ee eeN-ee_ ee eew-ee ee ee-e_ ee ew-e eweeeeweeweem eeeeeewe Peweeee wwemwm euee e eeeze eeeez eweeEeewwe e w e e .ww- ewew .ez .eeez ee eeeN .mweswew e4 meeeeewe eeueewem wew ee>weeew meewwe eeewewwwe we eeeewwe> we memeeeew:m eweEem Neeweem mmp.wm Nm¢.No~ Nmm.eo mmm.wm oem.mp me.mw eEeucw xpwseu mmm.w omo.e~ woc.o mom.w ovm.m omo.w eEeecw Ewewuwwo mmm.m~ Necemmm mwm.wm www.mm oom.m wm~.m eEeucw Ewew pez emewe>< e; cowaoop e: oopnom e; omuop e; opuo mweeeewoewegm euez ee eceN mweewew we eeeww eeeeee we eeweem e.mewwe~:we wwmw cw .wwuowm_ .ez .euez ee eceN .mweEwew we eseeew awwEew ece .eEeecw Ewewnwwe .eseecw Ewew pezu-.om.m eweew 96 level, the arithmetic means of off-farm income differed statistically among groups of farmers. In Table 3.31, the off-farm activities of the Zona da Mata producers are summarized. Sharecroppers worked off of their contract land on other farms an average of 169 days per year. An average of 70 percent of that time they worked as daily hired labor, and an average of 3 percent of that time they worked as sharecroppers on other properties. The other groups of farmers worked less time off their own land and in the agricultural sector. Large farmers did not have any income generated in the agricultural sector besides that from their own farms. However, the income this group of farmers earned outside the agricultural sector was the highest one among all groups of farmers. The side occupations in which this group of farmers was engaged were basically service and trade oriented. The 10-50 hectares group of farmers also had trade as their principal occupation in addition to the farm business. Other incomes generated off the farm included rent on houses owned in town, rent on vehicles owned by the farmer, contributions from relatives, retirement compensation, and small-scale farm indus- tries. The aggregation of all these sources of income indicates that large farmers (100-200 hectares farm size) had the highest incomes from "other income" sources, followed by farmers in the 0-10 hectares farm-size group. The sharecroppers had the least income from nonagri- cultural sector and "other income“ sources. 97 weeee mewweNewe m.mw xpepeewxewee< .wewpee .m.: P .mewweeewe wwme cw eemmewexe ewe meseeew Ewew-wwoe .ee>wem eeeEem neeweem emo.em wee.e mom.w oee.m ome.w eeeeew Ewewuwwe wepew mwo.m mow.e mmm.e mmm.m mem eseeew wesuo weueem wem.me mem.e mmm.m mNN.P emu eeweeweewwme we pee eee ewew wwe eeueweeem eeeecH 0 Ne wN em m weueem weweeeeewwme esp we pee eexwe: whee we weasez o New.m wow.m emo.m Nmm.e weueem weweeweewwme cw pee ewew wwe eeueweeem eeeeeH e e N m m Awe weeeeweewezm m< o m m mm ow way mwmee eeww; epwee co wepeem weweeeeewwme egu cw 0 mm oe we owe eexwe: meee we weasez e; eomuoow e; oowuom e; omnoe e; oeuo mweeeewoewecm mweewew we eeewo Ewen wwo xwez e.wwuowm_ .ez .eue: ee eeeN .mweeeeewe we meeewm e>ww we meseeew Ewewuwwe emewe>e>w=m ePQEem "eeweem e.e e.o e.e me o_ e.e me me meeeeewe eeweeeeewwme weguo e.e eeeeeewe xwwz e 4ewz w wweweee w eew:m P eeeeeu w . meeeeueme> e epweww we eewwee mm :wee me meeem N eewe e; oomaoop e; oopuom e; omuoe e; opuo mweeeeweeweem mweewew we eeewo mueeeewe .wmnowmp .wz .euez ee eceN .eEeeew mmewm .mweewew ea mueeeewe eeeeeeem we eeee>ew esp we cewpeewweeee emepeeeweenu.mm.m eeeew 100 revenue contributed the highest percentage--about 45 percent--to farmers' gross income. The share of corn sales to farmers' gross income decreased with an increase in the size of the farms, constituting about 3 per- cent of the gross income of the 100-200 hectares farmers. However, much is fed to livestock on the farm. Vegetables and poultry had relative importance for the share- croppers and 0-10 hectares landowners; the relative shares of these products in gross farm income decreased with an increase in farm size. Milk products, an important potential source of income for the region through small-scale industry operation, still had minor importance for the Zona da Mata farmers. Fruit revenue, as with revenue from milk products, was of relatively minor importance to gross farm income. Sumnary and conclusions. In this section family income and its components were presented, i.e., the part of that income generated on farm and off-farm. The net farm income and the family incomes of sharecroppers and 0-10 hectares landowners were below the average of the study area. The 10-50 hectares landowners' incomes, i.e., net farm income, off-farm income, and family income, were about the average of the study area. The family incomes of the 50-100 hectares and the 100-200 hectares groups, on the other hand, were about two times and seven times the average of the Zona da Mata family income. Off-farm income is of special importance for sharecroppers and landowners with 0-10 hectares. About 55 percent and 43 percent 101 of sharecroppers' and 0-10 hectares landowners' family incomes came from labor wages, donations from relatives, retirement compensation, etc. The computed Gini Ratio of the farmers' gross income was about .637, expressing relatively high income concentration in rural Zona da Mata. The analysis of the percentage contribution of the revenue of selected products to farmers' gross income confirmed the hypothe- sis developed in the previous sections about the relative importance of grain production for smaller producers (see Table 3.33). About two-thirds of all producers raise rice and beans, and about 80 percent grow corn. With the exception of sharecroppers, the average of rice and beans consumption as a percentage of production declines as farm size increases. This trend is also observed for corn; however, it is not as pronounced as for the former products. Beans, corn, and rice revenues contributed 40 percent, 24 per- cent, and 22 percent of sharecroppers', 0-10 hectares and 10-50 hec- tares landowners' gross income, respectively. The relative importance of grains for the gross income of 50-100 hectares and 100-200 hectares landowners was relatively small. Besides grains, other products of importance for the share- cr0ppers' gross income were coffee, vegetables, swine, poultry, and milk. Considering that production of coffee and milk involves special contracts because of the land and capital requirements and that few producers have the opportunity to engage in such contracts because they primarily involve family partnerships, the production of other 102 .>e>w:m eweEem "eeweem emewe> we me em ee em ee Be: ee eeem wm mw em no em me meweeee; semico— wm om em on wN ew mewepee; eowuom we we we ww oe we meweeee; omnow mw ow mw Nm em em meweuee; o_-e ee ee we ew eN me eweeeeweeweee Ewew :e eweeeeewe ewew ee wweeeeewe Ewew ee mweeeeewe eee=meee :e eee:meee we eeEemeee we mewwemeeeu emeeeeewee emeeeeewee emepeeewee emeueeewee emeeeeewee emeeeeewee .Ewew eweu meeem eewm .wwuewew .ez .eeez ee eeeN .mewwemeeee eewm Ewew we Ewew ee eew: eewueeeewe esp we emepeeewee eee eewewm mew:ewm mweEwew we emeuceewee--.mm.m eweew 103 enterprises should be regarded as potential income generators for the sharecroppers. The revenue from grains, coffee, cattle, and milk accounted for 73 percent of the 0-10 hectares landowners. It is worth noting that production of vegetables, swine, and poultry was also of rela- tive importance for these producers' gross income. The characteristics of the 10-50 hectares landowners are similar to those of the 0-10 hectares landowners. About 89 percent of the 10-50 hectares group was composed by grain, coffee, milk, and cattle revenues. About 92 percent of the 50-100 hectares group's gross income was constituted by grain, coffee, cattle, and milk sales. Milk revenue alone accounted for 45 percent of this group's gross revenue. Swine production also was of importance for this class of producers' gross revenue. The 100-200 hectares landowners' gross income was most depend- ent on coffee, milk, and cattle production. Coffee revenue alone accounted for 61 percent of the total gross income of this group of farmers. By examining the farm family income composition as well as the percentage of the farmers who raised the products described in this chapter and the share of those products' revenue on farmers' gross income, an attempt was made to identify typical farmers of each class of producers. These typical farmers are as follows: Sharecroppers: For this class of producers, production of grains (corn, beans, and rice) constitutes the most typical farm 104 enterprises. However, off-farm income constitutes the most important source of income for these producers--about 55 percent of their income was generated off the farm. Adjacent enterprises of importance for these producers are poultry and coffee production. 0-10 hectares landowners: The most typical farm of this class of producers is characterized by production of grains. Off-farm income is also of major importance for these producers--about 43 per- cent of their income was generated off the farm. Poultry, coffee, and swine production are adjacent enterprises of this group of pro- ducers. 10-50 hectares landowners: Production of grains, milk, poultry, coffee, cattle, and swine are the most common enterprises in a typical farm of this group of producers. Off-farm income is of relatively less importance for these producers than it is for share- croppers and 0-10 hectares landowners. 50-100 hectares landowners: The typical farm of this group of producers raises grains, milk cows, coffee, swine, and poultry. 100-200 hectares landowners: Production of milk and corn are the most common enterprises, followed by cattle, beans, rice, poultry, swine, and coffee. Similar enterprises were observed in both 50-100 hectares and 100-200 hectares groups of farmers. However, for the former group of farmers the highest share of their gross income was from milk sales and, for the latter, coffee generated the highest income share. CHAPTER IV ECONOMIC EFFICIENCY OF THE GRAIN SUBSECTOR In the preceding chapter, the process of income determination in the Zona da Mata farm sector was discussed. The importance of the grain subsector for the various classes of farmers was evident. This chapter deals with the production side of the grain subsector. The analysis that follows was motivated by the need for more knowledge about the economic efficiency of resource use in the grain-production process in the study area. The results of this analysis can be a valuable component of future development efforts for the area, par- ticularly in reallocating resources in the grain subsector with the objective of increasing farmers' income. The Grain Cropping System in the Zona da Mata It has been estimated that 75 percent of the total area used to produce rice in Minas Gerais State is upland, and the rest is lowland without controlled irrigation.1 Among the biological require- ments of the rice crop, soil moisture is the most limiting factor. It has been estimated that under controlled irrigation it is possible 1Empresa da Pesquisa Agropecuaria de Minas Gerais, "A Cultura de Arroz em Minas Gerais," Informe Agropecuario (Belo Horizonte) 5 (Julho 1979): 9. 105 106 to increase rice yields from four to five times relative to yields under the rainfall system.1 Rice production in the Zona da Mata is developed principally in lowland areas with no water control during the growing season. A traditional production system prevails; i.e., improved seeds, fer- tilizers, pesticides, and other modern inputs are not generally used. About 64 percent of the Zona da Mata farmers produced rice in the l976-77 agricultural year.2 The production of beans and corn is also widespread in the Zona da Mata. About 70 percent and 80 percent of the producers in that region grew beans and corn, respectively, in the l976-77 agri- 3 Vieira estimated that 90 percent of the bean pro- cultural period. duction in the Zona da Mata is intercropped with corn. This intercropping system is used primarily by small farmers, and hand- cultivation systems are the rule. Referring to the corn-beans intercropping system, Vieira pointed out that "it reduces the inci- dence of pests, utilizes family labor more intensively (which is relatively abundant in the region), reduces risk, and guarantees diversity of diet and income sources. On the other hand, this produc- tion system impedes the utilization of agricultural practices that 1Orlando Peixoto Morais, Fernando Linho, and Plinio César Soares, "Exigencias Climaticas da Cultura do Arroz," Informe Agro- pecuario (Belo Horizonte) 5 (Julhy 1979): 16-19. 2See Table 3.33. 31bid. 107 lead to higher yields."1 Research efforts have been concentrated on the corn-beans combination since researchers and extension workers have failed in their attempts to encourage farmers to plant improved crop varieties in sole stands. In both cropping systems, corn and beans planted separately and intercropped, the use of relatively flat land has been observed. As the topography becomes more hilly, the intercropped system is recommended because it protects the soil from erosion. In summary, the grain subsector of the Zona da Mata, which is the focus of this chapter, is composed of rice, beans, and corn sole cropped, as well as corn and beans intercropped. The production of all these crops largely uses traditional technologies. Specification of Grain-Production Functions and Some Theoretical Considerations Assuming neoclassical pure competition, a firm will select its levels of production and input use so as to maximize its profit function, where the total revenue (TR) minus total cost (TC) equals profit. Total revenue equals product price (PY) multiplied by out- put (Y). Output (Y) is a function of the input combinations 0(Xi, i=1...n). Total cost equals the sum over all the inputs of unit 1Clibas Vieira, "Cultivo Consorciado de Milho con Feijao," Informe Agropecuario (Belo Horizonte) 4 (Out. 1978): p. 42. For simi- lar evidence of advantages of the intercropping systems in other tra- ditional agriculture, see 0. W. Norman, D. Pryor, and C. J. N. Gibbs, "Technical Change and the Small Farmer, Hausaland, Northern Nigeria," African Rural Economy Paper No. 21 (East Lansing: Department of Agri- cultural Economics, Michigan State University, 1979), p. 59. 