Mexico - Ejido reform : avenues of adjustment five years later (Vol. 1 of 2) : Main report

Document ot The World Bank Report No. 18897 MEXICO EJIDO REFORM AVENUES OF ADJUSTMENT - FIVE YEARS LATER MAIN REPORT DECISION DRAFT July 1, 1999 Environmentally and Socially Sustainable Development Sector Management Unit Mexico Country Management Unit Latin America and the Caribbean Region This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. CURRENCY EQUIVALENTS Currency Unit = Mexican New Peso $1.0 = N$ FISCAL YEAR January 1 to December 31 WEIGHTS MEASURES Metric System MAIN ABBREVIATIONS AND ACRONYMS BANRURAL National Rural Bank ECLA Economic Commission for Latin America GDP Gross Domestic Product GOM Govemment of Mexico INDA National Institute for Agrarian Reform (Instituto Nacional de Desarrollo Agrario) NRE National Rainfed Equivalent NPR Nominal Protection Rate NAFTA North American Free Trade Agreement PROCEDE Program to Title Ejiditario Rights and Parcels (Programa de Certificaci6n de Derechos Ejidales y Titulos de Solares Urbanos) PROCAMPO Direct Income Support Program for Producers RAN National Agrarian Registry (Registro Agrario Nacional) SAGAR Ministry of Agriculture, Livestock and Rural Development (Secretaria de Agricultura, Ganaderia y Desarrollo Rural -formerly SARH) SARH Ministry of Agriculture and Water Resources (Secretaria de Agricultura y Recursos Hydraulicos) SRA Ministry of Agrarian Reform (Secretaria de Reforma Agraria) Vice President Shahid Javed Burki Country Director Olivier Lafourcade Acting Sector Director John Redwood Sector Leader Adolfo Brizzi Task Team Leader Louise Cord The following individuals contributed to the preparation of this study: Messrs./Mmes. Louise Cord (task manager); Benjamin Davis (agriculture, capital markets, income and migration analysis); Pete Lanjouw (non- agricultural activities and poverty); Pedro Olinto (land markets); Elizabeth Katz, Ana Lucia Moreno, Maria Correia (gender); Estanislao Gacitua- Mario, Carlos Munoz, Quentin Wodon, Fernando Rello, Juana Sandoval, Josefina de la Soledad (social capital and case studies); Louise Cord, Quentin Wodon (government programs) Nathalie Olsen, Gladys Lopez-Acevedo, Alberto Valdes (incentive framework); Claudia Binder, Ronna Montgomery, Emma Calderon, Todd Diehl (research assistance); Patricia Eyenga (production assistance). The team gratefully acknowledges the advice and support received from Alberto Valdes (World Bank), Alain de Janvry and Elisabeth Sadoulet (University of California, Berkeley). Financial assistance was received from the Mexico Country Department, DEC and LCSPR, as well as the Finnish and Spanish trust funds. Peer reviewers are John Nash and John Heath. Finally, the team expresses its deepest appreciation to the staff of the Secretariat of Agrarian Reforn, and in particular, Lic. Leopoldo Zorilla Omelas, Lic. Hector Robles Berlanga, Lic. Marco Antonio del Castillo, Lic. Laura Munoz Rojas and Lic. Leticia Manzanera Herera y Cairo, for their assistance and support. 111 MEXICO EJIDO REFORM AVENUES OF ADJUSTMENT - FIVE YEARS LATER Table of Contents Executive Summary .................................................... vii Introduction ....................................................................................................................................1 Institutional and Economic Reforms: 1988-1994....................................................................3 The Second Agrarian Reform and Its Antecedents ........................3 Reform in an Unfavorable Economic Environment..............6..........6 Distinguishing the Impact of the Reforms ........................ ...............11 Avenues of Adjustment: 1994-1997 ...................................................................................... 11 Increasing Labor Market Participation and Migration.............l..........1] Decreasing Dependence on Local Capital Markets ............................16 Increasing Diversity in Agricultural Strategies...... .........................19 Deepening Ejido Land Markets ................... .................22 Dynamic Ejidos Facilitate Adjustment .....................................26 PROCAMPO and ALIANZA - A New Generation of Government Programs...... ........29 Regional Forces of Adjustment ...........................................35 Implications for Rural Poverty ............................................................................................. 37 Factors Associated with Poverty in the Ejido Sector.......................37 Emerging from Poverty........................................... .....41 Signs ofIncreasing Integration...........................................43 Gains from Integration ....................... ..................45 Easing the Transition Process: What is the Government's Role?......................................46 Bibliography ................................................................................................................................. 53 Annex 1: Chronology ................................................................................................................... 55 A nnex 2: D ata Sources................................................................................................................. 56 iv List of Tables Table 1 Evolution of Real Producer Prices in Mexico 9 Table 2 Nominal Protection Rate 1992-97 9 Table 3 Annual Changes in Acreage of Key Crops 10 Table 4 Participation in Off-farm Activities 12 Table 5 Estimate of Non-Agricultural Earnings Per Individual 13 Table 6 Household Temporary Migration Rates 15 Table 7 Percentage of Households with Credit by Year, Land Size and Source 17 Table 8 Ejido Household Characteristics 19 Table 9 Household Use of Inputs and Sources of Procurement 20 Table 10 Currents of Change in Ejido Agriculture 21 Table 11 Household Participation in Rental Markets in 1994 and 1997 22 Table 12 Participation in Land Rental Markets by Farm Size and Region, 1997 23 Table 13 Breakdown of Land Ownership Transactions by Category and Year 24 Table 14 Maximum Likelihood Estimation of the "Friction" Model for Land Rental Markets 25 Table 15 National Distribution of PROCAMPO by Region and Size of Producer (1997) 31 Table 16 Perceived Land Use Requirements for PROCAMPO 31 Table 17 Household Characteristics by Region 36 Table 18 Correlates of Poverty in 1997 39 Table 19 Regression of Changes in Household Income Ranking 43 Table 20 Sources of Income by Farm Size 45 List of Figures Figure 1 Contribution of PROCAMPO to Household Income by Decile 32 Figure 2 Sources of Income 44 List of Boxes Box 1 Productivity in the Ejido Sector 4 Box 2 Migration and Gender 16 Box 3 Participation in Non-Ejido Organizations 29 Box 4 Indigenous Households In Mexico's Ejido Sector 41 Box 5 Voucher Program in Romania 50 MAP Mexico, Regional Breakdown (IBRD No. 29701) List of Background Papers in Volume II 1. The Evolution of the Incentive Framework for Mexican Agriculture: 1992-1997, Nathalie Olsen, FAO/CP 2. Poverty and the Non-Farm Economy in Mexico's Ejidos: 1994-1997, Peter Lanjouw, World Bank 3. The Impact of the Ejido Reforms on Land Markets in Mexico, Pedro Olinto, World Bank 4. Social Capital and Land Titling in Mexico's Ejido Sector, Louise Cord, Estanislao Gacitua- Mario and Quentin Wodon, World Bank. 5. Gender and Ejido Reform, Elizabeth Katz, Consultant V 6. Do Mexico's Agricultural Programs Alleviate Poverty? Evidence from the Ejido Sector, Louise Cord and Quentin Wodon, World Bank Appendices available upon request: Ejido: Case Studies, Estanislao Gacitua-Mario, LCSES The Adjustment Strategies of Mexican Ejidatarios in the Face of Neoliberal Reform, Benjamin Davis, IFPRI Appendix Tables: Compendium of Descriptive Statistics on 1994-97 Ejido Survey vi Mexico Ejido Reform Avenues of Adjustment - Five Years Later Executive Summary "The characteristics of the ejido sector on the eve of the 21St century are greatly different from those of its beginnings over 60 years ago. While the importance of community and an attachment to the land remain, the exclusively agricultural economy of the 1930s has given way to a tremendously diversified economy with a range of household income generation strategies. Today, ejidatarios and their children are just as likely to be working in off-farm wage activities, such as the maquila, or in a restaurant in Los Angeles, as growing corn and beans on the family plot." Davis, 1999 1. During the past twelve years, Mexico has witnessed a major transformation of its economic panorama. Since 1985, the Government of Mexico (GOM) has implemented sweeping reforms that have led to the emergence of an economic system based on market orientation, openness and competition. For the agricultural sector, the new economic paradigm has brought dramatic changes, including trade and price liberalization and the privatization of most of the sector's input and output parastatals. 2. To allow the ejido sector to adjust to economic liberalization, the Government initiated a bold program of agrarian reform in 1992. The ejido sector had grown out of the Mexican Constitution of 1917, which embodied a strong program of land reform. Under this program, the Government granted land and water resources to communities of producers (known as ejidos). The community's members, or ejidatarios, had usufruct rights to the land contingent on occupation and cultivation - prohibiting even the hiring of labor in the process. Under the ejido system, land could not be alienated, rented or mortgaged and absences from the ejido of more than two years led to the loss of land rights. By the early 1990s, the ejido sector accounted for approximately half of Mexico's farmland and three quarter's of the nation's producers. However, also by this time, the sector had become obsolete and characterized by productive inflexibility and increasing non-compliance with the sector's legal framework. If agriculture was to adjust to an open economy, reform of the ejido sector was considered critical and in 1992, the second agrarian reform eliminated more than half a century of restrictions for the ejido system. 3. The purpose of this paper is to analyze how ejido households are adjusting to the institutional and economic reforms of the 1990s and the implications of their adjustment strategies on the emergence from poverty. The principal questions to be addressed are: * What were the specific changes in the institutional and economic context of the ejido sector and what was their rationale? Section II illustrates the increasing ineffectiveness and inefficiency of the ejido system that prompted the reform of Article 27, and highlights the unfavorable and uncertain economic environment in which the institutional changes were unfolded. Viii * What are the principal avenues of adjustment used by the ejidatarios, and which households have participated in them? Section III examines the participation of ejido households in labor, land and capital markets as well as agricultural activities, and analyzes how individual, household, community and regional factors molded this participation. * What are the key factors associated with poverty and the emergence from poverty in the ejido sector? Section IV assesses the principal features associated with poverty in the ejido sector in 1994 and 1997 and identifies the key characteristics linked to emerging from poverty over the course of this period. * What steps can be taken to ease the ejido sector's transition and what is the role of Government? This final section presents recommendations on how Government can ease the transition process. 4. Data sources. The analysis draws upon three national, ejido household surveys, which were carried-out in 1990, 1994 and 1997 (see Annex 2). The household surveys examine participation in agriculture, non-farm activities and factor markets (the latter two are covered only in 1994 and 1997). A panel of 1287 households was created in 1997 to allow for an intensive analysis of the adjustment process since 1994. The panel controls for many unobservable variables that are not adequately reflected in cross-sectional analysis. In addition, two community surveys were carried out in 1994 and 1997 focussing on the structure and operation of the ejido community. Finally, to complement the quantitative data, six qualitative case studies were carried out in 1998, which focussed specifically on government programs, gender differentiation, natural resource management issues, land use and community organization. 5. It is beyond the scope of this study to assess the aggregate impact of the economic and institutional reforms on poverty and productivity in the ejido sector. The adjustment process is still in the early phases as most of the reforms were introduced after 1992. In addition, aggregate information on the evolution of agricultural production and rural poverty in the ejido sector is not available. Moreover, this study did not have access to comparable information on farm households outside of the ejido system. Without this information, it was not possible to distinguish between the impact of the economic reforms from the impact of the institutional reforms on the ejido household. 6. The principal conclusions of the study are: * The institutional and economic reforms have led to the increasing integration of the ejido with the non-farm sector as well as with the non-ejido farm sector. A relatively unattractive economic environment for agriculture, coupled with the removal of restrictions on land and labor market participation under the agrarian reform, has pushed ejido households to become involved in the non-farm sector. These growing connections with the non-farm and non-ejido economies will ultimately erode differential returns to land, labor and capital across the sectors and reduce rural poverty in Mexico. However, the realization of the benefits from integration will depend upon the level of development lx of factor markets and the extent to which all ejido households have good access to these markets. Participation in off-farm activities and PROCAMPO, residence in a dynamic ejido and an intensification of agricultural activities all offer avenues for ejido households to emerge from poverty. The analysis also suggests that how a household uses its land has become more important than the amount of land it uses in emerging from poverty. Also, participation in off-farm activities is not a guaranteed route to exit poverty in the ejido sector, as the returns from different types of off-farm activities vary significantly (Table 18 and Table 19). Main Findings and Recommendations by Topic 7. Incentive Framework for Agriculture. The legal and economic reforms were carried out in a period of economic adversity, characterized by trade and price liberalization, stagnant sectoral growth, turbulent macro economic conditions, high interest rates, and uncompetitive domestic marketing channels. This unfavorable environment is reflected in the declining real prices and the negative nominal protection coefficients for key commodities in Mexico's agricultural sector in the 1990s (Tables 1 and 2). 8. Increasing participation in off-farm activities and migration. Participation in off-farm activities (wage and self-employment activities) has expanded sharply in the ejido sector since 1994 for both men and women (Table 4). In 1997, 60 percent of the households participated in off-farm activities and they provided (including migration remittances) an average of 44 percent of household income. The contribution of off-farm income and remittances was relatively constant across all income deciles; illustrating that both the poor and the non-poor are involved in the off-farm sector (Figure 2). 9. The returns to different types of off-farm labor vary significantly. This suggests non-poor households participate in activities with the highest returns, such as self-employment and commerce related wage activities, while the poor participate in activities with the lowest returns such as casual non-agricultural and manufacturing wage activities. While women participate less than men do in off-farm activities, their participation is increasing. Preliminary evidence suggests that women may earn lower than similarly qualified men and that it is more difficult for a woman to acquire off-farm employment. 10. Despite the variability of the returns to non-farm activities, they are clearly associated with lower poverty and this relationship grew stronger in the ejido in 1997 compared to 1994. Moreover residents of ejidos where the members increased their involvement in off-farm activities since 1994 were more likely to emerge from poverty (Table 18 and Table 19). 11. International migration to the U.S. has also increased, while domestic migration appears on the decline, suggesting that the devaluation made U.S. real wages more attractive than Mexican wages (Table 6). U.S. migration is important, as family members who migrate to the U.S. are more likely to send remittances than Mexico migrants. International migration assets were x associated with lower poverty in both 1994 and 1997. On the negative side, households (often headed by women) left behind in ejidos where there is a lot of migration may have less access to social services and basic infrastructure than other households. 12. Government could undertake several initiatives to enhance prospects for ejido households to participate in off-farm employment and migration. The main priority areas for action are: * ensuring the availability of quality education to all ejidatarios and their families, as education is strongly correlated with access to non-agricultural employment and income, as well as with lower poverty -in addition to having an intrinsic value as a tool of empowerment (Table 5 and Table 18 ); * providing key infrastructure packages (roads, electricity, communications facilities, etc) to improve the quality of access to urban centers and to encourage firms, banks and other facilities to locate in rural areas (Binswanger, Khnadker and Rosenzweig, 1986); * offering tariff rate subsidies to private providers of infrastructure and services (roads, communications, etc.) to encourage them to deliver services and invest in low income areas where the high unit costs and low ability to pay often lead to under-investment; these subsidies should be well-targeted towards low income households so as to limit their fiscal burden; * evaluating the pilot programs that provide information on regional employment possibilities in low income isolated areas; and * expanding efforts to provide migrants with social services (mobile education and health care, child care, etc.) to ensure that the next generation has the appropriate skills to participate fully in labor markets, especially given the tendency of women to migrate for longer periods of time and with their children. 13. Developing Rural Financial Markets. Formal and informal rural financial markets in Mexico continue to remain shallow. Only 20 percent of the households received credit of any kind in 1997 compared to 30 percent in 1994 (Table 7). The decline in credit reflects the reduction of the Government's PRONASOL credito a la palabra program, as access to informal credit actually increased slightly over the period. Households with a greater involvement in agriculture, higher participation rates in Alianza and PROCAMPO, and with residence in the Pacific North and the Gulf region were more likely to receive credit in 1997. In contrast, households increasing their participation in off-farm activities and migration were less likely to have credit in 1997. 14. The low rates of participation in financial markets may be explained by several factors working to depress both the supply of and the demand for credit. One of these factors, unique to the ejido sector, is the fact that the PROCEDE title does not give the ejidatario the right to mortgage his or her land or sell out it outside of the ejido, without the approval of the ejido assembly. As a result, financial institutions may be less willing to lend to an ejidatario than to a private farmer with a full title. However, access to legal collateral is an important necessary condition for efficient rural financial markets, but it is not a sufficient condition. In the case of rural Mexico, many of the sufficient conditions are not fully in place to allow financial institutions X1 use collateral instruments effectively. On the demand side, many for efficient financial markets high transactions costs and interest rates depress the demand for financial resources. 15. While the increasing integration of the ejido economy with the non-farm sector has provided alternative capital flows to the ejido sector, weak financial markets remain a constraint for the household to move into more lucrative non-farm areas such as self-employment activities and modernize agriculture. In other countries, non-formal financial institutions have often filled the gap with services more suited to the needs of rural entrepreneurs. These services are not limited to agricultural activities and offer savings as well as credit instruments. To deepen access to credit and savings instruments in rural Mexico, constraints to the development of non-formal financial markets (trader credit, savings and loans, credit unions, non-government organizations, etc.) should be explored and addressed. 16. Agriculture: Diversification and Specialization. Overall, agricultural strategies experienced little change in the ejido sector since 1994, with three exceptions: (a) an increase in input use, with a recovery to the levels of 1990; (b) an increase in irrigated corn, and to a lesser extent, basic grains; and (c) an increase in land holdings, reflecting mainly (i) the accumulation of large tracts of land by a few households in the Gulf; (ii) the transfer of some communal land to individual households; and (iii) land quality improvements (Table 8). 17. These relatively mild aggregate changes in agricultural strategies mask underlying shifts of resources in, out and within the sector. These changes, largely regional specific, illustrate an increasingly heterogeneous agricultural sector within the ejido, as: (a) some ejidos modernize their production -mainly in the Gulf and North Pacific; (b) others continue to cultivate, but at a reduced intensity and look increasingly outside of agriculture for their income - mainly in the North and Center; and (c) yet another, larger group, scattered throughout the country, but highly concentrated in the Pacific South, remain as relatively small producers heavily dependent upon subsistence agriculture. 18. To support the overall agricultural adjustment process, Government must recognize the sector's increasing diversity. In particular, priority objectives for the Government to consider are: (a) ensuring that government policies do not impede the adjustment process and the shift of some households out of agriculture; (b) supporting low income households with productive potential in agriculture to invest in and raise their agricultural productivity; and (c) improving the efficiency and competitiveness of Mexico's agricultural markets. 19. To implement these objectives, three sets of recommendations are offered on PROCAMPO, Alianza and product markets. PROCAMPO is a generally well-conceived program. It seeks to provide producers with income support that is non-distortionary and allows households to decide whether to intensify or reduce their involvement in agriculture. The outreach of PROCAMPO in the ejido sector is high (84 percent of all households receive PROCAMPO). On average, PROCAMPO payments provide about 8 percent of household income in the sector, although the program's contribution may reach as high as 40 percent for low-income families (Figure 1). 20. As to be expected with an income transfer program, the analysis reveals that PROCAMPO has a strong positive impact on poverty reduction in the ejido sector. Moreover, although further research is necessary, the analysis suggests a multiplier effect in that PROCAMPO's income effect is greater than a one to one transfer of Pesos, implying that the program has a positive Xii impact on the household's investment behavior as well as on the local economy and employment opportunities. 21. To enhance participation in the program, as well as its apparent multiplier effect on income, the principal recommendations are to: * distribute the payment earlier in the crop cycle or at least announcing the amount of payment prior to planting (even if the actual payment is made after planting, this will add some certainty to the producers income stream); * allow ejidatarios to designate half beneficiaries - either banks or traders that could receive half of their PROCAMPO payment, while they retain the other half for consumption purposes or emergency use; * eliminate planting and land use requirements, which would lower administrative costs, add certainty to the producers' income stream (they will not have to requalify each year) and allow for earlier payment of PROCAMPO. If this does not seem viable, the planting and land use requirements could be disseminated more widely in the rural sector to increase the likelihood that banks and local traders would accept PROCAMPO payment as collateral; * increase the transparency of the certification process and make the list of eligible producers open to the public at the local level, which will also add certainty to households' stream of PROCAMPO payments and facilitate long-term investments; and * provide vouchers or some other form of financial support for technical assistance to ejidatarios with small plots to assist and encourage these producers use their PROCAMPO payments to invest in agriculture. 22. Alianza Para el Campo aims to promote agricultural investment and modernization, which as noted above is an important avenue of poverty reduction. The program began in 1996 and by the time of the survey in 1997 the target population only had access to the program for one, or at most two, crop seasons. In 1997, 11 percent of ejido households were participating in the program. This relatively low participation rate may be explained by: (a) the relatively recent introduction of the program; (b) a lack of program dissemination in the ejido sector by 1997; (c) counterpart resource constraints on the part of the ejido households; and (d) two structural aspects, which may lower the impact of the program and hence its demand among ejidatarios. First, some subprograms called for group membership, which the case studies suggest may have raised, in some instances, the transaction costs of receiving Alianza resources and may have undermined the likelihood that producers will be able to participate in the programs of most interest to them. Second, the fact that government officials, in some cases, may procure the inputs (as opposed to the ejidatarios) has led to some concerns that inputs received under the program may not meet the full specifications of the ejidatarios and that the program could be undermining the development of local retail outlets. 