108 price (Pxi) multiplied by the quantity used (Xi). THus, profit can be defined as: "M: Profit = PY-0(X],...,Xn) - PX. - x. (4.1) i 1 ' ' where 0(X1,...,Xn) is the production function of Y. Assuming the law of diminishing returns, profit is maximized when the first-order conditions are met: 3Profit _ _ T - PY DhXi - PX]. — 0 (4.2) where plzxi is the first derivative of the production function with respect to input Xi (i=l,...,n), and FY ”szi is the marginal value product (MVPxizY) of Xi in production of Y. Thus, under the assump- tion of pure competition, profit is maximized when the level of input use is such that MVPX1:Y equals the factor cost. Production Function4§pecification To verify whether grain producers of the Zona da Mata combine inputs in order to maximize profits, the marginal value product of Y for X1 must be determined. This can be done by estimating the pro- duction functions of the grains under study. The production function as stated before is as follows: Yj = 0j(X]j,...,an) (4.3) where Yj is rice (j=l), bean (j=2), corn (j=3), and corn-beans combination (j=4), and Xi (i=l,...,n) are the inputs. The produc- tion function above constitutes what Aigner and Chu called an "average 109 production function."] According to them, it would be correct to use the concept of "average" when one wishes to estimate how much output, "on the average," could be obtained for a firm in the industry with a certain set of inputs. Other important uses of the average function are that (a) in some cases one can approximate the industry's aggre- gate production function when aggregate data cannot be obtained but data at the firm level are available and (b) one can approximate an "average" firm production function when he has data only on industry aggregates. The explanatory variables included in estimating equation (4.3) are as follows: Y. = production of grain j, measured in 50 kg bags of rice, J 60 kg bags of beans, 60 kg bags of corn, and the value of corn-beans combination, measured in 1977 cruzeiros; xij = quantity of land used to produce commodity 3, measured in hectares; ij = quantity of labor used to produce commodity j, measured in man-days; X3j = quantity of seeds used to produce commodity j, measured in kg of seeds; >< 1| 4j value of pesticides used in production of commodity j, measured in 1977 cruzeiros; >< ll 53 value of fertilizers used to produce commodity j, measured in 1977 cruzeiros; 10. J. Aigner and S. F. Chu, "On Estimating the Industry Pro- duction Function," American Economic Review 58 (1968): 826-39. 110 X6j = value of machinery services used in production of commodity j, measured in 1977 cruzeiros; X7j = value of bullock plowing services used in production of commodity j, measured in 1977 cruzeiros; X . = proxy for management, which is the average of years of formal education of the head of the family and his wife, measured in years of education; & DZj = dummy variables. These variables were included in the function to capture environmental differences among sub- regions of the Zona da Mata. Three subregions were con- sidered: Juiz de Fora, Muriaé, and Vicosa. For D]j=l, Juiz de Fora is identified, and for 02j=l, Muriaé is identified. A possibility of management and environmental bias may occur in estimated production functions when proxies for management ability and environmental differences are not included in the equations. As discussed in Chapter III, the education of the farmer may be a vari- able associated with management efficiency and is included in the present analysis. This procedure was used successfully by Yotopoulos in a study of efficiency of resource use in subsistence agriculture.1 Many other techniques for dealing with the problem of management bias have been presented in related literature.2 1Pan A. Yotopoulos, "On the Efficiency of Resource Utilization in Subsistence Agriculture," Food Research Institute Studies, Stanford University 7(2) (1968): 125-35. 2See, for example, B. F. Massell, "Elimination of Management Bias From Production Functions Fitted to Cross-Section Data: A Model 111 In considering efficiency, it is worth noting that different researchers have used their own definitions. Hall and Ninsten used various concepts or types of efficiency to reflect judgments on dif- ferent aspects of farming. Their primary concern was, however, mana- gerial efficiency. They stressed that especially when making judgments about the relative performance of managers, allowance must be made for the nature of the physical environment facing each manager. If differ- ent managers face different constraints on their maximizing behavior, judgments about their relative performance will be useless unless these constraints are understood.1 Farrell, on the other hand, used the term "technical efficiency," which he defined as judging a firm's success "in producing maximum output from a given set of inputs." He also defined "price efficiency" as meaning to judge the firm's "success in choosing an optimal set of inputs."2 Marshak and Andrews seemed to use the same concepts and terms as Farrell employed but called them . . . . . 3 "technical" and "economic" eff1c1enc1es. and an Application to African Agriculture," Econometrica 35 (July- October 1967): 495-508; Irving Hoch, "Estimation of Production Func- tions Parameters Combining Time-Series and Cross-Section Data," Econo- metrica 30 (January 1962): 34-53; Yair Mundlak, "Empirical Production Function Free of Management Bias," Journal of Farm Economics 43 (February 1961): 44-56. 1M. Hall and C. Ninsten, "The Ambiguous Notion of Efficiency," Economic Journal 14 (March 1959): 71-86. 2M. J. Farrell, "The Measurement of a Productive Efficiency," Journal of the Royal Statistical Society, Series A, General, Part 3 120 (1957): 252-81. 3J. Marshak and N. Andrews, "Random Simultaneous Equations and the Theory of Production," Econometrica 12 (July 1944): 143-205. For a detailed revision on the definition and measurement of technical efficiency, see C. Peter Timmer, "0n Measuring Technical Efficiency," 112 Production-Function Estimation Procedure Generally, in production-function studies in agricultural economics, much of the concern is about economic efficiency. Some statistical difficulties are present in estimating the production functions which could result in misleading conclusions about effi- ciency. These difficulties include identification problems, specifi- cation bias, simultaneous equation bias, and multicollinearity. Equation (4.3) in a standard Cobb-Douglas production function is: 109 ij = E Bij 109 xljk + ujk (4.4) where ij = output j of farm k xijk 31.3- ujk = stochastic term amount of input i of farm k used to product output j elasticity of production of input i of output j To obtain consistent and efficient estimates of Bij’ it would be useful to achieve variations in the inputs from a stochastic pro- cess.1 However, under perfect competition, the decision of how much to produce as well as the choice of inputs is made according to the rules of profit maximization. Considering a specific product for any farmer, we have Food Research Institute Studies, Stanford University 9(2) (1970). See also Peter Schmidt and C. A. Knox Lovell, "Estimating Technical and Allocative Inefficiency Relative to Stochastic Production and Cost Frontiers," Journal of Econometrics 9 (1979): 343-66. 1Andrew B. Tench, Socioeconomic Factors Influencing Agricul- tural Output (Sozialfikonomische schrissten zur agrarentwicklunj Heft 12 Saarercken, 1975). 113 M s.v ——-— = A = R. (4.5) k axik xik 1k where Rik = the price of input i to firm k, divided by the price of output. The factor prices and product prices are the same for all farmers, and the elasticity of production of various inputs is known. Equation (4.5), written logarithmically, and with an error term added, becomes Xik = -1og Ri + yk + log Bi + wik (4.6) where xik = log xik’ yk = log Yk’ and wik is the error term. Then, equation (4.6) is unidentifiable.1 However, relaxation of some assumptions of the perfect- competition model allows for identifiability. For instance, if the assumption of certainty is relaxed, entrepreneurial ability will vary the inputs used in the production process. On the other hand, imper- fect factor markets and different elasticities of supply of inputs affect input levels. Differences in physical environment--the stock of capital, for instance--would also affect the use of inputs. The constancy of production elasticities and profit-maximization principle retained are aspects of the model. The former aspect is retained to make the function measurable, and the latter is retained to interpret the results against established neoclassical economic theory. Specification bias occurs because the true functional form and the complete range of variables that it should contain are 1Hoch, op. cit. 114 unknown. Heady and Dillon showed how the ordinary least-squares method, the most frequently used method in estimating production functions, may bring bias to the estimations when: (a) an incorrect functional form is used, (b) some variables are omitted from the model specification, and (c) when aggregation within and aggregation over inputs occur.1 Simultaneous equation bias results from the situation in which the disturbance term is not truly independent of the inputs. We can consider that the production function is one equation in a 1 system, and if it is functionally related to other equations, the single equation estimation will generate inconsistent estimations for the parameters.2 However, consider that a farmer makes a decision on planned output. The planned output in some sense determines the levels of inputs to be used, and it differs from the actual output by the farmers' degree of success. One way to avoid the problem of simultaneous equation bias is to assume that firms make input deci- sions on the basis of anticipated output rather than current output. 1Earl o. Heady and John L. Dillon, Agricultural Production Functions (Ames: Iowa State University Press, 1966). For other approaches dealing with specification bias, see 2. Griliches, "Specification Bias in Estimates of Production Functions," Journal of Farm Economics 39 (1957): 8-20. See also H. Theil, "The Analysis of Disturbances in Regression Analysis," Journal of the American Statistical Association 60 (Dec. 1965): 1067-79; J. B. Ramsey, "Tests for Specification Errors in Classical Linear Least-Squares Regression Analysis," Journal of Royal Statistical Society, Series B, 31,2 (1969): 350-71; James B. Ramsey and Peter Schmidt, "Some Further Results on the Use of OLS and BLUS Residuals in Specification Error Tests," Journal of the American Statistical Association 71 (June 1976): 389-90. ZG. S. Maddala, Econometrics (New York: McGraw-Hill Book Co., 1977), p. 220. 115 This assumption, according to Tench, seems reasonable in the agricul- tural sector, and the ordinary least-squares technique can be used to estimate the parameters.‘ The fourth problem of concern in this section related to the production-function estimation is multicollinearity, which exists if the columns of the matrix of the explanatory variable observations are linearly dependent.2 This could be the case, for instance, if there is a fixed method of cultivation with less chance of between- input substitution. In such a situation it is likely that as one input increases, another input also increases. The main consequence of multicollinarity is that the preci- sion of estimation falls so that it becomes very difficult, and some- times impossible, to separate the relative influences of the various explanatory variables. This loss of precision has three aspects: "Specific estimates may have very large errors; these errors may be 1Tench, op. cit. A classical article referring to the prob- lem of simultaneous equation bias was presented by A. Zellner, ' J. Kmenta, and J. Dréze, "Specification and Estimation of Cobb-Douglas Production Function Models," Econometrica 34 (Oct. 1966): 784-95; and I. Hoch, "Simultaneous Equation Bias in the Context of the Cobb-Doublas Production Function," Econometrica 26 (Oct. 1958): 566-78. Other related references are J. Johnston, Econometric Models (New York: McGraw-Hill Book Co., 1972); Y. Mundlack and I. Hoch, "Consequences of Alternative Specifications in Estimation of Cobb-Douglas Produc- tion Functions," Econometrica 33 (Oct. 1965): 814-28; and M. Nerlove, Estimation and Identification of Cobb-Douglas Production Functions (Chicago: Rand McNally, 1965). 2Peter Schmidt, Econometrics (New York and Basel: Marcel Dekker, Inc., 1976). 