23. Several modifications in program design are suggested to raise the benefits households derive from Alianza and thereby increase the demand for the program in the ejido sector: * promote the purchasing of Alianza inputs directly by the ejidatarios from local distributors, in order to facilitate the development of local private sector retail outlets; Xiii * provide program beneficiaries with vouchers that they could redeem at any participating input supplier for the input of their choice. Voucher programs are widely supported by farmers in Eastern Europe because they are more flexible and allow the producer more freedom in deciding when and how to use the subsidy (Box 5); * increase the dissemination of information on the program in the ejido sector so that a larger proportion of ejidatarios are aware of the program, understand its objectives and are clear on how to access the resources; * carefully assess the impact of group participation on the various subprograms where it is applied to ensure that it is an effective mechanism to channel resources and does not raise transaction costs of program participation; and, * increase the sensitivity of the program staff and design to ensure that women are offered the opportunity to participate in and benefit from the program. 24. The analysis of the incentive framework revealed high levels of implicit taxation for agricultural products (Table 2). The significant price wedge between Mexican producer prices and equivalent world market prices revealed by the NPR analysis is highly suggestive of uncompetitive agricultural markets Mexico. The lower producer prices, brought on in part by the market inefficiencies, undermine incentives for agriculture as well as and the ability of markets to provide effective resource allocation decisions. Improving the competitiveness of agricultural markets is a key area for action, and in particular Government should consider: * Exploring the causes behind the noncompetitive markets. Do they reflect the influence of a monopolistic market structure for farm markets? Do the costs reflect high transport costs due to local monopolies and/or regulation, or do they reflect inefficiencies and distortions along the marketing channel. What is the role played by expensive storage costs due to the high interest rates, a scarcity of rural and wholesale infrastructure, and the poor regulatory and financial framework for storage? 25. Improving the Sustainability and Growth of Ejido Land Markets. Since 1994, the frequency of land market transactions in the ejido sector has increased. About 20 percent of the households participated in rental transaction in 1997, a 22 percent increase over 1994 (Table 11). While the share of households engaging in ownership transactions remains small (less than 5 percent of all households in 1997), the number of transactions has risen gradually since 1992 (Table 13). Moreover, preliminary evidence suggests that transaction costs for land rentals have declined (Table 14). Both the increasing number of transactions and the decline in transaction costs indicate that the ejido land market has grown. This deepening of ejido land markets has facilitated the increasing integration of the ejido with the non-farm and non-ejido economies and is suggestive of a more efficient allocation of land both within and outside of the sector. 26. The role of PROCEDE in fostering the ejido land markets remains unclear. Ejidatarios with PROCEDE titles in 1997 engaged in more rentals in both 1994 and 1997 than ejidatarios without titles. Moreover, households with and without PROCEDE experienced similar percent increases in the frequency of land rentals between 1994 and 1997. The higher rate of land transactions amongst PROCEDE ejidatarios probably reflects the selection bias with which PROCEDE ejidos Xiv were selected (para. 19). Finally, the informal nature of many land market transactions implies that over time the official PROCEDE records and titles could become less reflective of actual land ownership patterns. 27. Government must address three challenges in this area during the next five years. First, it must complete PROCEDE in the remaining 41 percent of the ejidos (as of December 1997), which may pose more challenges than the first half of the titled ejidos, as the larger and more isolated ejidos and those with pending land issues were left for the end of the titling program. Second, the informal nature of most land ownership transactions could render the PROCEDE cadastre out of date in a few years time and could threaten the continued reduction in transaction costs (as boundaries and ownership rights are not properly recorded). Third, ensure that the program treats female and males equally, particularly in light of the growing role of women in agriculture. To address these three challenges, Government could consider: * assisting communities so that they can properly document internal and external land transactions at the ejido level and then transmit the relevant records to the RAN; moreover, the information on land transactions should be made public locally and nationally to ensure the transparency of the emerging ejido land market; * modifying certain aspects of PROCEDE to facilitate its implementation in the more remote regions and indigenous communities (South Pacific and North; and * initiating a comprehensive gender-training program for PROCEDE officials and beneficiaries, which would cover inheritance issues and the rights of female ejidatarios. 28. Learning the Lessons ofDynamic Ejidos and Regions. Although the direction of causality is not clear, the analysis indicates that dynamic ejidos are associated with lower levels of poverty and a higher share of the members emerging from poverty. Dynamic ejidos are those with high levels of ejido social capital, which reflects the extent to which the ejido has regular assemblies, norms and practices to manage common resources (e.g., pasture) as well as links to higher level umbrella organizations in the social sector. Ejidos with higher levels of social capital are also more likely to have higher access to basic social services, infrastructure and private sector contracts, as well as have a more favorable attitude towards PROCEDE. 29. The analysis also highlights that indigenous households in the Gulf region have increased their participation in off-farm activities, modernized their agriculture and been more successful than other households in emerging from poverty. 30. To understand the role of social capital and the factors driving the development process in indigenous households in the Gulf region since 1994, the following recommendations are provided: * Provide training and technical assistance to interested communities to assist them improve the management of common resources as well as the delivery of social services, while avoiding the temptation to provide them with direct financial resources, whi-h often creates distorted organizational incentives. * Analyze the role of ejido social capital in the development process as well as the effects of government programs such as PROCAMPO, PROCEDE, Alianza, and Solidarity Municipal Funds on ejido social capital. XV Analyze the recent development spurt in indigenous households in the Gulf region to better understand what were the key factors behind the increases in labor market participation and the modernization of agriculture experienced by many households in the region. The study also reveals that this region had significantly above average levels of access to credit, Alianza and PROCAMPO and the role of these programs in the development process, especially for indigenous households, should be clarified. xvi 1 Introduction 1. During the past twelve years, Mexico has witnessed a major transformation of its economic panorama. Since 1985, the Government of Mexico (GOM) has implemented sweeping reforms that have led to the emergence of an economic system based increasingly on market orientation, openness and competition. For the agricultural sector, the new economic paradigm has brought dramatic changes, including trade and price liberalization and the privatization of most of the sector's input and output parastatals. 2. To allow the ejido sector to adjust to economic liberalization, the Government initiated a bold program of agrarian reform in 1992. The ejido sector had grown out of the Mexican Constitution of 1917, which embodied a strong program of land reform. Under this program, the Government granted land and water resources to communities of producers (known as ejidos). The community's members, or ejidatarios, had usufruct rights to the land contingent on occupation and cultivation - prohibiting even the hiring of labor in the process. Under the ejido system, land could not be alienated, rented or mortgaged and absences from the ejido of more than two years led to the loss of land rights. By the early 1990s, the ejido sector accounted for approximately half of Mexico's farmland and three quarter's of the nation's producers. However, also by this time, the sector had become obsolete and characterized by productive inflexibility and increasing non-compliance with the sector's legal framework. If agriculture was to adjust to an open economy, reform of the ejido sector was considered critical and in 1992, the second agrarian reform eliminated more than half a century of restrictions for the ejido system. 3. The purpose of this paper is to analyze how ejido households are adjusting to the institutional and economic reforms of the 1990s and the implications of their adjustment strategies on their emergence from poverty. The principal questions to be addressed are: * What were the specific changes in the institutional and economic context of the ejido sector and what was their rationale? Section II illustrates the increasing ineffectiveness and inefficiency of the ejido system that prompted the reform of Article 27, presents the key elements of the legal reforms, and highlights the unfavorable and uncertain economic environment in which the institutional changes were unfolded. * What are the principal avenues of adjustment used by the ejidatarios, and which households have participated in them? Section III examines the participation of ejido households in labor, land and capital markets as well as agricultural activities, and analyzes how individual, household, community and regional factors have molded this participation. * What are the key factors associated with poverty and the emergence from poverty in the ejido sector? Section IV assesses the principal features associated with poverty in the 2 ejido sector in 1994 and 1997 and identifies the key characteristics linked to emerging from poverty over the course of this period. * What steps can be taken to ease the ejido sector's transition and what is the role of Government? This final section presents recommendations on how Government can ease the transition process. 4. Data sources. The analysis draws upon three national, ejido household surveys, which were carried out in 1990, 1994 and 1997 (Annex 2).2 The household surveys examine participation in agriculture, non-farm activities and factor markets (the latter two are covered only in 1994 and 1997). The 1990 survey covered approximately 35,090 households in 5,007 ejidos, while the 1994 and 1997 surveys reached 1,342 and 1,665 households from 275 and 286 ejidos, respectively. A panel approach was used in the 1997 survey to allow for an intensive analysis of the adjustment process since 1994. The panel controls for many unobservable variables that are not adequately reflected in cross-sectional analysis. The panel consists of 1,287 households that were surveyed in both 1994 and 1997. In addition, two community surveys were carried out in 1994 and 1997 with ejido authorities in 286 ejidos, focussing on the structure and operation of the ejido community. To complement the quantitative data, six qualitative case studies were carried out in 1998, which focussed specifically on government programs, gender differentiation, natural resource management issues, land use and community organization (Appendix, Gacitua-Mario). In addition, national and international data from SAGAR and the World Bank on price trends, marketing and other transactions costs is used to prepare the analysis of the incentive framework. 5. The principal conclusions of the study are: * The institutional and economic reforms have led to the increasing integration of the ejido with the non-farm sector as well as with the non-ejido farm sector. A relatively unattractive economic environment for agriculture, coupled with the removal of restrictions on land and labor market participation under the agrarian reform, has pushed ejido households to become involved in the non-farm sector. These growing connections with the non-farm and non-ejido economies will ultimately erode differential returns to land, labor and capital across the sectors and reduce rural poverty in Mexico. However, the realization of the benefits from integration will depend upon the level of development of factor markets and the extent to which all ejido households have good access to these markets. * Participation in off-farm activities, participation in PROCAMPO residence in a dynamic ejido and an intensification of agricultural activities all offer avenues for ejido households to emerge from poverty (Table 18 and Table 19). The latter suggests that how the household uses its land is more important than the amount of land used by an household in emerging from poverty. Fine Darticipation in off-farm activities is not a guaranteed route to exit poverty in the ej sector, as different off-farm activities offer very different returns. 2 The 1994 data have received extensive analysis to date (see de Janvry, Gordillo and Sadoulet, 1997 and Davis, 1997). 3 6. It is beyond the scope of this study to assess the aggregate impact of the economic and institutional reforms on poverty and productivity in the ejido sector. The adjustment process is still in the early phases as most of the reforms were introduced after 1992. In addition, aggregate information on the evolution of agricultural production and rural poverty in the ejido sector is not available. Moreover, this study did not have access to comparable information on farm households outside of the ejido system. Without this information, it is not possible to distinguish between the impact of the economic reforms from the impact of the institutional reforms on the ejido household. Institutional and Economic Reforms: 1988-1994 7. During the 1990s, the ejido sector faced two important and almost simultaneous reforms. First, the Second Agrarian Reform of 1992 officially ended the era of land reform and removed most restrictions against land transactions and participation in labor markets. Second, the economy wide and agricultural economic liberalization programs that began in the late 1980s were reinforced in 1992 with the passage of NAFTA. Both sets of reforms were unfolded into a difficult economic environment, characterized by falling real prices and economic turbulence, which would increase the challenges of adjustment. This section first evaluates the origins and the content of the Second Agrarian Reform and then describes the liberalization package and the overall evolution of the incentive framework for agriculture during the 1990s. The Second Agrarian Reform and Its Antecedents 8. The ejido system and the first agrarian reform originated early in this century following the Mexican revolution when the peasantry's dissatisfaction over lack of access to land rights was one of the main catalysts of the conflict. Following several decades of strong political support for the ejido system and significant land reform, the ejido system became a critical instrument to ensure political control in rural areas and to organize agricultural production. By the late 1980s, the system accounted for 30,000 ejidos with 3.2 million ejidatarios, about 18 percent of which were women (INEGI, 1991). The sector controlled the majority of the country's agricultural resources, with approximately half of Mexico's farm land, 70 percent of the nation's forests and three quarters of the nation's producers. The sector was responsible for more than 70 percent of the nation's corn production and 80 percent of bean production (de Janvry, et al, 1995). 9. The Ejido System Under Stress. Under the ejido system, the President of the Republic granted land and water resources not to individuals, but rather to a community or group of producers (known as the ejido). Each ejidatario, or comunero (in the case of indigenous agrarian communities), was given usufruct rights over a parcel of land, access to communal land, the right to a house plot and voting rights in the ejido assembly.3 2 Agrarian communities have similar rights to ejidos and hereinafter, all references to ejidos also refer to agrarian communities, unless specifically mentioned otherwise. 4 10. While the relationship between the ejido and the state has evolved over time, there have been two constant themes: first, the use of the ejido to assure political control in rural areas; and second, the use of the ejido as an instrument to organize and control agricultural production. The state utilized a number of mechanisms to control the rights and resources of ejidos (Yates, 1981; de Janvry, Gordillo and Sadoulet, 1997). Ejidos functioned under a strict legal code with detailed rights and obligations for ejidatarios, which limited their mobility and ability to diversify out of agriculture. For example, ejidatarios had to work the land themselves; they were prohibited from hiring wage labor; land could not be rented or sold; and absences from the ejido of more than two years led to the loss of land rights. 11. The state also sought to control what the ejidatarios produced as well as how they produced and marketed their output. Public credit was not provided to individual ejidatarios, but rather to the ejido as a whole. It was provided in-kind and the ejidatarios had to grow specific crops with specific inputs in order to receive the credit. Similarly, state enterprises controlled the marketing and production technologies not only of basic grains and oilseeds, but also of specialized crops heavily grown in the ejido sector, such as coffee, cocoa and sugar. The impact of the economic and institutional framework on the sector's productivity was unclear (Box 1). Box 1 Productivity in the Ejido Sector Much discourse has been dedicated to the question of whether ejido producers are more or less productive than farmers in the private sector. Both Dovring (1969) and Heath (1990) have shown that, controlling for farm size, few differences in land and labor productivity can be found between ejido and private producers. Since the 1970s, ejido peasants as a group have been the recipients of much state largesse and patronage, but this has come at the cost of productive inflexibility and political control. Further, it is unclear whether the community structure of the ejido has played a positive or negative role on productivity. On the positive side, the ejido facilitated access to credit, the management of common natural resources and the use of pooled labor resources, and offered economies of scale for the purchase of agricultural machinery. On the negative side, the ejido often was undermined by corruption and inefficiency, as local caciques and government agents used the legal structure to further their personal interests at the expense of the rest of the members of the ejido- 12. State control over the ejido was not absolute, however, and there was increasing non- compliance with the law. Informal and illegal markets emerged, particularly in the renting and selling of land, and in the use of ejido communal land. Transactions were often legalized under different guises by ejido and government authorities. Nonetheless, these markets often played a functional role in permitting the peasant economy to exist and adapt to the pervasive political and legal interventions of the state. 13. Mexico's Second Agrarian Reform. Given the increasing ineffectiveness of the ejido system and its incompatibility with a liberalized agricultural sector adjusting to an open economy, the overhaul of the legal framework for Mexico's ejido sector was considered vital. A legal land market, increased security of land tenure and greater economic collaboration with the private sector were argued to be the key ingredients in a reform package that would enable the ejido sector to modernize and adjust to the economic reforms. Thus, in 1992, the Government modified 5 Article 27 of the Mexican Constitution governing all land use in Mexico and the related Agrarian Laws. With respect to the ejido sector, the four most important changes were: * An end to the agrarian reform program that had been in effect for the more than seven decades since the Mexican Revolution, although the concentration of land in large estates remains forbidden and a legal mechanism was created to distribute individual landholdings in excess of the legal size limits; * The removal of prohibitions against the sale, rental and sharecropping of parceled ejido farm land and land for human settlement, although the sale of parceled farm land to outsiders still requires the approval of the ejido assembly (unless the latter had previously approved the passage of the land parcel to dominio pleno or "full title" status)'; * Ejido members were now allowed to redefine the boundaries of communal land, in addition to their traditional right that allows them to assign common land individually (even though it cannot be appropriated individually); and * Economic associations between private sector entrepreneurs and ejidatarios were now allowed. 14. Implications of the Reforms. Expected Benefits. The revision of Article 27 of the Constitution and the subsequent land-titling program (PROCEDE) were expected to have three main benefits. First, they would encourage investment in ejido land, as farmers gained greater land security and therefore a higher expected value of future income and investments in the land. Second, the reforms would increase the supply of credit, as farmers could now use their land as collateral for a loan. Third, the ability to engage in rental and sale transactions would promote a more efficient allocation of land among agricultural producers, as land would be passed from less to more productive farmers. 15. Gender Implications. The new Article 27 of the Constitution gave ejidatarios greater control and freedom over their land resources. However, the individualization of property rights has some gender implications that are likely to affect the intra-household bargaining power and resource allocation decisions. With the advent of individual property rights, wives and children of male ejidatarios no longer have formal legal rights over land transactions involving parcels titled to their husband/father. In other words, male household heads can lease or sell their property without the consent of family members. Wives and children are legally entitled to "first dibs" on purchasing their husbands' or fathers' property (derecho de tanto) for a period of thirty days, but this supposes that they have access to sufficient resources to engage in this sort of intra-family transaction. 16. Another aspect of privatization is that female ejidatarias may be vulnerable to dispossession (Stephen 1994). Women are more likely to lease or lend their land to male relatives and ejido Even with the reforms agrarian communities can not sell their land to outsiders, although they now have the option to convert their status to ejidos and ultimately adopt dominio pleno should they desire. 6 residents as opposed to cultivating it themselves,s and PROCEDE privileges long-term users of land parcels over other claims.6 This problem is exacerbated by the lack of accurate legal information, on the part of many women, regarding their land rights as well as the tendency of male ejido authorities to side with men in the event of conflicting claims. On the positive side, the legalization of land transactions, insofar as it facilitates a widening and deepening of the market, may particularly benefit women, who more frequently exploit their parcels indirectly via tenancy contracts. 17. Implementing the Second Agrarian Reform. The "Program to Certify Ejido Rights and to Title House Plots" (Programa de Certificaci6n de Derechos Ejidales y Titulos de Solares Urbanos - PROCEDE) originated in late 1992 to assist the GOM implement the revised Article 27 of the Constitution. PROCEDE provides certificates and titles to qualified individuals as recognized by the ejido community and in particular grants: (a) farm parcel certificates; (b) certificates of access to communal lands; and (c) titles to house plots. PROCEDE documents are legally required for all land transactions, including rentals. It is a voluntary and free program available to all ejidos and, since 1996, to agrarian communities as well (Vol. II, Cord and Wodon). 18. A PROCEDE certificate to parceled farmland, nevertheless, does not grant the ejidatario full land rights nor can it be used as legal collateral (although it useful to obtain assistance from certain government programs such as PROCAMPO). While ejidatarios can rent and sharecrop their land freely, they can only sell their land to non-ejido members if they have the approval of the ejido assembly or if they have dominio pleno (granted either for a specific plot or the entire ejido - excluding communal lands). 