116 highly correlated, one with another; and the sampling variances of the coefficients will be very large."1 Selecting Subsamples for the Production-Function Analysis The first step in determining the subsamples for the grain- production-functions analysis was to identify the farmers who grow corn, beans, and rice, sole cropped, and corn and beans combined in the same stand. Because farmers who own more than 100 hectares are not potential members of PRODEMATA, l9 farmers in this category were eliminated from the 550 original cases in the sample. After veri- fication of all values of variables included in the production function stated in the preceding section, and eliminating those cases that pre- sented enumeration or key-punching problems,2 the total number of grain producers was estimated and is presented as "total number of producers" in Table 4.1. Under the inputs grouping specified in the production func- tions, a substantial number of zero observations occurred. Because of these observations, some procedure should be used to estimate the log-log production function specified. King and Byerlee had a similar problem, and among the possible solutions they cited was to drop households with zero observations in a particular variable from the 1 2Some of the eliminated cases had zero observations in inputs such as land or labor. Johnston, op. cit., p. 160. 117 analysis.1 Alternatively, they preferred to replace zero observa- tions with some arbitrary small number;2 however, they recognized that parameter estimates based on the log-log model must be inter- preted with caution. Table 4.1.--Total number of grain producers of the Zona da Mata and number of cases selected for grain-production-function analysis. Number of Cases Selecteda Total Number Enterprises of Producers Small Large Total Producers Producers Rice 334 138 188 326 Beans 48 20 25 45 Corn 92 48 42 9O Corn-beans 274 112 148 260 Source: Sample survey. aSmall producers = sharecroppers and O-lO hectares landowners; large producers = 10-50 hectares and 50-100 hectares landowners. 1Robert P. King and Derek Byerlee, "Income Distribution, Consumption Patterns, and Consumption Linkages in Rural Sierra Leone," African Rural Economy Paper No. 16 (East Lansing: Department of Agri- cultural Economics, Michigan State University, 1979). 2This procedure was also adopted by Massell, "Elimination of Management Bias From Production Functions," and Benton F. Massell and R. N. M. Johnson, "Economics of Smallholder Farming in Rhodesia. A Cross-Section Analysis of Two Areas," Food Research Institute Studies in Agricultural Economics, Trade, and Development, Supplement to Vol. VIII, 1968. 118 For the purpose of the present research, it was decided to aggregate some inputs in order to eliminate all zero observations. Two classes of modern inputs (MI) were considered: MI], which included value of fertilizers and pesticides, and M12, which included machinery services and value of bullock plowing services. For some products in the analysis, there were still zero observations, which would imply small degrees of freedom for further estimations. To eliminate this problem, M11 and M12 were aggregated with the value of seeds used in the production process of commodity j. The subsample sizes are shown in Table 4.1. Because of the aggregation of inputs, the production function explicative variables of equation (4.3) were . = quantity of land used to produce commodity j, meas- 1J ured in hectares; ij = quantity of labor used to produce commodity 3, reduced to: X measured in man-days; MIj = value of “modern inputs"--seed, machinery and bullock plowing services, and pesticides and fertilizers; XBj = average years of education of the head of the family and his wife; and dummy variables Dlj and DZj’ introduced in the function to capture the environmental differences among the three subregions of the Zona da Mata: Juiz de Fora, Muriaé, and Vicosa. Estimations of the Grain-Production Functions In the process of estimating average production functions of grains for the Zona da Mata, an attempt was made to verify possible differences between small and large farms' production functions. This was done because of the evidence of different resource endowment 119 and resource use by various classes of producers, presented in Chapter III. Tables 4.2 and 4.3 summarize the production functions of beans, corn, rice, and corn-beans intercropped for small and large farmers, respectively. The ordinary least-squares method was used for the estimations; with the exception of education and the two dummy variables, all variables were taken in logarithms. (This func- tional form provided expected coefficient signs of all variable inputs as well as desirable levels of significance for the coefficients. For the small farmers' grain-production functions, the corrected coefficient of determination (R2) ranged from .4882 to .7080. In all equations land and labor were statistically significant at levels ranging from 1 percent to 10 percent. (See Table 4.2.) The coefficient of the modern input variable in the bean equation was not statistically significant at the 20 percent level and presented an unexpected negative sign. However, it was significant and positive in the other equations. Education was statistically significant at the 1 percent level in the rice equation but was not significant in the other equations. Ths estimated coefficient of the dummy variable 01 in the rice equation was statistically significant at the 1 percent level, indicating a distinct intercept for the Juiz de Fora subregion among other subregions of the Zona da Mata. For the larger farmers' grain-production functions, the cor- rected coefficient of determination (R2) ranged from .4983 to .7587. The estimated coefficients of the variables land, labor, and modern inputs were statistically significant at the 20 percent level or less. 120 .we>ew peeewee or we eweN Eeww ucewewwwe ewuceewwwemwm peewewwweeoeew .we>ew peeuwee m we eweN Eeww acewewwwe xwueeewwwcmwm weewewwweeuee ._e>e_ eeeewee _ we eweN seww ucewewwwe ewpeeewwwemwm ecewewwweeuw .mpcewewwweee ecu we wewwe eweeeeum ewe memecpeewee cw mewemww eew .mwee:eeeew meweuee; owie ece eweeeeweewezm u wweewew wweEme weee.Ne eewe.ee Ne__.ww eeee.e wee eew we. _e e_ we Neee. Neee. eeew. wewe. Ne wNwe_.v NNwe.- wNNew.V eewe.- weeew.v eeNe.- wwNee. e eeeN.- Ne we_e_.v wwee.- wewww.e wweeNee.- weewN.v wNNN.- weeee. e ewee.- Pe wewee.v eeNe. weeee.v weeew. weeee.v eeew. wemeN. V wwwe. eeweeeeee wwwee.e eweee. weeee.e «wNNeN. wNeee.e weeew. wweNN. V eeee.- eeeeew eweeez wweew.e wNeeN. wweew.v ewNee. weeww.e wwwNe. wNeee. e weeeNee. weee4 weNee.e weeee. wewee.v eeeee. weNN_.v «eweeN. weeee. e eweeee. eee4 wewee.e ewee.e weeNe.V eeww.- weNee.v wwee.- wNeee.wv meme. ewee eeeeeeee meeemieweu eewm cweu meeem meeeew e.wwlowmw .ez .epez ee eceN esp we mweewew wweEm wew mcwewm we mcewueeew eewueeeeweiu.m.e eweew .we>e_ eeeewee em ee eweN eeww eeewewwwe eweeeewwwemwm aeewewwweeueewe ._e>e_ peeewee ow we eweN eeww eeewewwwe ewueeewwwemwm eeewewwweeueee .we>e_ eceewee m we eweN eeww eeewewwwe eweeeewwwemwm peewewwweeeee .we>e— aceewee e we eweN aeww peewewwwe ewpeeewwwemwm weewewwweeoe .mecewewwweee esp we wewwe eweecepm ewe memeeueewee cw mewemww eew .mwee:eeeew meweuee; ooeaom ece meweeee; omiow u eweewew emwe4e ewew.ee erN.ew ewwe.ew eeee.e wee wee eew me e_ we _.2. Neee. weew. eeee. weew. we .. wewe_.v wweeeN. weNew.e eweeNew. weeeN.e eewwewe.- weeeN.v eewe.- Ne wwee_.v eeee.- wewww.v meee.- wweee.e wweewe.- wweee.v eweNNew.- we wewNe.e eeNe. weNNe.v ewwee. wwwee.e eNee. weeee.e weeeww. eeweeoeee weeee.v eeweN. wweee.v wwweN. wwewe.e eeNeN. weer.v wwweeewN. eeeeew eweeez weeee.v eweee. weeee.e weewe. wweew.v weere. weeew.e wNNee. weee4 weewe.v weeee. weewe.v ewewe. wNee_.e weweeeN. wwmeN.e eweeeNe. eee4 weeee.v eewN.e weewe.e ewee.- weeew.v eeew. weNee.v ewee.w- ewee eeeeeeee meeemucweu eewm :weu meeem mueeew e.ww-ewme .ez .eeez ee eeeN es» we wweewew emwee wew mewewm we meewueeew eewpeeeeweiu.m.e eweew 122 Education as proxy for management was statistically significant in the beans equation (at the 10 percent level) and in the rice equation (at the 1 percent level). Dummy variable 0] as intercept shifter was statistically significant in the beans and corn equations, and dummy variable 02 was statistically significant in the corn, rice, and corn-beans combination equations. The same procedure was used to estimate production functions including both subgroups of farmers. These pooled functions are pre- sented in Table 4.4. Tests conducted on the results of these equations failed to reject the hypothesis that both groups of observations, those for small and large farmers, belong to the same regression model.1 Under the assumption that the variance of the error term is proportional to the square of the independent variable, a test sug- gested by Park found no heteroscedasticity problems in any of the equations presented in Table 4.4.2 1For details about the tests used, see Gregory C. Chow, “Tests of Equality Between Sets of Coefficients in Two Linear Regressions," Econometrica 28 (July 1960): 591-605. The ratio [SSET - ($551 + SSE2)]/K ($55] + SSE2)/(T - 2K) is distributed as F(K, T-2K) under the null hypothesis that both sub- groups of farmers belong to the same regression model. The computed ratios for beans, corn, rice, and corn-beans combination were 1.06, 1.01, 1.08, and 1.09, respectively. For rejection of the null hypothe- sis at the 5 percent level of significance, F would have to be at least 2.33, 2.17, 2.01, and 2.01 for beans, corn, rice, and beans-corn combi- nation, respectively. 2R. E. Park, "Estimation With Heteoscedastic Error Terms," Econometrica 34 (Oct. 1966): 888. Heteroscedasticity tests proposed ' by Park are presented in Appendix C. i 123 ._e>ew peeewee om ._e>ee eeeewee ow ._e>ew eeeewee m .ee>ew uceewee _ memegpeewee cw mewemww eew .ewee:ee:ew we eweN seww weewewwwe eweeeewwwemwm eeewewwweeeeeew we eweN Eeww ueewewwwe eweceewwwemwm peewewwweeeeee we eweN eeww eeewewwwe ewueeewwwemwm eeewewwweeutw we eweN eeww peewewwwe ewueeewwwemwm acewewwweeee .mueewewwweee esp we wewwe eweeceum ewe meweeeeeleowieegeummew eee wweeeeweewecm eeeeeeHe Nwee.ew eeNe.eew eeew.eN ee_e.Nw wee eeN ewe me em we wNwe. eeee. ewee. weee. Ne wwNee.v wewweeew. wNeee.v wNew. weeew.v ewew.- weeeN.v eweN.- Ne weeee.e ewee.- weeww.v wwewNwN.- wweww.e wNeee.- wwmwN.e weeeeee.- we weewe.e eewe. wwwNe.e weeee. weeee.e wewwewwe. wwewe.e weweew. eeweeeeee wewee.e eeNeN. weeee.e weewN. wNeee.e weeeN. wee__.v eeww. eeeeew eweeez weewe.e eewee. wweee.e weeee. wewee.v wewee. weeew.e weeeee. weee4 wewee.v wNeem. wwwee.v wwwee. weeee.v eNeeN. weeeN.v «eeNe. eee4 wewwe.v eewe.e weeeN.v Neee.- wewee.e eeee. weeee.v eeew.- ewee eeeeeeee meeemiewee eewm ewee meeem menace e.ww-ewe_ .ez .eeez ee eeeN eew wew eewewe we eeeweoeew eeweoeeewe--.e.e eweew 124 The corrected multiple determination coefficients ranged from .6127 (for corn-beans combination) to .6940 (for rice); all four regressions were significant at the 1 percent level. Overall, land and labor are important variables in explaining interfarm output differences. Consistently, for the products analyzed, land appeared to be a limiting factor, as one would expect. Labor was statistically significant at the 5 percent level in the estimated equation for beans, and it was significant at the 1 percent level in the other equations. The coefficients of modern inputs were sta- tistically significant at the 1 percent level in the corn, rice, and corn-beans combination equations and was not significant in the beans production equation. Because of the large standard error of modern inputs in the bean equation relative to the estimated parameter, the results that follow concerning this variable must be interpreted with caution. Education (as proxy for management) was significant in the bean regression (at the 5 percent level of significance), in the corn regression (at the 20 percent level of significance), and in the rice regression (at the 1 percent level of significance). Dummy l, which stands for the Juiz de Fora subregion, was significant in the bean and rice equations at the 10 percent level, and in the corn equation at the 1 percent level. In all of these three regressions this variable had a negative sign, indicating that regional differences among farms did have an effect on production of beans, corn, and rice, which were lower in the Juiz de Fora subregion 125 than in the rest of the Zona da Mata.1 Turning to the corn-beans combination, it was observed that Dummy 2, which stands for the Muriaé subregion, was positive and significant at the 20 percent level. This result suggests that regional differences among farms did have an effect on corn-beans combination production.2 Elasticities of Production In the functional form used in this study, the regression . coefficients of land, labor, and modern inputs are equal to the pro- 3 duction elasticities and independent of factor ratios. Elasticities of production indicate the expected increase (or decrease) in produc- tion that would occur if the amount of the input resource was increased (or decreased) by 1 percent, other input levels being held constant. All production elasticities for the various types of enterprises are positive. The sum of the production elasticities is less than, equal to, or greater than unity for decreasing, constant, or increasing return to scale, respectively. To test for constant returns to scale, the null hypothesis was that the production elasticities sum to unity for each crop. A 1Note that for beans, corn, and rice regressions, variable DI (Dummy 2), which identifies the Muriaé subregion, was not significan at the 20 percent level. 