19. Participation in PROCEDE is high, as of December 1997, 59 percent of the ejidatarios have PROCEDE titles, while an additional 20 percent were participating in the program. In contrast, very few ejidatarios have expressed an interest in converting their parcels to dominio pleno and even fewer have gone forward and officially requested the full title status.' During the first four years of implementation, PROCEDE was most prevalent in ejidos that were within easy access of transportation, had electricity, did not have severe boundary problems and had obtained all of the necessary documentation. These ejidos also were more likely to already have active land markets and were home to the more prosperous ejidatarios (Vol. II, Olinto). Reform in an Unfavorable Economic Environment 20. Beginning in the mid-1980s, Mexico initiated a broad and dramatic economy-wide program of reform. This involved opening-up the Mexican economy with an unilateral program of trade For elderly widows, physical constraints lead them to tum cultivation over to male tenants. Younger women are often limited by a gender division of labor that inhibits them from engaging in plowing. 6 Article 48 of the new Agrarian Law established the conversion of use rights to full property rights after five years of continuous use. The participation rates are slightly lower for ejidatarios in the 1997 ejido survey, where only 48 percent had completed PROCEDE and an additional 18 percent were participating (Vol. II, Cord and Wodon). Only in 20 percent of the ejidos surveyed in 1997 did the authorities indicate that there was interest in dominio pleno. In most of these cases, the proposal to move towards dominio pleno had not been officially presented to the ejido assembly. Dominio pleno had been approved by the ejido assembly in eight cases and had been rejected in five cases. 7 liberalization, increasing the role of market forces via. privatization and deregulation, and steeply reducing fiscal expenditures through the elimination of expensive price supports. In the agricultural sector, an important liberalization program was initiated in the late 1980s, and included significant trade and price reforms as well as the privatization of key parastatals. 21. These economic reforms were carried out in a difficult environment characterized by largely stagnant world prices, economic instability, uncompetitive domestic agricultural markets, and high interest rates. The economic reforms, combined with these factors, created an unfavorable incentive framework for agriculture in the 1990s, as indicated by the sector's declining real producer prices and negative nominal protection rates (NPRs). This section reviews the economic reforms, their intended benefits and the overall evolution of the incentive framework for agriculture in the 1990s. The aim is to present the economic context in which the institutional reforms were unrolled in order to better understand the behavior of the ejido sector during the 1990s. 22. Economic Reforms. Trade and price liberalization were the cornerstones of the Mexican agricultural reform program. Agricultural trade liberalization began in 1986, was reinforced soon after when Mexico joined GATT, and culminated with the signing of NAFTA in 1992. Trade reforms included the elimination of quantitative restrictions and licensing requirements for imports and exports, and their selective replacement with tariffs. 23. Significant price reforms were initiated in 1988 and have continued through the present. The key reforms include: * the elimination in 1988/89 of guaranteed prices for wheat, sorghum, barley, rice and oilseeds (although a similar system of agreement prices was in place for many of these crops between 1992 and 1995); * while price subsidies for corn and beans were to be phased-out gradually during the early NAFTA years and replaced with a system of direct income support payments (PROCAMPO), they were precipitously eliminated by the currency devaluation in late 1994, at the same time the PROCAMPO program was introduced; * the withdrawal of the state from procurement and marketing functions (except for corn and beans, where Government sharply reduced its involvement after the 1994/95 devaluation and is scheduled to totally eliminate its participation in 1999); and * the elimination of input subsidies on seeds, fertilizer, pesticides, machinery and diesel fuel (the input subsidy on electricity for groundwater pumping is the only major input subsidy to remain in place). 24. Expected Impact of the Reforms. Economic and trade liberalization was expected to have three effects on the agricultural sector (World Bank, 1992). First, it would improve the allocation of resources, as factors of production would shift out of the inefficient production of corn and into other, more productive activities, such as the production of fruits and vegetables. Second, it 8 would yield significant fiscal savings due to the expected dismantling of producer and consumer subsidies. Third, it would affect income distribution both within rural areas and between urban and rural areas. The impact of liberalization was not expected to be experienced evenly across all farmers (World Bank, 1992; Levy and van Wijnbergen, 1994): * The better-off farmers, located mainly in irrigated areas, were expected to be the largest beneficiaries of the reforms as wage costs would decline and they would be able to take advantage of the new agricultural trade opportunities. * Subsistence farmers were expected to suffer as the value of their labor and rainfed land declined, although to the extent they were net buyers of corn, they would benefit from the lower prices of this basic commodity. * Medium-scale farmers on rainfed land, a not-so-poor and politically vocal group, were expected to incur the largest losses, as the value of their land declined and they lost access to important government subsidies. * With respect to urban residents, a major channel of adjustment was expected to be the displacement of rural labor to urban areas (and the U.S.). Depending on the sources, an estimated 300,000 to 800,000 workers were expected to leave rural Mexico during the liberalization process, creating downward pressure on urban wages and exacerbating urban poverty (Burfisher, Robinson and Thierfelder, 1994). 25. Incentive Framework for Agriculture. The legal and economic reforms described above were carried out in a period of economic adversity, characterized by trade and price liberalization, stagnant sectoral growth and world prices, turbulent macro economic conditions, high interest rates, and uncompetitive domestic marketing channels. This unfavorable environment is reflected in the declining real prices and the negative nominal protection coefficients for key commodities in Mexico's agricultural sector in the 1990s. 26. Real Price Trends. Overall real prices for key Mexican agricultural commodities declined steadily in the 1990s, with the exception of 1996, when prices rose sharply, reflecting the delayed impact of the devaluation and a marked temporary increase in world market commodity prices. Table 1 presents the annual fluctuations in real producer prices between 1991 and 1997 for six basic crops (corn, beans, wheat, sorghum, coffee and tomatoes). Two trends are apparent. First, the steady decline in the 1990s of real commodity prices (with the exception of coffee prices) and second, the tremendous price instability facing Mexican agricultural producers, particularly following the devaluation. For example, prices for corn, Mexico's most important crop, experienced annual real price fluctuations ranging from 3 to 15 percent in the first half of the 1990s, which grew to 20 to 50 percent after the devaluation. 9 Table 1 Evolution of Real Producer Prices in Mexico for Key Crops (in %) 91 92 93 94 95 96 97 91-94 95-97 Corn 8.6 -6.5 -2.9 -8.3 -22.3 32.3 -22.8 -5.3 5.9 Beans -14.2 -13.9 -2.4 -10.6 -25.6 39.4 -2.2 -7.7 22.4 Wheat -0.8 -5.8 -5.2 -4.4 -8.8 26.6 -17.2 -4.6 5.0 Sorghum -9.3 4.8 -8.0 -11.7 -14.7 46.3 -5.9 -4.5 21.0 Coffee -10.3 -14.4 2.9 51.2 36.8 -39.4 43.6 18.3 1.5 Tomatoes -4.1 50.2 -7.1 -2.9 - 18.8 -3.8 0.5 18.6 -2.1 1. Compiled as the annual compounded rate of change in prices for the start and end years. Source: Banco de Mexico producer price series. The prices are deflated by the national producer price index minus the services sector. 27. Nominal Protection Rates. The nominal protection rates (NPRs) for key crops further underscore the existence of an unstable and relatively poor incentive framework facing Mexican agricultural producers in the 1990s (Table 2). The NPR Table 2 measures the difference between Nominal Protection Rates (in percent) 1992-97 nominal producer prices and the 92 93 94 95 96 97 equivalent world market price at Corn 18% 16% -15% -34% -45% -23% the point of collection.9 The Beans -5% -8% -38% -67% -66% N/A Wheat -8% -12% -21% -58% -17% -29% NPR is a reflection of domestic Sorghum -31% -26% -37% -28% -36% -29% output prices, interest rates, Coffee N/A 18% -52% -34% 4% -26% international price trends, trade Tomatoes N/A 15% -26% -67% -63% N/A policy, market transaction costs Source: Olsen (1999), producer prices and adjustment costs from SAGAR. policy,NPR = I - (domestic priceladjusted equivalent world market price). (transportation, storage, fees, etc.) and the exchange rate. The NPR does not take into account subsidies or taxes on inputs (such as the effective protection rate), nor government expenditures on public goods and services (such as the producer subsidy equivalent). 28. The NPR trends indicate that Mexican agricultural prices were below equivalent international prices between 1992 and 1997. First, with the exception of corn in 1992 and 1993, the NPRs for the basic gains and oilseeds were negative throughout the 1 990s. Second, after the devaluation in early 1995, the shortfall between Mexican and equivalent world market prices increased. In 1994, the NPR for corn was -15 percent, while in 1996, it reached -45 percent. Third, by 1997, the gap between world market equivalent prices and Mexican prices declined somewhat and became more homogenous among the different crops, although Mexican prices remained between 20 and 30 percent below world market levels.'0 29. Following 1994, Government rural development expenditures were only partially able to offset the negative implicit taxation suggested by the NPRs. Applying the average level of To derive the equivalent world market price, the CIF prices is adjusted to reflect the marketing costs (e.g., storage, finance, transport costs, etc.) to bring the imported commodity to the producers' point of collection. The NPRs presented in Table 2 are negative 23 % for 1997, compared to 12% for 1998, noted in the AMEC/GAM report on "Mexico's Grain Sector: the Case of Chiapas.' The principal reasons for the differencevare:. (a) higher transport, storage and finance changes, in part due to a more detailed breakdown of the various steps in the marketing chain; (b) the inclusion of importer and trader margins of 5%; (c) a higher import price for GAM/AMEC, who uses March 1998 CBOT prices, compared to the world market average for 1997; (d) higher CONASUPO prices in 1998; and (e) a lower premium used by GAM/AMEC for yellow maize (15% instead of 20%). The detailed cost estimates used by Olsen were provided by SAGAR. 10 taxation for the six crops across Mexico's entire crop sector and comparing this estimate of total implicit taxation to total rural development expenditures reveals that only in 1993 and 1994 were government rural development expenditures greater than the level of implicit negative taxation (Vol. II, Olsen). The burden of implicit taxation was particularly strong in 1995 and 1996, where even after subtracting government expenditures, it equaled around 50 percent of crop GDP (excludes livestock, fisheries and forestry). 30. Explanations for the negative NPRs and Declining Producer Prices. The negative price trends and the price gap between Mexican and international prices reflect several factors, some of which are direct policy variables (trade, exchange rate and price policy) and others which Government only partially affects (interest rates, marketing margins). * Liberalization. First, in the early half of the 1990s, the removal of trade and exchange rate protection depressed domestic producer prices. In 1994, quantitative import restrictions on corn, wheat and beans were removed and replaced with a duty free quota (and an above quota tariff that has never been implemented), and the tariff rates for sorghum, rice and soybeans were reduced. * Exchange Rate. Second, prior to the 1994/95 devaluation, the exchange rate was considered to be highly overvalued, which facilitated cheap imports and also undermined domestic prices. * Inefficient Marketing Channels. Third, given the relatively liberalized environment of agriculture in Mexico, much of the taxation implied by the negative NPRs is likely to be implicit, stemming from unaccounted marketing margins. In this light, the negative NPRs may signal the presence of uncompetitive agricultural markets. * Interest Rates. Fourth, the relatively high interest rates in Mexico raise long-term storage costs significantly (the NPR takes only short-term storage costs into account), which lowers domestic prices. * Adjustment Noise. Fifth, the decrease in the NPRs following the devaluation and the uneven behavior of the NPRs with respect to one another suggest that markets do not adjust smoothly and efficiently to large external shocks (such as a 100 percent devaluation) and that there was a lot of unexplained noise in agricultural markets following the devaluation. The unexplained residual grew after the devaluation, suggesting that potential gains from the currency re-evaluation benefited traders and agro- processors rather than producers. 3 1. Overall Cropping Patterns. A review of broad Table 3 Annul Chngesin Acreage of Key Crops acreage trends indicates a highly responsive 1990-94 1995-97 agricultural sector in which producers seek to adapt (%) 0/0) to the dramatically changing price environment, as Corn 3.8 -0.4 well as to variations in supply shifters (e.g., weather, Beans 1.2 -0.9 input prices) by constantly changing their cropping Sogu 1.5 -5.4 patterns (Table 3 ). A comparison of annual Coffee 2.1 -0.3 changes in cropping patterns between 1990-94 and Tomatoes -5.6 3.4 Source. SAGAR/CEA. * Compound growth rates based on price changes between the start and end years. 11 1995-97 reveals large annual fluctuations in crop area, as well as significant differences in the cropping trends among the different crops. For example, between 1990 and 1994, sorghum acreage decreased at an annual rate of 7.0 percent, compared to a 13.8 percent annual growth rate between 1995 and 1997. 32. There appears to be less variation in the acreage of crops grown by the smallholder sector (beans, corn and coffee) than for the other crops - perhaps suggesting that smallholders are less responsive or were less able to respond during the two periods in question. An alternative explanation for the stability of these traditional crops is that they were attractive to all producers - subsistence and commercial alike - as they offer known prices, production functions and marketing channels, especially in the case of corn and beans, where either a floor price or a subsidized price has been effect through 1999. Distinguishing the Impact of the Reforms 33. The economic and institutional reforms of the 1990s created a new paradigm for the ejido sector that was expected to lead to increased agricultural productivity as well as to facilitate the shift of many ejido producers into non-farm activities. The relatively more attractive non-farm economy due to the removal of agricultural subsidies was expected to pull households away from the farm sector. At the same time, the removal of restrictions on land and labor market participation under the agrarian reform was expected to provide a push effect encouraging ejido households to leave the rural sector. Given the lack of information on private sector producers, which would serve as a control group, and the similar overarching themes of both reforms, it is not be possible to distinguish between the effects of the economic and institutional reforms on the ejido sector. Avenues of Adjustment: 1994-1997 34. The story of adjustment in the ejido sector cannot be told from an individual farmer's perspective, but rather must be seen from the standpoint of a household, or an extended family, that is constantly redeploying its resources between sectors and activities according to where they may be used most efficiently. The analysis suggests that several avenues of change have emerged during the 1990s. The most important are: (a) participation in off-farm activities and migration, which not only stabilize and in some cases increase household income, but has also served as a substitute to Mexico's largely absent rural financial markets; (b) increasing heterogeneity in agricultural strategies, as some producers intensify their agricultural production (often with government support from Alianza and PROCAMPO) and others reduce it, while yet another group experiences little change; (c) an increasing frequency of land transactions; and (d) and residence in a dynamic and well organized ejido. These avenues of adjustment are examined below. Increasing Labor Market Participation and Migration 35. Faced with an unfavorable economic environment for agriculture and a legal framework that no longer discriminated against participation in labor and land markets, ejido households have increased their involvement in off-farm activities (wage and self-employment) and migration. 12 This section first examines participation in off-farm activities, then analyzes the characteristics associated with non-agricultural income, before assessing the evolution of migration in the ejido sector. 36. Participation in Off-farm Activities. Table 4 Since 1994, the share of households involved in Household Participation in Off-farm Activities Share of households in off 1994 1997 off-farm activities rose by 50 percent, reaching fa activities 60 percent of all households in 1997. The largest increase in labor market participation ay or a i% 61% was directed towards the non-farm sector (Table 4). While the increased involvement in off- Ag. Wage labor 15% 18% farm activities was prevalent across all farm Self-employment 9% 24% sizes, the push towards the off-farm sector was Source: 1997 Ejido Study most significant for small households with less than five national rainfed equivalent hectares (NRE) (Appendix Table V-bl). 37. Participation in off-farm activities varies by gender and age. f Women participate less than men in off-farm activities. Thirty-eight percent of the men participate in off-farm employment, compared to 19 percent of the women. The lower participation rates for women may reflect the fact that women have to travel further outside of the ejido to locate non-agricultural employment: 19 percent of salaried women reported working in other parts of Mexico as opposed to in the ejido or a nearby town, compared to 14 percent of salaried men. This most likely reflects men's greater involvement in agricultural production as farmers and day laborers. 3 Moreover, women who do participate in non-agricultural activities tend to do so for longer periods than men. Women tended to work at a salaried job for 8.6 months in 1996, compared to 5.8 months for men (Vol. 11, Katz). * Younger family members participate more heavily in off-farm activities than their parents, suggesting that the importance of off-farm activities is likely to grow. Wile 36 percent of the heads of household participate in non-agrcultural activities, 40 percent of the sons participate. Similarly, while 14 percent of the wives of the head of household participate in non- agricultural activities, 29 percent of the daughters participate in such activities. All measurements of land are national rainfed equivalents (NRE). A hectare of NRE is equivalent to the amount of land necessary to produce the average national com yield on rainfed land where cor is planted in pure stands. An average yield of 1.08 tons of com per hectare is assumed in the calculations (Annex 3). 13 Table 5 Estimate of Non-Agricultural Earnings Per Individual Coefficient p>/t/ Education of HH head (#years) 0.57 0.00 Age of HH head 0.64 0.00 Age Squared -0.01 0.00 Female HH head -2.60 0.00 Self Employ. (SE) (# of HH workers SE) 5.64 0.00 Indigenous SE -0.58 0.79 Female SE 3.07 0.12 Household (HH) Size 0.07 0.52 # of HH Dependents -0.21 0.03 Indigenous HH -0.25 0.73 Ejido Social Cap. Index 0.18 0.16 % Ejido Land Irrigated -3.68 0.00 % Ejido Access Road Paved 2.68 0.00 North Pacific -2.78 0.00 Center -0.41 0.37 Gulf -3.92 0.00 South Pacific -4.18 0.00 # of relatives in MX 0.17 0.57 # of relatives in US -1.21 0.00 % HH land in corn 2.27 0.03 % HH land in basic grains 0.95 0.76 % HH land in fruits & veg. 0.71 0.71 Land/capita 0.15 0.65 HH leasing in land -10.58 0.01 HH leasing out land 5.11 0.19 constant -6.23 0.00 Source: 1997 Ejido Study, Vol. II, Lanjouw Notes: Excluded region is North. 38. Non-Agricultural Income. The level of non-agricultural income (excludes income from agricultural wage labor and self-employment) received by individuals participating in such activities in the ejido sector depends upon four broad sets of factors: (a) first, specific characteristics of the individual, namely age, gender and education; (b) second, the extent to which the household is involved in other income generating activities; (c) third, the type of non- agricultural activity in which the individual is engaged; and (d) fourth, characteristics of the local labor market and economy. These four sets of factors are reviewed below, drawing upon a regression of non-agricultural income on a set of explanatory variables (Table 5).2 The dependent variable comprises annual non-agricultural wages and earnings from self-employment activities for each individual engaged in such activities. 39. Individual characteristics. Characteristics of the individual participating in the non- agricultural activity are key determinants of the level of income received by the said individual. The three characteristics examined in the regression were education, gender and age: a The model uses the censored least absolute deviation (CLAD) technique to overcome problems associated with a regression of non-agricultural incomes on a range of explanatory variables using simple ordinary least squares, which yields biased estimates of the explanatory variables. This is because the regression does not properly take into account the censoring of the dependent variable at zero (corresponding to all households which do not have any non-agricultural sources of income). For more on the rationale of the CLAD and or the methodology see Vol. II, Lanjouw, 1999. Individual earnings from self-employment activities have been calculated as household income from self- employment activity, by type, divided by the number of household members employed in the respective type of self-employment activity. All reported results are significant at least at the ten percent level, unless otherwise indicated. 14 * Education is positively associated with non-agricultural income. For example, an additional year of schooling is associated with a 57 percent increase in earnings from non-agricultural activities, holding all other variables constant. The fact that the younger generation is more educated than their parents suggests that non-agricultural income will become even more important in the future. Older sons and daughters of ejido households are, on average, more than twice as educated as their parents, with a mean of 7 vs. 3 years of schooling." * Gender is an important determinant of non-agricultural income, as women earn dramatically less then men and were less likely to obtain a non-agricultural job.14 The results suggest that a woman could earn up to 92 percent less than a man given the same educational level, age and other explanatory variables.s These results need not necessarily indicate lower wages for the same occupation, as they are partially explained by the fact that women participate less than men in off-farm activities and the fact that the analysis relates to total agricultural income earned per individual and not per unit of time worked. Pagan and Sanchez (1998) also suggest that significant gender-based rural labor market discrimination may exist, although they also note that lower total wage income for women may reflect supply-side constraints, in part as young children have a negative impact on women's participation in the labor force. * Age. Initially, earnings rise with age, but after age 40 they start to decline. 40. Involvement in other income generating activities. The extent to which an individual's household has access to other income generating activities also affects the level of non- agricultural income. The more involved the household is in migration or agriculture (as defined by the leasing in and out variables) the smaller an individual's non-agricultural income. More specifically, if the household leases in land, holding other variables constant, non-farm earnings virtually disappear (a decline of 99 percent), while leasing land out is (weakly) associated with higher non-farm earnings. These results highlight a possible household labor constraint, as households heavily engaged in migration and/or agriculture may not have excess labor for non- agricultural activities. 41. Market Characteristics. The characteristics of the local and regional economy are also important determinants of non-agricultural income. * Region. Relative to the North and Center, non-agricultural incomes are generally lower in other parts of Mexico, controlling for other households' variables (for a description of the regions see para. 112 and Map). Again, these differences may be explained by the higher household off-farm participation rates in the North and Center (Table 17). The higher rates of participation in the North and Center may reflect the fact that these regions are more urbanized and there are more non-agricultural income generating opportunities. "Note that this figure may be biased since it only includes children living in their natal households. Less educated young people may be more likely to marry and form their own households. The Pearson correlation coefficient between age and education for all ejido residents over 14 years old is -.59, which is significant at the 0.01 level. See Vol. II, Lanjouw for a regression on characteristics associated with acquisition of a non-agricultural job. "A coefficient c multiplying a dummy variable can be interpreted as a percent change in the endogenous variable only as long as c is close to zero. For larger values, in absolute terms, the percent change in the endogenous variable is given by 100[exp(c)-l]. 