2Note that D] (Dummy l), which stands for regional differences of the Juiz de Fora subregion, was not statistically significant at the 20 percent level in the corn-beans combination regression. 3Earl O. Heady, "Technical Considerations in Estimating Pro- duction Functions,” in Resource Productivity, Returns to Scale, and Farm Size, ed. Earl O. Heady, Glenn L. Johnson, and Lowell F. Hardin (Ames: Iowa State College Press, 1956). 126 two-tailed t-test was used, and none of the sums of the elasticities were significantly different from unity at the 5 percent level.1 This finding was consistent with those of other comparable studies.2 Marginal Productivities The marginal productivity of factor i in producing crop j is denoted by ”ij and is given by _ Y- ”ij Eij iTJL (4.7) ii where Eij = the elasticity of factor i in producing crop j; Yj = the output of crop j; and Xij = the amount of input i used to produce grain j. The estimated marginal productivities of land, labor, and modern inputs were calculated at the geometric means of these vari- ables and the outputs. The factor marginal productivities as well as the marginal value products of land, labor, and modern inputs are presented in Table 4.5. The marginal value product of each input indicates the expected increase in output forthcoming from the use of an additional unit of the input, the level of other inputs remaining unchanged. In a production process in which more than one production factor is 1The standard errors computed for the tests were .3969 for the bean equation; .1658 for the corn equation; .6451 for the rice equa- tion; and .1284 for the corn-bean combination equation. 2See, for instance, Massell, op. cit., and A. A. Halters, "Production and Cost Functions: An Econometric Survey," Econometrica 31 (January-April 1963): l-66. 127 .mN.m e—eew Eeww eeeweeeoe .mewweeewe wwmw :w meeeeiewee eee .mmee ex om cw eeww .mmee ex on cw eewemeee ewe: ewee eee mceeme .eeees ewpeE:pwwwuoeeewe eewe> weewmwez ww~.p mmo.o mpo.o Foo.o mueecw :weeez Nmo.m~ ow_.o «mm.o mow.o weee4 omm.mww mmw.¢ mmw.¢ wom.~ ece4 mewuw>wpeeeewe wecwmwez weueew eee.w eNe.eew eee.ee ew_.eee w.weV eeowwe eeeeee weee. V ewe.e weee. V eee.e weee. V NNe.e wwee. V eee.e oNe weee. V eeN.e wwee. V ee_.e weee. V wwN.e weNe. V eeN.e owe weee._V eNe._ weew.wV eee.w wNew.wV eee.w wNee._V eee.w weweeeV eeweeoeee weeN.NV eee.wew.w weeN.eV eNe.eew weee.eV eee.eee weee.NV eee.eee w.weV eeeemew eweeez weee._V eee.we weeN.NV wew.ee wwwe.NV eew.ee wwee._V ewe.eN wewee-eeeV eweee4 wee_.NV wNN.N weee.NV wee.w wee_.NV eNe._ wewe._V wwe._ weeweeeeeV neee4 weee.NV eew.eNe.e weeN.eV Nee.ew wNNe.NV ee_.wN weee.NV eem.e e.ee eeee Ameeweew>ee eweeeeemV meeez ePQEem mceemueweu eewm eweu meeem meweewwe> e.ww-ewmw .ez .epez ee eeeN .mewpw>wueeeewe eewe> weewmwee eee .mewuw>wue=eewe wecwmwee weueew .mceee eweEem unease eee eeeceli.m.e e—eew 128 involved, the marginal productivity of any resource depends on the quantity used and on the proportion of the other resources with which it is combined. There was considerable variation in the marginal productivity of inputs in the grains analyzed. The marginal value product of land was higher in the beans and corn-beans combination than in the corn and rice enterprises. The same tendency was also observed with the marginal value product of labor; however, a small difference was observed between corn and the corn-beans combination. It was found that the marginal value product of modern inputs was higher in the rice enterprise, followed by corn, corn-beans combination, and bean enterprises. From the figures presented above, relating to the input marginal products, it follows that the greatest income-raising possi- bility is indicated by increasing crop land. The analysis that follows in the next section, however, deals more closely with factor limitations in grain production. The Economic Efficiency of Input Use For the study area there is no empirical evidence that farmers efficiently allocate inputs so as to maximize output value at market prices. It is, though, of interest to examine the extent to which the actual allocation deviates from an output-maximizing allocation, which is a developmental policy issue. The opportunity costs used to analyze economic efficiency of resources were as follows: (a) for land, the average rent paid 129 per hectare of crop land in the survey year was used, which amounted to Cr$583.00; (b) for labor, the average wage rate paid for daily labor during the agricultural year was used, which was Cr$35.00; and (c) for modern inputs, an interest rate of 13 percent was assumed to be relevant, or Cr$l.l3 per Cr$l.00 of input. This was believed to be a reasonable cost for "capital" and a price at which additional funds could be acquired.1 Table 4.6 presents the marginal returns/factor cost ratios and the test for economic efficiency of resource use for each enter- prise.2 According to the statistical results, producers of beans, on the average, were using land, labor, and modern inputs inefficiently. Modern inputs seemed to be used in excess, and land and labor were underused. The marginal return to opportunity cost of land ratio was higher than the ratio for labor, implying that land was scarcer than labor in the study region. 1The average rent per hectare of crop land was reported by EPAMIG, Informe Agropecuario, ano 3, various issues; the wage rate was reported in the questionnaire. 2A t-test: t = MVPi-Pi/S(MVP1) was used to test for effi- ciency, in which MVPi is the marginal value product of resource X; at the geometric mean; Pi is the price or opportunity cost of that resource, and S(MVP{) is the standard error of the marginal value product obtained as below. The null hypothesis was that MVPi-P1=O, or that marginal return/price ratio was statistically not different from unity, which is the criterion used for efficiency. For computa- tion of S(MVPi), Heady and Dillon's formula was used: S(MVPi)=Y/X - bei’ where prices of product and input are included in the calcula- tions in order to compute MVP's. (See Heady and Dillon, Agricultural Production Functions, p. 231.) For an alternative computation of S(MVPi). see H. 0. Carter and H. O. Hartley, "A Variance Formula for Marginal Productivity Estimates Using Cobb-Douglas Function," Econo- metrica 26 (Jan. 1958): 306-13. 130 Table 4.6.--Marginal returns/factor cost ratios and tests for differences of the ratios from unity of selected grains a37gheygeometric means of resources, Zona da Mata, MG, Variables Beans Corn Rice Corn-Beans Marginal Return/Factor Cost Ratiosa Land 1.72 0.69 0.90 1.51 Labor 1.17 0.81 0.54 0.83 Modern inputs 0.35 1.35 3.72 1.14 Tests for Differences of Ratios From Unityb Land S S S S Labor S S S NS Modern inputs S S S NS aFactor costs were as follows: Crop land rent = Cr$583.00, labor (daily wage) = Cr$35.00, and modern inputs = 13 percent interest rate. bS = significant difference at the 5 percent level; NS = no significant difference at the 5 percent level. Statistical results for corn and rice seemed to be similar. Land and labor were used in excess, whereas modern inputs were under- used. In the case of rice, modern inputs were so underutilized in the production process that the value of marginal product was almost four times the assumed opportunity cost. The statistical results for the corn-beans combination were different. 0n the average, farmers were using labor and modern inputs efficiently. Land seemed to be a limiting factor, so that its margi- nal return was still higher than its respective opportunity cost. 131 In summary, the most limiting resource for beans and corn- beans intercropped seemed to be land, and for corn and rice producers, modern inputs. Labor seemed also to be a limiting production factor for bean producers. This may be related to the fact that sole-cropped bean production may require more skilled labor because this production system requires more attention, such as pest control, improved seeds, fertilizers, and so on.1 The results presented above suggest that in order to increase net returns, the use of modern inputs can be increased in corn and rice production until its marginal return is equal to the assumed opportunity cost of 13 percent, or Cr$l.13 per Cr$l.00 of input. Similarly, more land should be used for bean and corn-bean inter- cropped producers, whereas for bean producers labor use also should be expanded, ceteris paribus. On the other hand, labor and land used in corn and rice production should be reduced, and modern inputs used in bean production should also be reduced. However, with an increase in the use of modern inputs, for instance, the productivity of both land and labor in corn and rice production would be increased, which would imply that these conclusions could change substantially. Minimum Cost Combination of Inputs Table 4.7 presents the optimum input combination and deviations of the actual combination from the optimum at the geometric mean of production of small, large, and the average farms in the grain sub- sector of the study area. For these computations, factor-demand 1Vieira, op. cit. 1132 .eeweeewesee seewuee Eeww .eeuee we eewuew>ee emeueeewee ewe memeeueewee cw mewemww eew e .mwee:eeee— meweueez oo—uem eee meweueez om-o_ u meew emwe— unwee:ee:e— neweuue: epic eee mweeeewoewegm u «swew .weEme ww.wee.w ew.eNe.e ww.wee.e weweV eeweoeeewe eeeeeweee weV ee.__e.w e_.ewe._ we-V e_.eeN._ ee.we_._ we_-V ew.ewe we.eew weweV eeeeew eweeez weeV we.ew ee.eew eeV ee.ee ew.we weeV we.ee ee.we weeee-eeeV weee4 wwe-V ee.e Ne.N ee-V we.e eN.N wee-V w_.N ee._ weeV eee4 meeemueweu weeee ex eeV ee.eN ee.e_ we.ew eeweoeeewe eeeeeweee eNN-V ee.ewe e_.ee. eeN-V .w.eee ee.ee_ weee-V ew.eee ee.ee weweV eeeeew eweeez we eN.ee ee.ee we ee.eN ew.em AeeV ee._N e_.we wewee-eeeV weee4 N. Ne.w Nw._ e. NN._ ee._ we_V .e. ee._ weeV eee4 oowe weeee ew eeV ee.ee eN.wN ee._N eeweoeeewe eeeeeweee we- ew.eew e_.eee ee-V ee.eee ee.eee ee-V Ne.eee ee.weN weweV eeeeew eweeez ew- ee.ee eN.ee ewV we.ee e_.ee eNV ee.wN __.we wewee-eeeV weee4 w_V Ne._ ee.. .NV e... ee._ eNV ee. .e._ weeV eee4 eweu weeee ex eeV ee.m ee.e ee.e eeweoeeewe eeeeeweee wew we.eew ee.eee wwwV Ne.ee. ee.eee wwwV ew.ee_ ee.Nwe weweV eeeeew eweeez we-V ew._N ee.eN eV ee.e_ ee.eN weNV NN.ew e...N wewee-eeeV weee4 wee-V we.w e_._ ee-V ee._ e_._ we_-V w... ee. weeV eee4 meeem Eeswueo peeee< seswueo peeue< seewueo weeue< ewew eewe4 Ewew emewe>< ewew wweEm ecewueeweeeu euweemea e.wwuewm. .ez .eeez ee eeeN .eeeeeweweuew eeeeuewee eee .eeww .ewee .eeee we Ewew emwe. eee .ewew emewe>e .Ewew —_e5m we eewueeeewe we :eee ewwueeeem en» ae Eeswuee esp eeww eeweeeweeee useew peeuee we meewuew>ee eee eewuecweeee useew Eeewueoii.w.e e—eew 133 equations were used. It is observed that relatively small land adjustments are required for small farm producers of beans. About six man-days and about 70 percent of expenditure in modern inputs could be allocated in other activities. About the same percentage of larger farms' expenditures in modern inputs could also be reallo- cated to other activities, whereas about 45 percent increase in planted land would be required to operate in the optimum input combination. Observe that, in contrast to the other farm sizes, the adjustment on large farms would require a 9 percent increase in labor used. In the case of corn and rice, it is observed that relatively little land has been overused. However, especially in the case of rice, to attain minimum cost combination of inputs, about 50 percent and 40 percent of the labor of small and large farms, respectively, should be reallocated in other activities. About 26 percent of the labor force used for small farms exceeds the optimum labor level required in corn production, whereas it would be necessary to increase the labor used by large farms by 13 percent to attain that optimum. A possible intervention of the PRODEMATA for achievement of minimum combination of resources is associated with adjustments of modern inputs in production of corn and rice. The average farm of the Zona da Mata would require an additional 64 percent and 270 percent of modern inputs in corn and rice production, respectively. The small farms require an 80 percent increase in modern inputs for corn produc- tion and 340 percent for rice. About 40 percent and 225 percent additional modern inputs would be required for the large producers of corn and rice, respectively. 