15 * Access to urban areas. Also, the extent to which the ejido is connected to urban areas is also important (as indicated by the share of the ejido access road which is paved). A ten percent increase in the fraction of the road that is paved is associated with a 26 percent increase in income from non-agricultural sources. Good road access may broaden the local economy as more non-agricultural enterprises locate in the community and also reduces the transaction costs of traveling to urban areas to participate in non-agricultural activities. 42. Type of Off-farm Activity. The results suggest that off-farm activities can be highly heterogeneous and associated with very different returns. At the individual level, self-employment activities, as a whole, are more lucrative than wage-labor activities (although they may be accounted for in part by the way in which self-employment activities have been calculated - see Vol. II, Lanjouw). An interaction term between female gender and self-employment activities indicates (weakly) that the lower earnings for females are primarily associated with wage-labor activities, not self-employment activities. The heterogeneous nature of returns to non-agricultural activities suggest that both the very poor and very wealthy participate in non-farm activities in the ejido sector, which is the case in other countries as well (see Reardon et al, 1998) (para. 119). 43. Migration. Complete out-migration from the ejido community appears limited and lower Table 6 than originally expected. Only 18 households out Household Temporary Migration Rates of 1342 appear to have completely migrated out 1994 1997 of their ejidos since 1994. Instead, households Temporary Household Migration us 3 % 8% have chosen to participate selectively in Mexico 10% 7% migration, with one or more (but not all) of the Source: 1997 Ejido Study household members migrating, particularly to the Percentage of HH with current migrants in the US or Mexico. U.S. In 1994, only 3 percent of panel households had current migrants in the U.S., but by 1997, this number had grown to 8 percent, almost all of whom were new migrants.16 In contrast, participation in Mexican migration declined between 1994 and 1997, down from 10 percent to 7 percent (Table 6). The increase in U.S. migration is important due to the high incidence of remitters (more than 40 percent) compared to Mexican migration (25 percent) (Appendix Table VII-4). The rise in U.S. migration and the decline in Mexican migration suggests that real wages in the US rose more than real wages in Mexico, probably due to the impact of the devaluation. 44. As a result of the increase in migration, especially to the U.S., many households have a family member or immediate relative residing outside of the ejido. Almost 45 percent of all ejido households have either a member of the family who has migrated to the U.S., or have children and siblings living there. In the case of Mexican migration assets, more than 80 percent of the households have a family member residing outside of the community. 45. What are the principal characteristics of the migrants? Medium sized households in the North region were the most likely to have migrants to the U.S., while subsistence households in the Center were most likely to have migrants in Mexico (Appendix Table VII-4 and Table 17). In addition, for the first time in 1997, involvement In U.S. migration activities included a small but Migration is defined as having relocated for work purposes for at last one month. Note that children who were non-residents at the time of the survey wee not included in the analysis. 16 significant number of panel households from the South and Gulf. Households in the agriculturally successful Pacific North are the least likely to participate in U.S. migration. 46. In general, men have higher participation than women Box 2 in migration activities, both Migration and Gender within Mexico as well as to the The higher levels of male migration have had consequences for women United States (Vol. II, Katz, on the ejidos. On the one hand, it has induced a "feminization of Table 4). While 13 percent of agriculture" at a time when agriculture is becoming less profitable. And insofar as women remain users but not owners of their husband's all men have migrated to the land, they are constrained in their abilities to increase the productivity U.S. and 15 percent of men have of the parcels. in addition, the higher male out-migration has reduced migrated within Mexico at some the internal ejido level demand for services previously provided by timewomen, such as the sale of, washing and ironing and women must now timein teir ives theincreasingly leave the ejido for off-farm employment. On the other corresponding figures for hand, there is some evidence that the lack of resident men may open up women are less than 2 percent. ejidatario status to women (Vol. II, Katz). The data also suggest that__________________________ women migrate less frequently than men: over the past ten years, the average male migrant has left and returned almost four times within Mexico and twice to the U.S., while a the average female migrant has moved only two times within the country and 1.4 times to the U.S. The widespread male migration has had important consequences for women on the ejidos (see Box 2). 47. The household decision to migrate to the U.S. is also examined using a probit analysis, where U.S. migration is considered a funlction of household and regional characteri StiCS.17 The results imply that land assets, education, migration assets and previous experience in off-farm activities are associated with a higher probability of U.S. migration (Appendix Table Vi-5). It is not the land-poor or destitute that are likely to migrate to the U.S., but rather those who have the land and other assets required to cover the significant fixed costs of U.S. migration. Migration assets are also determinants, as they provide a network and an information source that reduces the transaction costs associated with migration. The significant and negative coefficients on technical assistance (as well as the effect of the agriculturally modernized Pacific North) hint that the relative profitability of agriculture may serve to dissuade U.S. migration. Finally, ethnicity is also associated with a lower probability of U.S. migration. Decreasing Dependence on Local Capital Markets 48. Rural financial markets have traditionally been very weak in Mexico. A major objective of the 1992 reform was to increase access to credit by allowing ejidatarios to use land titles as collateral to obtain a loan. Expanded access to credit was considered vital to raising rural investment and productivity in the ejido sector and supporting the adjustment process. However, rural financial markets have not played a key role in the adjustment process, as the participation of the ejido sector in rural financial markets did not deepen in the 1990s, for a variety of reasons. Consequently, the dependency of the ejido sector on rural financial markets for external capital and savings instruments has remained low. As a result, non-agricultural income (especially e to the low number of migrants, we employ a technique suggested by Maddala, 1984, Limited-Dependent and Qualitative Variables in Econometrics (Cambridge University Press: Cambridge) and randomly select half of the non- migrant households to exclude ftom the equation. 17 remittances) as well as livestock investments are emerging as partial replacements for external capital and formal savings opportunities. 49. This section evaluates the relationship between the ejido sector and rural financial markets and assesses a key alternative that has emerged to compensate for weak rural financial markets. In particular, the section: (a) examines participation in rural financial markets within the ejido sector between 1994 and 1997; (b) offers several explanations for the low participation rates; (c) analyzes the principal characteristics of ejido households that received credit during this period; and (d) briefly considers the role of livestock as a substitute for formal savings opportunities. 50. Participation in Rural Financial Markets. Credit is scarce in the ejido sector, as it is in the rest of the Mexican economy. In 1994, 30 percent of the ejido households received a loan from formal or informal sources compared to 20 percent for 1996/1997. The decline in credit between 1994 and 1997 mainly reflects the decline of the PRONASOL Credito a la Palabra program, which offers small amounts of credit (around US$100) to corn and bean farmers (Table 7). On the other hand, there was a small increase in the share of households with informal credit (mainly from friends and relatives), especially in the case of the households with more land. 51. The low participation rates in capital markets are not unique to the ejido sector. A 1994 study of rural microentrepreneurs in Mexico revealed participation rates only slightly higher than in the ejido sector." The main conclusion of that study was that rural financial markets in Mexico were shallow and inefficient, reflecting both a low supply of financial services and a low demand for financial services. Table 7 Percentage of Households with Credit By Year, Land Size and Source 1994 1996 1997 All 0 e-5 >5 All 0 e-5 >5 All 0 e-5 >5 Formal Credit Banrural 5 7 1 8 4 1 2 7 7 0 3 8 Pronasol 19 21 20 18 4 0 5 3 3 0 5 2 Other govt. 1 3 1 2 0 3 1 1 0 1 1 Commercial bank 1 4 0 2 1 0 0 1 1 0 0 1 Other formal 2 3 1 3 1 0 0 2 1 0 0 2 Informal Credit Friends andrelatives 2 3 1 2 7 3 5 9 5 1 5 6 Comerciante 0 0 0 0 0 0 1 0 0 0 0 1 Prestanista 1 2 1 1 2 2 3 1 1 2 1 1 Other informal 0 0 0 0 2 1 1 2 1 0 1 2 Any credit 30 40 25 33 21 7 18 24 20 4 19 22 Source: 1997 Ejido Study. 52. Why such low participation in rural financial markets? The low participation of the ejido sector in rural financial markets may be explained by several factors that suppress both the demand for and the supply of financial services. In some cases, these factors are unique to the ejido sector, but in many cases they affect all rural households. ' The study revealed that 45 percent of the rural micro-entrepreneurs interviewed in parts of Guanajuato, Puebla and Veracruz had access to credit in 1992-94 (World Bank, 1995). 18 53. One of the most important factors depressing the supply of financial services is the lack of collateral in the ejido sector. Even after the 1992 constitutional reform and the PROCEDE program, ejidatarios do not have full title over their land, which would allow them to use this asset as collateral to obtain credit. Ejido land (even with a PROCEDE title) cannot be sold or expropriated by someone outside of the ejido unless the title is for a parcel with dominio pleno, or unless the ejido assembly approves the sale (which the authorities are unlikely to do in the case of a bank seeking to seize its collateral)." Other factors depressing the supply of financial services are: (a) the weak incentive framework facing agriculture in the 1990s; (b) the relatively high transactions costs associated with the provision of financial services to small farmers; (c) the high levels financial stress on the commercial banks due to the 1995/96 economic crisis; and (d) years of government interventions in rural financial markets, which distorted incentives for rural financial institutions. The demand for financial services was suppressed by the high interest rates, and the absence of financial services suited to the needs of small farmers/entrepreneurs. 54. Characteristics of households receiving credit. What are the characteristics of the households that did receive credit, either in 1994 or in 1997? First, participation in formal credit markets in the ejido sector is associated with a clear focus on agriculture (Appendix Table IV-2). Those households with formal credit in 1997, or which had credit in both 1997 and 1994, had significantly higher areas of corn, fruit and vegetable cultivation, used high yield technologies and increasingly were renting in land. Second, these ejidatarios with formal credit (mainly from BANRURAL) also had a significantly higher incidence of participation in Alianza and PROCAMPO. Third, households receiving credit in both years were associated with residence in the agriculturally advanced Pacific North, while new 1997 entrants into the credit market were significantly concentrated in the Gulf region. 55. In contrast, households that exited from capital markets since 1994 are reducing their participation in agriculture in favor of off-farm activities and migration. This suggests that off- farm income and remittances may be substitutes for credit (Appendix Table IV-2). The households that exited the capital markets tended to be concentrated in the North and Pacific North. 56. Livestock as an instrument of savings. Given the relative absence of rural financial markets, cattle ownership is an important instrument of savings and the participation of ejido households in the livestock sector is rising. Investment in livestock assets, particularly cattle, has increased in the ejido sector since 1994, both in terms of the number of households with cattle and the average size of cattle herds (Table 8). In 1997, almost half the ejido households owned cattle. Households of all land sizes entered the cattle sector between 1994 and 1997, while between 1990 and 1994, it was mainly households with larger plots that entered cattle production (Appendix Table V-5). 