134 For the producers of corn-bean combination, land should be increased about 40 percent, the smaller farms require an additional 16 percent modern inputs, and about 40 percent of the labor force used in corn-bean combination should be allocated to other activi- ties. If the required land adjustment can be attained in practice and PRODEMATA is able to supply credit for increasing modern inputs to achieve the optimum input combination, the main concern falls in the labor arena. The increase in labor use in some activities analyzed seems to present no real problem because of the relatively abundant supply of labor in the region. However, the question that remains to be answered relates to the alternative employment for the labor to be released. Gains From Operating at the Least-Cost Combination Gains for the average farm from producing with the least-cost combination of resources are based on several assumptions. From the side of the input market, perfectly elastic supply is assumed. In the case of labor, for instance, one more required unit of this input in the productive process will be fully used and will receive its opportunity cost. On the other hand, if one unit of labor is released from the productive process, this input will be allocated into other farm activities or off-farm employment receiving the opportunity cost considered in the analysis (Cr$35). This assumption is extended mutatis mutandis to other inputs. From the side of the output market, 135 perfectly elastic demand and no constraints in marketing the product in neighboring regions are assumed. Under these assumptions, the average farm would gain Cr$640 from rice production and about Cr$600, Cr$210, and Cr$80, respectively, from corn-bean combination, bean, and corn production. These figures are especially important for sharecroppers and 0-10 hectares land- owners. For these small farmers who grow rice, that gain represents a 9 percent increase in their net income. For those who grow corn- beans combined in the same stand, beans, and corn, their net income would increase by 9 percent, 3 percent, and 1 percent, respectively. Summary and Conclusions In attempting to estimate production functions of grains (corn, beans, rice, and corn-bean combination) for the Zona da Mata, it was observed that the groups of small and large farms belong to the same regression model. In the pooled production functions, the estimated coefficients of land, labor, and modern inputs were statistically significant at the 5 percent level or less, with the exception of modern inputs in the bean equation, which was not significant even at the 20 percent level. The average years of education of the head of the family and his wife, as proxy for management, was not signifi- cant in the corn-bean equation even at the 20 percent level. The pro- duction functions of beans, corn, and rice for Juiz de Fora and the production function of corn-bean combination for Muriaé differed from the function for the rest of the study area. Tests conducted on the sum of the production elasticities indicated constant returns to 136 scale of grain production in the Zona da Mata, which is consistent with many empirical studies reported in the literature.1 The analysis of marginal return/factor cost ratios indicated economic inefficiencies in land, labor, and modern inputs allocation in the production of beans, corn, and rice. In the case of corn-bean combination production, economic efficiency in labor and modern inputs allocation was observed; however, land was underused in this production process. The figures presented in Table 4.8 illustrate the per hectare value of inputs and outputs in the study area. These figures help to understand the reallocation of resources suggested by the analysis developed in this chapter. In light of the opportunity cost used in the analysis, for the average farm, it is suggested that the use of land for bean and corn- bean combination production be increased to 30 percent and 40 percent, respectively. The use of this factor should be decreased by 20 and 13 percent, respectively, in corn and in rice production. Labor use should be expanded on large farms in bean and corn production, whereas for other enterprises and for the small farms it should be decreased. Decreasing labor use, on the other hand, is compensated by increasing the use of modern inputs. It is suggested by the analysis that in the combination corn-bean, corn, and rice production, for the average farm, modern inputs should be augmented by 4 percent, 64 percent, and 269 percent, respectively. 1See Walters, 1963, op. cit. 137 Table 4.8.--Value of output and input per hectare of selected grains, Zona da Mata, MG, l976-77. Types of Enterprisesa Output-Input Beans Corn Rice Corn-Beans Output 1,841.25 1,633.93 32,100.20 2,886.22 Labor 657.30 861.25 1,305.41 1,294.15 Modern inputs: 536.34 442.42 258.73 630.90 Seeds 296.76 52.21 113.62 225.16 Fertilizersb 167.96 191.97 18.46 249.11 Pesticides 8.70 6.44 .74 7.49 Machinery services 10.59 72.78 18.30 52.05 Animal services 52.33 119.02 107.61 97.09 (1) Variable input cost 1,193.64 1,303.67 1,564.14 1,925.05 (2) Gross margin 647.61 330.26 536.06 961.17 Returns to variable 54 25 34 50 cost (percentage) aMeasured in cruzeiros per hectare. Approximately 12.50 cruzeiros equal 1 U.S. dollar. bIncludes chemical fertilizer and manure. cFor computation, the following formula was used: [(2) e (1)] x 100. Under the assumptions of perfectly elastic supply and per- fectly elastic demand of input and output markets, respectively, the gains from operating at the least cost combination were estimated. For the average farm they were estimated to be about Cr$210, Cr$80, Cr$640, and Cr$600 for bean, corn, rice, and corn-bean combination 138 farms, respectively. The smaller farmers would benefit more from operating at the least cost combination. For producers of rice, net income would increase by 9 percent, whereas for producers of corn- bean combination, beans, and corn, their net income would increase by 9 percent, 3 percent, and 1 percent, respectively. CHAPTER V SUMMARY AND CONCLUSIONS Introduction The purpose of this chapter is to present the summary and conclusions of the study. The first section provides an overview of the study with a focus on the problem, objectives, and methodology. In the second section the summary of findings is presented. The implications and policy issues are presented in the third section. Finally, some limitations and suggestions for further research are included in the fourth and last section. Summary of Problem, Objectives,pand Methodology Rural poverty has been of increasing concern among Brazilian policy makers. Among the many backward areas of rural Brazil is the Zona da Mata of Minas Gerais State, which is the study area of this research. In 1975 the rural per capita income of the Zona da Mata was about US $250, which was 25 percent of the per capita income of the country as a whole. Per capita income in the study region was below the national poverty level, which was estimated to be about us $340.1 The region's social infrastructure is poor, and available health and education services are deficient. The health status of 1World Bank, op. cit., p. 14. 139 140 the inhabitants of the region is characterized by high mortality and morbidity rates, caused by communicable diseases; a high infant mortality rate, caused mainly by infectious diseases; and a serious incidence of malnutrition. Investment in education in the rural Zona da Mata is rela- tively low. The limitations of educational opportunities seem to be reflected in the low educational attainment of the labor force. In 1970, about 60 percent of the agricultural workers of the region had not received any formal education. Action toward bringing about changes in the study area has been implemented. Several state organizations have combined their efforts with those of the World Bank to implement a conscious develop- ment program (PRODEMATA). The cost of the project is estimated at US $139 million; agricultural credit is its main component (about 60 percent of the project cost). The beneficiaries of this service are primarily sharecroppers and small farmers with fewer than 100 hectares of rainfed land. Other important components of the PRODEMATA are rural electrification; land reclamation; production support services, including agricultural research, technical assistance, and cooperative services; and social services, including investments in basic aspects of health and education. Because of the many activities developed on different-sized farms as well as different opportunities for off-farm jobs, it was considered important to study the process by which farm income is determined. Also, because grains are the principal source of food and income, especially for smaller farmers, coupled with the fact 141 that this subsector is a potential recipient of a large percentage of credit from the PRODEMATA, major attention was centered on the economic efficiency of production of food and feed grains. These factors constitute the central incentive for this study. The overall objective of this research is to generate better knowledge about the process of income determination in the Zona da Mata farm sector. Particular emphasis was placed on microeconomic aspects of the farming system and managerial abilities of the farmer that constitute the center of all interactions of the components of the system. The specific objectives include: development of a con- ceptual framework of the income-determination process for the Zona da Mata farmers; analysis of the grain-production subsystem of the study area, especially the economic efficiency of this subsector; and, finally, in light of the results, suggestion of alternative actions to be incorporated in the project to improve small-farm income in the study area. Small-farm groups include sharecroppers and farmers with farms ranging in size from 0 to 100 hectares, who are potential beneficiaries of the PRODEMATA. The instruments of the analysis used to reach the objectives of the study are mainly tabular analysis coupled with analysis of variance, neoclassical theory of production, and multiple regression analysis. The conceptual framework developed in this study (Figure 3.1) identifies the major determinants of the farm income. The primary aim of the farming system research approach is ”to increase the overall efficiency of the farming system; this can be interpreted as developing 142 technology that increases productivity in a way that is useful and acceptable to the farm family, given its goal(s), resources and l constraints." This was the main motivation for analyzing the income- determination conceptual framework and the economic efficiency of the grain-production subsector of the study area. The results of these analyses are presented in the following section. Summary of Findipg§_ This research followed the sample stratification used by the monitoring and evaluating team of the PRODEMATA. The strata analyzed were sharecroppers, and landowners with 0-10 hectares, 10-50 hectares, 50-100 hectares, and 100-200 hectares. The summary of findings pre- sented below follows the income-determination conceptual framework presented in Figure 3.1. The analysis of resource endowment in the study area sug- gested that land was a relatively scarce factor for the majority of farmers. Even within a particular class of producers, relatively high variation of land ownership was observed, suggesting a probable high concentration of ownership of this production factor. Indeed, ' empirical computation of the Gini Ratio of land ownership for the study area (about .68) suggested a high concentration of land owner- ship.2 Conversely, family labor seemed to be a relatively abundant production factor in the study area. The major source of labor is 1Norman, op. cit. For more information about this approach see John L. Dillon et al., "Farming Systems Research at the Interna- tional Agricultural Research Centers" (Armidale: The University of New England, September 1978). (Mimeographed.) 2Silva, op. cit. 143 from males over 16 years old, who comprised 83 percent of the labor force of the study region. In general, investment in farm buildings and equipment was low. It was estimated that investment in housing (proprietary and employee houses) accounted for 70 percent of the total investments in buildings, including housing, storage, and ani- mal facilities. Investment in machinery and equipment is generally low, especially among sharecroppers and 0-10 hectares landowners. As expected, investment in animals increased with farm size and, conse- quently, with availability of pasture and investment in animal facilities. The analysis of resource use and productivity revealed that about 60 percent of the land of the region was used for pasture. Coffee, another nonsubsistence activity, occupied a considerable part of the region--about 7 percent. About 21 percent of the land was used for grain production; about 9 percent of the total farm land was used to produce corn, 6 percent to produce beans, and the same percentage to produce rice. About 40 percent of the cropped area was used to produce corn, and about 26 percent and 23 percent were used to produce beans and rice, respectively. Smaller farmers used a larger percentage of land in grain production. In attempting to verify differences in output per hectare of the several enterprises (corn, beans, rice, sugar cane, tobacco, coffee, fruits, and vegetables), results of the analysis of variance suggested no statistical differences at the 5 percent level among the different classes of farmers. There was great variability of labor used per hectare of the most common crops grown in the study area, 144 especially in the grain subsector. However, production of rice per man-day was not statistically different among farmer groups at the 5 percent level. Capital was divided into four categories: farm assets (K1), livestock (K2), operating expenses (K3), and permanent structures (K4). These various forms of capital were analyzed using ratios, with land and labor as the basis. The main conclusions drawn from the analysis are that investments in machinery, equipment, and work animals per hectare of land do not differ statistically at the 5 percent level among farmer groups. In turn, all ratios of capital/utilized labor were statistically different among farmer classes at the 5 percent level. The analysis of labor/land ratios, on the other hand, suggested a decreasing tendency of labor/land ratios as farm size increased. Finally, the smaller farmers used less credit and participated less in cooperatives. Four proxies were used for management efficiency, including: (1) farmer age, (2) farmer education, (3) number of days worked off the farm, and (4) commercialization index. The arithmetic mean of farmers' age was about 50 years, and it did not differ statistically among groups of farmers at the 5 percent level of significance. In general, the percentage of illiteracy decreased as farm size increased. About 43 percent of the sharecroppers were illiterate, while all of the 100-200 hectares landowners group were literate. Sharecroppers also worked off the farm but in the agricultural sector more than did other groups of producers. On the other hand, that group of producers worked fewer days off the farm and outside the agricultural sector. 145 Finally, analysis of the commercialization index indicated that those farmers who had the highest commercialization index were those who owned the most land. The sharecroppers, however, had a higher com- mercialization index than did the 0-10 hectares landowners. Education seemed to be a good proxy for management and was the variable used in the grain subsector economic efficiency study developed in Chapter IV. Average family size in the Zona da Mata comprised 5.71 per- . sons (including other people living with the family). Sharecroppers I had the most children, followed by farmers with 50-100 hectares and -1 those with 100-200 hectares. Those with sons and daughters older than 15 years were more common among larger farmers. Analyses of production, consumption, and marketable surplus were conducted for 11 products: rice, beans, corn, coffee, fruits, vegetables, cattle, swine, poultry, milk, and milk products. Initially, the analysis was conducted for the whole Zona da Mata; then the sample was broken down into the five categories of producers. In aggregated terms, it was observed that more than 50 percent of the production of beans, corn, and poultry was intended for on-farm con- sumption. Also, production of grains was very common among farmers. About 80 percent of the sampled farmers produced corn, and about 70 percent and 60 percent produced beans and rice, respectively. The sharecroppers seemed to be more involved in producing subsistence crops, such as corn, beans, and rice. There was a tendency, however, to raise poultry and to participate in coffee and milk-production contracts. Besides grain production, a large percentage of 0-10 hectares landowners grew coffee and raised poultry and swine. This 146 group of farmers had smaller amounts of grains as marketable surplus. In turn, production of coffee, cattle, milk, and milk products had characteristics of commercial enterprises. For farms larger than 10 hectares (besides grains), production of coffee, cattle, and milk was the most common in the study area. Analysis of variance was conducted on prices received by farmers for eight products of the study area. The results suggested that, with the exception of swine production, prices did not differ among farmer classes at the 5 percent level of significance. These results seem to indicate that farmers had similar storage facilities or sold their commodities in a similar period of the year (except for swine, which seems to be the reverse). The Zona da Mata average net farm income was about Cr$29,500, several times more than sharecroppers' and 0-10 hectares landowners' net farm income--about Cr$6,200 and Cr$7,600, respectively. The 50-100 hectares and 100-200 hectares landowners' net farm incomes were above the average net farm income for the region. The 10-50 hectares land- owners' net farm income was about the average level for the region. Off-farm income represented a substantial percentage of the total family income for sharecroppers and 0-10 hectares landowners--about 55 percent and 43 percent, respectively. The family incomes of these groups of farmers were below the regional average. The Gini Ratio of family income was about .64, implying a relatively high income concentration. Considering the share of all products grown by the various classes of producers of the study region, as well as the percentage 147 of producers who grow these products, a typical farm was identified for each stratum analyzed. For the sharecroppers, production of grains (corn, rice, and beans) was the typical farm enterprise. Adjacent enterprises for these producers are poultry and coffee pro- duction. However, off-farm income constitutes their most important source of income. The typical farm of 0-10 hectares landowners also is characterized by production of grains. Other enterprises of importance for this category of farmers are poultry, coffee, and swine. Nevertheless, close to 50 percent of the family income of these pro- ducers is from off-farm sources. Off-farm income becomes relatively less important for the groups of farmers with more than 10 hectares of land. In the case of the 10-50 hectares landowners, production of grains, milk, poultry, coffee, cattle, and swine is the most common activity on a typical farm of this group. A typical farm of 50-100 hectares landowners raises grains, milk cows, coffee, swine, and poultry. A similar combination of enterprises is observed on a typi- cal farm of 100-200 hectares landowners. However, for the 50-100 hectares farmers, the highest share of their gross income is from milk sales, and for the 100-200 hectares farmers, coffee generates the highest income share. In the process of analyzing the production economics of the grain subsector of the study area, it was observed that the groups of small and large farmers belong to the same regression model. In the grain-production functions of beans, rice, corn, and corn-beans intercropped, the estimated coefficients of land, labor, and modern inputs were statistically significant at the 5 percent level or less, 148 with the exception of modern inputs in the beans equation, which was not significant even at the 20 percent level. The average years of education of the head of the family and his wife, as a proxy for management, was significant (at the 20 percent level or less) except in the corn-beans equation. The estimated coefficients of the dummy variables included in those production functions indicated that envi- ronmental differences of Muriaé did not have an effect on the produc- tion of beans, corn, and rice. The same effect was observed for envi- ronmental differences in Juiz de Fora on the production of corn-beans intercropped. Economic inefficiencies in the allocation of land, labor, and modern inputs were observed in the production of beans, corn, and rice. In the case of corn-beans combination production, economic efficiency in labor and modern input allocation was observed; however, land was underused in this production process. Factor demand equations were used to determine optimum input combinations at the geometric means of production for each enterprise of small farms, the average-sized farm, and large farms.1 For the average Zona da Mata farm to operate at the optimum input combination, land should be increased about 30 percent in beans and about 40 percent in corn-beans combination production. In the case of the production of corn and rice, the average farm should decrease land use by about 20 percent and 13 percent, respectively. Labor use should be decreased 1See footnote to Table 4.1 for the definition of farmer classes as well as the sample used in estimating the production functions. 149 in all enterprises and in all sizes of farms, with an exception in the case of beans and corn production on large farms, which would require a labor increase of about 10 percent. Modern input use should be decreased by farmers producing beans and among the large producers of corn-beans combination. For the other producers of corn-beans combination as well as corn and rice, a substantial increase of this input is required, especially in rice and corn production. Implications and Policy Issues The most striking characteristics of the Zona da Mata farming system, evidenced in this study, are the different resource endow- ments, resource uses, and emphasis on different enterprise combinations, determining various levels of farm family incomes. It becomes logical, though, to define specific target groups and to design and test poli- cies that meet their characteristics and needs so that developmental actions can be effective in changing the actual poverty scenario of the study region. A group of producers that deserves the special attention of policy makers is the sharecroppers. This group of producers consti- tutes an important labor source for landowners. However, it is not clear to what extent that group constitutes a barrier' to landowners for augmenting resource ownership, especially capital, and enabling them to move toward commercial agriculture. 0n the other hand, it is questionable whether the presence of sharecroppers on large farms is not a means whereby landowners can gain control over greater quan- tities of land and capital. Besides, sharecroppers might be the 150 part of the rural population that can migrate more easily to the urban areas since they do not have a high investment in their businesses. The problems brought about by migration are twofold: First, since the migrants have a low level of education and no training to engage in urban jobs, they will aggravate social problems in the large cities. The second problem is associated with adjustments farmers have to make to fill the outmigrant's position. Because it is reasonable to assume that sharecroppers will continue to constitute a large proportion of the Zona da Mata popula- tion in the long run, actions should be taken to improve this group's level of living. Yet, because 0-10 hectares landowners' farming- system characteristics were found to be similar to those of the sharecroppers, policies aimed at increasing the incomes of both groups are examined together. As defined by INCRA, the minimum farm size in the region that would provide full employment and income for a family with four workers is about the average farm size found in this study (about 27 hectares). Considering land as a scarce factor for these producers, it would be desirable to promote changes in the existent pattern of land distribution. Needless to mention, this measure would imply practical difficulties. However, gradual changes could be pro- moted to avoid the proliferation of small farms, such as prohibiting division of farms (in the case of inheritance, for instance). Also, the economic feasibility of cooperative farming should be evaluated 151 in the study area since it has been proven to have economic advan- tages for farm members in other regions of the country.1 Other measures that could be undertaken and that could produce faster results are leasing contracts. Deals involving larger areas could be promoted, as should long-life contracts. These arrangements are likely to bring better resource efficiency in the study area as well as to increase agricultural output.2 Despite the peculiar resource limitations of these producers and the traditional agriculture they practice, in aggregated terms they represent a substantial proportion of the food suppliers for the region. Considering that the great majority of these producers have not received credit or technical assistance, it is believed that the implementation of the PRODEMATA might increase substantially the participation of these farmers in regional agriculture. However, all innovations brought by the project should be consistent with the farm- ing system of these producers and in no way should the priorities of the family and the characteristics of their natural environment be disregarded. The level of education is a bottleneck for efficient credit use and adoption of new agricultural techniques. As much as 'Dias, 1979, op. cit. 2Berry and Cline empirically observed that the small-farm sector makes better use of its available land than does the large- farm sector. Developmental strategies focusing on small farms-- whether they involve land redistribution or improved access of small holders to credit, new technology. etc.--are likely not only to have beneficial distribution and employment effects but also to be efficient means of increasing output. See R. Albert Berry and William Cline, Agrarian Structure and Productivity in Developing Countries (Baltimore: The Johns Hopkins University Press, 1979). 152 possible, the extension service should increase their interaction with farmers in their decision making. It is suggested that off-farm job opportunities for these producers be increased. Small agro-industries should be created in the study area, using regional raw material as well as the local labor force. 0n the other hand, out of the peak labor demand for agricultural production, surplus labor could be allocated to main- tenance of secondary roads (as has been done in other regions of the country). Training should also be provided to these producers to increase their skills to be used in the farm sector as well as in the urban sector so that potential migrants could achieve better income levels. Attention is now drawn to alternatives to increase income of 10-50 hectares landowners. Most of the considerations addressed to the sharecroppers and 0-10 hectares landowners also apply to the 10-50 hectares landowners. From this farm size category to larger ones, a certain level of specialization is observed, principally toward milk and coffee production. The increase in milk production of the 10-50 hectares landowners is limited primarily by the pasture land and animal facilities. Introduction of new technologies, especially in improved animal feeding systems such as silage, improved quality of the herd, and improvements of the animal facilities, is suggested. Coffee sales alone account for one-third of the income of those producers. A promising source of income for this category of farmers as well as a way to increase labor demand in the region would be to expand coffee plantations. A detailed economic-feasibility 153 study should be conducted and, because of the relatively high initial cost of such plantations, new forms of credit for this purpose should be explored. Technologies that meet the characteristics of these producers' farming system should be coupled with agricultural extension service and agricultural credit to promote the harmonious flow of cash-crop and food-crop production. Labor-intensive technologies and partner- ships should be emphasized. Among the potential beneficiaries of the PRODEMATA, the 50-100 hectares landowners had the highest income. For some of these farmers, the observations made above about milk and coffee production are plausibleneesures to increase their incomes. The importance of this group is recognized for accomplishing suggested contracts with smaller producers, principally if labor-intensive technologies are emphasized in the area. Finally, action toward verifying the possibility of reallo- cating resources in the enterprises of the region, especially in the grain subsector (which was empirically analyzed in this study), is recommended. Such action would provide economic benefits for the producers, especially if agricultural credit from the PRODEMATA could be coupled with the creation of new job opportunities in the area. Limitations and Suggestions for Further Research The data analyzed in this study were intended to capture general views of the region limiting further extensive analysis of the farming system. Additional characteristics of the exogenous and 154 endogenous factors of the human element of the analysis (see Appen- dix 0) should be incorporated in future analyses. They might suggest new social relationships with other variables of the framework that could not be accomplished with the data analyzed. It would also be relevant to investigate the farming systems in a dynamic perspective. In this sense, an interdisciplinary team (comprising both social and technical scientists) could generate periodic information about the system, providing a solid basis for possible policy intervention. The Universidade Federal de Vicosa, which is also responsible for evaluation of the PRODEMATA, is located in the Zona da Mata and could with least cost lead efforts to imple- ment such a system. Very little is known about the effects of Brazil's current inflationary trend on the farm sector, especially on small producers. Emphasis for research should be placed on determining better combi- nations of inputs, enterprises, and farm investments to hedge against inflation. Questions remain to be answered about the use of credit, especially the correlation that exists between the use of credit and growth, as well as production efficiency. The role of capital in the farming system and priorities of credit application as well as effective policies and programs for the delivery and repayment of credit should be investigated. As stressed in the above section, more should be learned about the regional labor market. The use of labor in periods of peak 155 demand and slack periods should be determined for actions toward creat- ing off-farm jobs, and the potential effects of introducing small tractors and machinery into the region's labor market should be analyzed. Research is also needed on the various markets in which the smaller farmers can participate. The crops and enterprises espe- cially suited to small-scale production should be identified and investigated. Following this, the economic feasibility of creating cooperatives to provide inputs as well as to market farmers' output should be investigated. Apart from the formal agricultural extension service, commu- nications research for various enterprises could have high payoffs in the Zona da Mata. The radio could be tested as a means of inform- ing farmers about prices and the proper times to employ various agri- cultural practices. Pricing of agricultural products and inputs is a major area that demands the attention of policy-oriented research. Alternative pricing policies and their impact on production and consumption of major products (as well as inputs) deserve special research priority. Finally, no risk component was considered in this study. It is suggested that the risk component of the farming system be identi- fied and investigated so that developmental policies addressed to a target group of farmers could implicitly consider that group's risk preferences. APPENDICES 156 APPENDIX A SHARECROPPERS DEFINITION 157 APPENDIX A SHARECROPPERS DEFINITION In this study the sharecroppers are defined as landless or tenants. Under contract, they use landowners' land in exchange for payment in kind, in cash and/or production costs. The institutional contracts between the landowners and sharecroppers are regulated by Law No. 4.504, "Estatuto da Terra," established on November 30, 1964. This law determines the rights and obligations of both parties under several circumstances the contracts are set. Specifically, it deter- mines the minimum period of time for those deals, each party's share of the production under various arrangements of production cost sharing, etc. However, in practice, it is observed landowners exert- ing their power, imposing favorable deals for themselves. 0n the other hand, one may observe arrangements between landowners and their relatives benefiting the latter. This situation is more common for perennial crops in father-son sharecropping contracts.1 Table A.l presents the participation of the surveyed share- croppers in some crop deals. In the case of grains, the most common contracts involve sharecroppers' payment in kind of 50 percent of the production. Few sharecroppers were engaged in contracts involving production of coffee, tobacco, and sugar cane. Most of these 1Universidade Federal de Vicosa, DER, Programa Integrado do Desenvolvimento da Zona da Mata--MG. Primeiro Relatdrio Anual de Avaliach (Vicosa, MG: Marco 1979). 158 159 contracts established one-third of the production as payment to the landowners. Table A.l.--Participation of sharecroppers in contracts with land- owners 0; selected enterprises, Zona da Mata, MG, l976-77. Production Proportion Paid Number of Crops Contracts to the Landowner b 1/2 1/2 1/4 Others Rice 77 32 c 39 4 2 (41.6) (50.6) (5.2) (2.6) Beans 94 17 74 2 1 (18.1) (78.7) (2.1) (1.1) Corn 109 17 88 l 3 (15.6) (80.7) ( .9) (2.8) Coffee 22 19 3 -- -- (86.4) (13.6) Tobacco 6 6 -- -- -- (100.0) Sugar cane 5 4 1 -- -- (80.0) (20.0) Source: Sample survey. aBased on a sample of 129 sharecroppers. bRefers to other arrangements between landowners and share- croppers. cThe figures in parentheses are percentages of contracts of each kind. APPENDIX B THE GINI INDEX OF INCOME CONCENTRATION OF THE ZONA DA MATA 160 APPENDIX B THE GINI INDEX OF INCOME CONCENTRATION OF THE ZONA DA MATA In determining the distribution of income among the Zona da Mata surveyed farms, gross farm income was computed and divided into 15 classes of income (Table B.l). The figures in Table B.l indicate that as farm size increased, gross farm income also tended to increase. About 60 percent of the sharecroppers and 0-10 hectares landowners had gross income less than Cr$10,000. The percentage of the sample included in that income cate- gory was about 35.5 percent, which accounted for 4.4 percent of the total sample gross income. The concentration of income becomes more evident as one moves to subsequent classes of gross income. The first three classes of gross income, for example, which included about two- thirds of the entire sample, had only 19.5 percent of the total sample gross income. About 11 percent of the farmers had 54 percent of the gross income generated in the Zona da Mata. The understanding of the computation of the Gini Ratio is facilitated by using the Lorenz Curve shown in Figure B.l. This curve is derived by plotting the cumulative fraction of the total income against the cumulative fraction of the units receiving this income, where the income-receiving units are arranged from poorest to richest income classes. If the Lorenz Curve coincides with the Line of Equality, every unit has the same income. 0n the other hand, in the absence of complete income equality, the Lorenz Curve lies below the 161 162 .ze>w:m eweEem Neeweem e.eew e. e.e w.: e. e e ewee eee eee.eNe e.ee e. w.wN e e. e e eee.eNe-eee.eeN e.ee e. e.e. e e. e e eee.eeN-eee.eeN e.ee e. e w.w e. e e eee.eeN-eee.eeN e.we N.N e.e_ e.e e.e e e eee.eeN-eee.ee_ e.ee e.N e.e e.e, e.N e e eee.eew-eee.er e.Ne e.e e.ew e.ew e.w e e. eee.eN_-eee.ee e.we e.N e.e w.e e.e e. e. eee.ee -eee.ew e.ee e.N e.e e.e w.e e. e eee.ew -eee.ee e.Ne w.e e.e e.ee e.e e e.N eee.ee -eee.ee w.ww e.e e N.ew e.e e._ e. eee.ee -eee.ee e.ew e.e e w.w w.e e.N N.e eee.ee -eee.em w.we e.e e.ew w.e w.ew w.e N.e eee.ee -eee.eN w.we N.NN e e.e e.eN e.eN e.eN eee.eN -eee.e_ e.ee e.ee e e.e N.ew e.Ne e.we eee.eww -nuwwwe eweewe ee eeN-eew ee eee-ee ee ee-ew ee ew-e eweeeeweeweee weewweeeweV eeeeee mmeww AwV meweceeeeww memmewu weeeeewe we emeeeeewee we memmewo .wwuowmw .ez .eeez ee eeeN .mweeeeewe we memmewe e>ww meeEe eeeeew mmewm Ewew we eewpeewwumwou-.—.m eweew 163 Cumulative 1 Fraction of Income b----.-----.--. b--..---- O ‘h .4, 1 d u-lo + d Cumulative Fraction of Units Figure B.l.--Illustration of a Lorenz Curve. 164 Line of Equality, as is the case shown in Figure B.l. The Gini Ratio can be derived from the Lorenz Curve, where it is the proportion of the total area under the diagonal that is between the Lorenz Curve and the diagonal. Using Figure 8.1, the Gini Ratio is expressed as follows: Area between the Lorenz Curve A _ and the Line of Equality, (C 1) 9101 Ratlo = A + B ' Area under the Line of Equality For the computation of the Gini Index of income concentration or the Gini Ratio, the following formula was used: k Glnl Ratlo = l - 1:] (fi+l - f1) (yi + yi+]) (C.2) where: fi = cumulative fraction of units yi = cumulative fraction of income k = number of classes This is the formal mathematical presentation by Riemenscheider1 of Morgan'sz discussion of the Gini Ratio. The computed income Gini Ratio of the surveyed Zona da Mata farmers' farm income was about .64 in the l976-77 agricultural year. This ratio indicates more income concentration in the study area rela- tive to the state of Minas Gerais, whose estimated income Gini Ratio was about .54. 1Riemenscheider, op. cit. 2Morgan, op. cit. APPENDIX C TESTS FOR HETEROSCEDASTICITY 165 APPENDIX C TESTS FOR HETEROSCEDASTICITY This appendix presents tests for heteroscedasticity for the estimations presented in Table 4.4 of Chapter IV as well as the simple correlation matrices of the variables of those estimations. Test for Heteroscedasticity Park1 formalized the graphical method by suggesting that o? is some function of the explanatory variable Xi' The functional form he suggested was 0? = 02 )(i evi (0.1) 01" 1n 0? 1 1n 02 + 6 1n Xi + v. (C-2) 1 where Vi is the stochastic disturbance term. Park suggested use of e? as a proxy for 0? because this parameter is generally not known. The presence of heteroscedasticity in the data would be sug- gested if s is statistically significant. In turn, the homoscedas- ticity hypothesis may be accepted if B is not statistically signifi- cant. The residuals obtained from regression presented in Table 4.4 were regressed on X, (land, labor, modern inputs) as suggested in equation (C.2), presenting the following results: 1Park, op. cit. 166 Beans Corn Rice ln ln ln ln ln ln 1n 167 .2789 + .0191 In (.1025) t = .1861 .0873 + .0639 1n (.0999) t = .6394 .1402 + .0228 1n (.0706) t = .3230 .2540 + .0073 1n (.0527) t = .1392 .1525 + .0287 1n (.0538) t = .5340 .5103 - .0429 1n (.0279) t = -1.5375 R .4482 - .0566 In (.0382) t = -1.4809 R (Land) R2 = .0008 (Labor R2 = .0094 (Modern Inputs) R2 = .0024 (Land) R2 = .0002 (Labor) R2 = .0032 (Modern Inputs) 2 = .0262 (Land) 2 = .0067 168 1n e. = .5078 - .0196 1n (Labor) (.0426) t = -.4585 R2 = .0006 1n e? = .5473 - .0240 1n (Modern Inputs) (.0236) t = -l.016 R2 = .0032 Corn-Beans 1n e? = .3599 - .0863 1n (Land) (.0437) t = 1.9763 R2 = .0149 ln e? = .0431 + .0554 1n (Labor) (.0523) t = 1.0589 R2 = .0043 1n e. = .4178 - .0179 1n (Modern Inputs) (.0330) t = -.5419 R2 = .0011 As the estimated equations indicate, there is no statistically significant relationship between the two variables. Thus, following Park's test, one may conclude that there is no heteroscedasticity in the error variance. 169 eeee._ eeee.- eeew. eeNN. eeee.- eeew. eeew. N wEEee wwV eeee._ wNew. eee_.- eNee. eeee.- eeNN.- w essee weV eeee._ wwwN. weeN. eeNe. weee. eeeeew eweeez weV eeee.w wmee.- eeme.- eeeN. eeweeeeee weV eeee.w eNee. eewe. weee4 me eeee.w NNew. eee4 wNV eeee.w eeeeee wwV wwV weV weV weV weV wNV wwV e.eewwee .¢.¢ $7.05. a; mcmmn Low COED—2C. UwqumHmm .._.o mwpnmem> mnu mcoEm mcome—MLLOU eweewe--.w.e eweew 170 eeee.w eeee.- weee. eewe. eeew.- eeNe.- wewe. N weEee wwV eeee._ ewew. eeee.- eeNe.- Newe.- eeew.- w easee weV eeee._ weeN. eeeN. Neee. eeee. eeeeew eweeez weV eeee._ eeee.- ewew. ewNN. eeweeoeee weV eeee.w wewe. eewe. weee4 weV eeee._ weee. eee4 wNV eeee.w eeeeee wwV wwV weV weV weV me wNV wwV eweewwee .¢.¢ wPDMH .3. :L00 Low cowuuczw OmeEmem ..._.o mwpnmwgm> m...“ 9.05m mcowampwggou eweewe--.N.e eweew 171 oooo._ moo¢.u emmp. emmo. Nmmo. epwo. Nem—. N >.5:an va eeee._ wNNe.- eewe.- eeeN.- eeee.- e_e_.- _ wEEee weV eeee.w eeeN. meme. ewwe. weee. eeeeee eweeez weV eeee._ eeeN. wemN. eeee. eeweeeeee weV eeee.w ewee. weew. weee4 weV eeee.w wwee. eee4 wNV eeee.w eeeeee wwV wwV weV weV weV weV wNV wwV eweewwee .e.e eweew cw eeww wew :ewpeeew eeeeewpme we mepeewwe> ecu meeEe mcewuewewwee eeeewmil.m.u eweew 172 oooo.~ mpmm.l omwo.l memo. wmmo.u pmmo.i mmoo. N >553: va eeee._ Neee. weee.- eeeN.- ewee.- eeNe.- _ essee weV eeee.w eNNN. weee. weee. Neee. eeeeee eweeez weV eeee.w eeew. eNNN. eewN. eeweeeeee weV eeee.w wNee. wNee. weee4 weV eeee._ ewee. eee4 wNV eeee._ eeeeee wwV wwV weV weV weV weV wNV wwV eweewwee .e.e eweew ew eewuecweeee meeeeucwee wew :ewueeew eeeeeweme we meweewwe> esp meeEe meewuewewwee eweewmul.e.o eweew APPENDIX D FARMING SYSTEMS RESEARCH 173 APPENDIX D FARMING SYSTEMS RESEARCH In this study, especially in Chapter III, the objective was to relate all components of the farming system in order to verify the process of income determination in the study area. As defined by Dillon et al., A farming system (or farm system or whole-farm system) is not simply a collection of crops and animals to which one can apply this input or that and expect immediate results. Rather, it is a complicated interwoven mesh of soils, plants, animals, implements, workers, other inputs and environmental influences with the strands held and manipulated by a person called farmer who, given his preferences and aspirations, attempts to produce output from the inputs and technology available to him. It is the farmer's unique understanding of his immediate environment, both natural and socioeconomic, that results in his farming system. The conceptual framework used in this research is in some way related to a schematic representation of some determinants of the farming system presented by Norman2 (Figure 0.1). According to Norman, the total environment can be divided into two elements: tech- nical and human. Technical elements include the actual and potential livestock and crop enterprises, physical and biological factors that have been modified by man--often through technology development. Two types of factors characterize the human element: exogenous and endogenous. The exogenous factors are outside the control of the 1John L. Dillon, Donald L. Plucknett, and Guy J. Vallaeyes, "Farming Systems Research and International Agricultural Research Centers" (Armidale: The University of New England, Sept. 1978), p. 8. (Mimeographed.) 2Norman, op. cit., pp. 3-5. 174 175 individual farmer, and they include (a) community structures, norms, and beliefs; (b) external institutions, such as extension services, institutional credit, and price policies; and (c) miscellaneous influ- ences, such as population density and location. The endogenous fac- tors, on the other hand, are controlled by the farmer himself, "who ultimately decides on the farming system that will emerge, given the constraints imposed by the technical element and exogenous factors."1 In Chapter III some of the determinants of the farming system of the study area are analyzed, especially interaction of human and technical elements determining levels of income of five different groups of farmers. Ibid., p. 3. 176 .Eepmem mewewew we mewemew ucemeweew meeww :exewm .ceewez EewwV seemem mewewew en» we mueeewewepee eeem we eewpeueemeweew eweeeenemul.w.o eweeww I‘N h‘---u— d V 33334 .833 2.2!: V 711 f? '- neewu _ _ 114 .Av .Q ..ku .Qo stew.wwo V “cg-.5. w 4 4 Lee... .338 2.3 L V w .‘—-_—----_----_----_-_J \ 261.455.... / _ 23.83:. .833. +4/ \» 26:55 _ Sage—u weuwczuew LOZaO J 2.33335 4 03m cafe: 1 1' nvcv>em .1 AER: " " .mvwoae: . . Cva—Ua guc~ 000000000000 L — IIIIIIIIIIIIIIIIIIIIII A a I . . r 1' 39.926: .11 ' 839.528 1 9:52 pecweuxm ~l 33 ea... 32:: _ meeeeoeeeu 4 use .mee2 4 .meweaeewem wewwe—Eco meeeeeexm w _ mucgopu 69:31 BIBLIOGRAPHY 177 BIBLIOGRAPHY Aigner, D. J., and Chu, S. F. "0n Estimating the Industry Production Function." The American Economic Review 58 (1968): 1-10. Berry, R. Albert, and Cline, William. Agrarian Structure and Produc- tivity in Developing Countries. Baltimore: Johns Hopkins University Press, 1979. Bessel, J. "Measurement of Human Factor in Farm Management.“ Inter- national Journal of Agrarian Affairs Supplement (July 1969): 37-44. Brown, Maxuell. Farm Budgets--From Farm Income Analysis to Agricul- tural Project Analysis. World Bank Occasional Paper No. 29. Baltimore: Johns Hopkins University Press, 1977. Chenery, Hollins B. "Poverty and Progress--Choices for the Developing World." Finance and Development 17 (June 1980): 11-16. Chow, Gregory C. "Tests of Equality Between Sets of Coefficients in Two Linear Regressions." Econometrica 28(3) (1960): 591-605. Dias, Guilherme Leite da Silva. "Pobreza Rural no Brasil: Carac- terizacfio do Problema e Recomendacbes de Politica." Colegfio Analise e Pesquisa l6 (Agosto 1979). Brasilia: Comissao de Financiamento da Producao. Dillon, John L.; Plucknett, Donald L., and Vallaeys, Guy J. "Farming Systems Research at the International Research Centers." Armidale: The University of New England, 1978. (Mimeographed.) Empresa de Pesquisa Agropecuaria de Minas Gerais. "A Cultura de Arroz em Minas Gerais." Informe Agropecuario 5(5) (Julho 1979 : 9. Farrell, M. J. "The Measurement of a Productive Efficiency." Journal of the Royal Statistical Society, Series A, General, Part III, 120 (1957): 252-81. Fields, Gary S. "Who Benefits From Economic Development? A Reexamina- tion of Brazilian Growth in the 1960's." The American Economic Review 64(4) (September 1977): 570-82. 178 179 FundacSo Instituto Brasileiro de Geografia e Estatistica. Anuario Estatistico do Brazil (Rio de Janeiro), several issues. . Censos Econ6micos de l975--Censo Agropecuario de Minas Gerais. Secretaria de Planajamento da PresidGncia da Repdb- lica, 1975. Garcia, Joao Carlos. "Anélise de Alocacfio de Recursos por Proprie- térios e Parceiros em Areas de Agricultura de Subsistencia." M.S. thesis, Universidade Federal de Vicosa, Imprensa Uni- versitaria, 1975. Graber, Kenneth L. "Factors Explaining Farm Production and Family Earnings of Small Farmers in Brazil." Ph.D. dissertation, Purdue University, 1976. Griliches, Z. "Specification Bias in Estimates of Production Func- tions." Journal of Farm Economics 29 (1957): 8-20. Hall, M., and Winsten, C. "The Ambiguous Notion of Efficiency." Economic Journal 14 (March 1959): 71-86. Heady, Earl 0. "Technical Considerations in Estimating Production Functions." In Resource Productivity, Returns to Scale, and Farm Size. Edited by Earl 0. Heady, Glenn L. Johnson, and Lowell S. Hardin. Ames: The Iowa State College Press, 1956. Heady, Earl 0., and Dillon, John L. Agricultural Production Functions. Ames: Iowa State University Press, 1966. Herbst, J. H. Farm Management--Principles, Budgets, Plans. 4th ed. Champaign, 111.: Stipes Publishing Co., 1976. Hoch, I. "Simultaneous Equations Bias in the Context of the Cobb- Douglas Production Function." Econometrica 26(4) (October 1958): 566-78. "Estimation of Production Function Parameters Combining Time-Series and Cross-section Data." Econometrica 30(1) (January 1962): 34-53. International Monetary Fund. "International Financial Statistical." Various issues. Johnston, J. Econometric Models. New York: McGraw-Hill Book Co., 1972. King, Robert P., and Byerlee, Derek. "Income Distribution, Consump- tion Patterns and Consumption Linkages in Rural Sierra Leone." African Rural Economy Paper No. 16. East Lansing: Michigan State University, 1977. 180 Maddala, G. S. Econometrics. New York: McGraw-Hill Book Co., 1977. Marshak, J., and Andrews, W. "Random Simultaneous Equations and the Theory of Production." Econometrica 12 (July 1944): 143-205. Massell, Benton F. "Elimination of Management Bias From Production Functions Fitted to Cross-Section Data: A Model and an Appli- cation to African Agriculture." Econometrica 35(3-4) (July- October 1967): 495-508. Massell, Benton F., and Johnson, R. W. M. "Economics of Smallholder Farming in Rhodesia--A Cross-Section Analysis of Two Areas." Food Research Institute Studies in Agricultural Economics, Trade and Development, vol. 8 Supplement, 1968. Morais, Orlando Peixoto; Antunes, Fernando Zinho, and Soares, Plinio Cesar. "Exigéncias Climaticas da Cultura do Arroz." Informe Agropecuario 5(55) (Julho 1979): 16-19. Morgan, James. "The Anatomy of Income Distribution." Review of Economics and Statistics 44 (1962): 270-83. Mundlak, Yair. "Empirical Production Function Free of Management Bias." Journal of Farm Economics 43(1) (February 1961): 44-56. Mundlak, Y., and Hoch, 1. "Consequences of Alternative Specifications in Estimation of Cobb-Douglas Production Functions." Econo- metrica 33 (October 1965): 814-28. Nerlove, M. Estimation and Identification of Cobb-Douglas Production Functions. Chicago: Rand McNally, 1965. Norman, David W. "The Farming Systems Approach: Relevancy for the Small Farmer." Michigan State University Rural Development Paper No. 5. East Lansing: Michigan State University, 1980. Norman, David H., Pryor, David H., and Gibbs, Christopher, J. N. "Technical Change and the Small Farmer in Housaland, Northern Nigeria." African Rural Economy Program Paper No. 21. East Lansing: Michigan State University, 1979. Paniago, Euter; Ribon, Miguel; Silva, SebastiSo M. Ferreira da; and Teixeira Filho, Ant6nio Raphael. Estudos Sobre Uma Regifio Agricola: Zona da Mata de Minas Gerais (II), Instituto de Planejamento Econ0mico e Social--INPES. Monografia No. 11. Rio de Janeiro, 1973. Park, R. E. “Estimation With Heteroscedastic Error Terms." Econo- metrica 34(4) (October 1966): 888. 181 Ramsey, J. B. "Tests for Specification Errors in Classical Linear Least-Squares Regression Analysis." Journal of Royal Statis- tical Society, Series B, 31(2) (1969): 350-71. Ramsey, James B., and Schmidt, Peter. "Some Further Results on the Use of OLS and BLUS Residuals in Specification Error Tests." Journal of American Statistical Association 71(354) (June 1976): 389-90. Riemenschneider, Charles H. "The Use of the Gini Ratio in Measuring Distributional Impacts." M. 5. research report, Michigan State University, 1976. Schmidt, Peter. Econometrics. New York and Basel: Marcel Dekker, Inc., 1976. Schmidt, Peter, and Lovell, C. A. Knox. "Estimating Technician and Allocative Inefficiency Relative to Stochastic Production and Cost Frontiers." Journal of Econometrics 9 (1979): 343-66. Silva, Carlos Arthur B. da. "Factors Affecting Enterprise Choice: An Analysis of Traditional Food Production in Southeastern Minas Gerais, Brazil." Ph.D. dissertation, Michigan State Univer- sity, 1981. Souza, Antonio Fagundes de. "Pesquisa, Assisténcia Técnica e Extensfio Rural." In 0 Homem e o Campo. Fundacfio Milton Campos. BraSllia: Senado Federal--Centro Gréfico, 1976. Taveira, Tacito Claudio Andrade. "Anélise de Localizacfio da Producfio Agricola em Relacfio ao Mercado de Juiz de Fora Minas Gerais." M.S. thesis, Universidade Federal de Vicosa, 1976. Teixeira, Teot6nio Dias. "Resource Efficiency and the Market for Family Labor: Small Farms in the Sertao of Northeast Brazil." Ph.D. dissertation, Purdue University, 1976. Tench, Andrew B. Socio-economic Factors Influencing Agricultural Output. Sozialokonomische Schrifften zur Agrarentwicklung-- Heft 12. Saarbrucken, 1975. Theil, H. "The Analysis of Disturbances in Regression Analysis." Journal of the American Statistical Association 60 (December 1975): 1067-79. Timmer, C. Peter. “On Measuring Technical Efficiency." Food Research Institute Studies (Stanford University) 9(2) (1970): 99-171. Universidade Federal de Vicosa. Diagnostico Econ0mico da Zona da Mata. Vicosa: Imprensa Universitfiria, 1971. 182 Universidade Federal de Vicosa, Departamento de Economia Rural. Programa Integrado de Desenvolvimento da Zona da Mata--MG, Primeiro Relatério Anual de Avaliacfio. Vicosa, Marco 1979. Upton, Martin. "The Influence of Management on Farm Production on a Sample of Nigerian Farms." Farm Economist 11(12) (1970): 526-36. Vieira, Clibas. "Cultivo Consorciado de Milho com Feion." Informe Agrppecuario 4(46) (Outubro 1978): 42-45. Walters, A. A. "Production and Cost Functions: An Econometric Survey." Econometrica 31(1-2) (January-April 1963): 1-66. Wharton, Clifton R., Jr. "Subsistence Agriculture: Concepts and Scope." In Subsistence Agriculture and Economic Development. Edited by Clifton R. Wharton, Jr. Chicago: Aldine Publishing Co., 1969. World Bank. "Brazil--Staff Project Report of the Integrated Rural Development Project in the State of Minas Gerais." Report No. 1291 Br. Washington, 1976. (Mimeographed.) Yotopoulos, Pan A. "On the Efficiency of Resource Utilization in Subsistence Agriculture." Food Research Institute Studies (Stanford University) 7(2) (1968): 125-35. Zellner, A.; Kmenta, J.; and Dréze, J. "Specification and Estimation of Cobb-Douglas Production Function Models." Econometrica 34 (October 1966): 784-95. a—‘h-T-‘x a_.> .- 13")“