57. The importance of livestock ownership as a form of savings is supported by the fact that households with cattle or accumulating cattle tend to have higher levels of assets and are improving their asset base and financial situation. Households with cattle have higher than average land assets, higher than average levels of input use as well as higher than average access ' lowever, legal collateral is only an important asset in obtaining credit to the extent that the legal system efficiently allows financial institutions to seize that collateral. The fact that so few ejidatarios have requested dominio pleno (only one documented case was noted in the survey) suggests that lack of formal collateral may not be perceived by the ejidatario as the critical constraint to obtaining credit. 19 to PROCAMPO and Alianza. Table 8 They also have significantly Ejido Household Characteristics, 1994 and 1997 higher U.S. permanent migration assets, which may be Total assets NREhas 8.10 10.06 ++ financing their accumulation of Land Used has 7.44 9.10 cattle. In addition, first time Cattle cattle owners are also herds of cattle 5.70 6.66 + share with cattle %44 48 purchasing more land than other households and renting Capital Assets Tractor ownership %8 out less land. Households Truck ownership 1I owning cattle in 1997 for the Land_Markets first time appear to be mainly Purchaed 4 from the Gulf and include an Rented more out 7 average number of indigenous Rented more in 9 households, while households HouseholdIncome Pesos 25495 N/a with cattle in 1994 and 1997 Institutions and Organizations tend to be in the North Formal credit 28 19 (Appendix Total credit (including informal) 30 31 ApenixTalev-).Technical assistance %8 6- 58. In contrast, households with no cattle or who are Cor, irrigated has .38 .76 +- selling their cattle are likely to Cor, rainfed has 2.93 3.01 Fruits and vegetables, irrigated has .12 .11 be in financial stress, Fruits and vegetables, rainfed has .37 .46 decapatilizing their assets and Fodder, irrigated has .09 .17 Fodder, rainfed has .96 1.24 perhaps even reducing their Other basic grains has .27 .43 + involvement in agriculture. Oilseeds has .45 .35 All, irrigated has 1.02 1.40 -1* These households lost pasture All, rainfed has 5.63 6.25 + land and have increasingly Agricultural Technology rented out land. Over 10 HYV (improved seeds) 18 23 percent left corn production Chemicals 45 49 1. Number of weighted observations is 1308 for both years altogether and 14 percent left + implies a significantdifference between 1994 and 1997 with a90 percent confidence interval, while frui andvegtabl prducton. implies a 95 percent confidence interval. fruit and vegetable production.Ejido Study. No strong alternative to the agriculture-cattle strategy is observed among these households. While off-farm activities, including migration, increased significantly among this group compared to 1994, their 1997 levels remain significantly below average. Increasing Diversity in Agricultural Strategies 59. Following the institutional and economic reforms, agricultural strategies in the ejido sector were expected to experience major changes. Producers were expected to adjust their cropping patterns and technologies in line with market forces and some producers were expected to exit from agriculture entirely. However, overall agricultural strategies experienced little change in the 20 ejido sector since 1994, with three exceptions, which were driven mainly by households in the North Pacific and Gulf regions. They are: (a) an increase in input use; (b) an increase in irrigated corn acreage and to a lesser extent basic grains; and (c) an increase in land assets. 60. However, these relatively mild aggregate changes in agricultural strategies mask underlying shifts of resources in, out and within the agricultural sector. These changes, largely regional specific, illustrate an increasingly heterogeneous agricultural sector within the ejido. 61. Changes in Agricultural Strategies. Input Use. The first and most notable aggregate trend was an increase in input use. Between 1990 and 1994, there was a sharp decline noted in the use of inputs and technical assistance, suggesting that the sector was becoming increasingly traditional and subsistence oriented (de Janvry, Gordillo and Sadoulet, 1997). However, in 1997, input use recuperated, returning to the levels noted in 1990 (Table 9). For example, 52 percent of the households used chemical fertilizers in 1997, compared to 49 percent in 1994 and 56 percent in 1990. The increase was particularly strong for households with a lot of land and those residing in the Gulf (Table 17 On the other hand, technical assistance remains extremely scarce, as only 6 percent of the households used technical assistance in 1997 compared to 8 percent in 1994 and 61 percent in 1990. 62. The increase in input use may reflect: (a) the introduction of the Alianza matching grant program, which supports the purchase of inputs in some cases; (b) the deepening of private input markets following the privatization in the early 1990s of many of the input parastatals; and (c) increased liquidity due to the PROCAMPO income support payments. 63. While input use may have returned to levels similar to those of 1990, the procurement Household Use of Inputs and Sources of Procurement channels have changed significantly. The share 1990 1994 1997 of inputs procured through official sources has % % % declined from 64% to 13%, while the share of Input Use High Yielding Seed 24 19 25 households using commercial sources has risen Varieties (Table 9). The decline in official procurement Fertilizer 63 54 60 Natural 10 4 13 sources reflects the privatization of many input Chemical 56 51 53 and output parastatals (the latter also often Chemicals 56 47 52 Technical Assistance 61 9 7 provided input) as well as the removal of input subsidies, which had made the official channels Procurement Sources Self Provision 20 10 12 more attractive and stymied the development of Official 64 10 13 private input providers. Commercial 58 61 69 1Social/Ejido 12 16 6 64. Cropping Patterns. The second trend was Source: 1997 Ejido Study (refers to cultivating households only). a Small, but significant expansion of irrigated corn, and to a lesser extent, an expansion in other basic grains as well (Table 8). While the acreage of rainfed con remained roughly constant, irrigated cor acreage doubled (from 0.38 ha to 0.76 ha). The increase in corn acreage, albeit small, may reflect: (a) the fact that corn production requires relatively little labor, which may be important given the household's increasing participation in off-farm activities; (b) the producers' belief that they must remain in corn to receive PROCAMPO (p,ara. 97 or Table 17); and (c) the fact that corn production is less risky than other crops, given its subsistence value, and its known floor price (announced at 21 planting) and marketing channels. The known price is especially relevant given the overall instability of the incentive framework for agriculture in the 1990s (para. 25). 65. Cultivation patterns in the ejido sector are more stable than those noted in Mexico's overall agricultural sector (Table 3). This suggests that ejidatarios faced different supply constraints than other producers or had a more multi-faceted adjustment strategy based on changes in their non- agricultural strategies. While the study lacks the data to compare ejidatarios to non-ejidatario agricultural producers, the analysis does suggest that off farm strategies were important adjustment mechanisms. Another explanation for the stability of cropping patterns in the ejido sector is the fact that ejidatarios concentrate on a few crops, making it less likely for them to switch in and out of a variety of products. A study of the Procuraduria Agraria (based on interviews with 1,360 producers in 1997) revealed that about 90 percent of the ejidatarios grew only one or two crops (Procuraduria Agraria, 1998). 66. Land. The third aggregate trend is an increase in the amount of quality adjusted land owned and cultivated by ejido households, mainly reflecting the fact that ejidos have distributed much of their communal land to their members. From 1994 to 1997, the amount of quality adjusted land used by the average ejido household increased by over 25 percent (Table 8). This increase in quality adjusted land ownership and use reflects: (a) the accumulation of large tracts of land by a few households in the Gulf; (b) quality improvements; and (c) the transfer of communal land to individual control. In the latter case, since 1994, many ejidos have divided their communal land among their members, perhaps in order to title the land individually under the PROCEDE titling program. Land purchases were not a primary factor behind this accumulation of land assets (as they remained relatively small, para. 76). 67. Avenues of Change in Agriculture. Overall agricultural strategies were relatively constant between 1994 and 1997. However, underlying trends show that some ejido households were adjusting their involvement in agriculture. These emerging trends are suggestive of an increasingly heterogeneous agricultural sector within the ejido, and are in contrast to the pre-1992 reform era, when cropping patterns and technologies were fairly homogenous across the ejido sector. Table 10 Currents of Change in Ejido Agriculture (1) More Less Same None No Changes Some Changes Land 29% 9% 62% 6% (2) 62% 38% Inputs 19% (new 11% (exit) 38% (old 31% 64% 30% entrants) and (never current used users) inputs) Corn 28% 15% 48% 9% 57% 43% Fruits and 12% 11% 3% 74% 77% 23% Vegetables 1. The change categories include households with a more than 10% change in the particular variable. 2. Represents the share of households cultivating no land in 1997. Source: 1997 Ejido Study 22 68. Between 60 and 70 percent of the households did not significantly change their basic crop mix, use of inputs or the amount of land they cultivated (Table 10). However, some households increased and/or intensified their agricultural production, while others decreased their involvement in the sector. Between 10 and 30 percent of the producers expanded their involvement in agriculture and/or intensified their production strategies. These households cultivated more land, increased their input use or significantly expanded into fruit and vegetable production. On the other hand, about 10 to 15 percent of the households decreased their involvement in agriculture. These households reduced the amount of land they cultivated, ceased using inputs, and exited from com, and/or fruit and vegetable production. Deepening Ejido Land Markets 69. The participation of ejido households in land markets, particularly rental markets, has increased since 1994. In addition, preliminary evidence suggests that there has been a reduction in transaction costs for land rentals. Both of these trends indicate that the ejido land market has grown. This deepening of the ejido land market has facilitated the increased participation of ejido households in off-farm activities and has positive efficiency implications as well. Table 11 Household Participation in Rental Markets in 1994 and 1997 (in percent) 1990 1994 All Households Households W/O Households W/ PROCEDE Title in PROCEDE 1997 Title in 1997 % of ejidatar. % of ejidatar. % of ejidatar. % of ejidatar. Ejidatarios with land: Owned 98.3 97.0 97.5 96.7 Transacted (in or out) N/A 15.2 10.2 18.0 Not owned 4.9 8.8 6.0 10.3 Rented 1.2 3.0 2.9 3.1 In partnership 0.6 1.2 0.8 1.4 Loaned 1.0 2.3 1.8 2.5 Granted 1.2 0.1 0.0 0.2 Other 0.9 2.2 0.5 3.1 Rented to others 1.4 6.4 4.2 7.7 1997 All Households Households W/O PROCEDE Households W/ PROCEDE Title in 1997 Title in 1997 % of ejidatar. % of ejidatar. % of ejidatar. Ejidatarios with land: Owned 95.5 96.2 95.1 Rented (in or out) 18.5 12.6 21.7 Not owned 10.7 7.0 12.7 Rented 4.2 2.9 4.9 In partnership 2.5 1.4 3.2 Loaned 3.8 2.7 4.4 Granted 0.0 0.0 0.0 Other 0.2 0.0 0.2 Rented to others 7.8 5.6 9.0 Source: 1997 Ejido Study, Vol. II, Olinto. 23 70. However, the role of PROCEDE in fostering the ejido land markets remains unclear. Ejidatarios with PROCEDE titles in 1997 engaged in more rentals in both 1994 and 1997 than ejidatarios without titles. Moreover, households with and without PROCEDE experienced similar percent increases in the frequency of land rentals between 1994 and 1997. The higher rate of land transactions amongst PROCEDE ejidatarios probably reflects the selection bias with which PROCEDE ejidos were selected (para. 19). Finally, the informal nature of many land market transactions implies that over time the official PROCEDE records and titles could become less reflective of actual land ownership patterns. A land market characterized by many informal transactions could then undermine the realization of the expected benefits from PROCEDE and could also threaten the continued reduction in transaction costs that is noted below. These recent developments in ejido land markets are reviewed in more detail below. 71. Frequency ofLand Transactions. Rental Markets. Land rental markets became more active in the ejido sector as a whole between 1990 and 1997 (Table 11). By 1997, almost 19 percent of the households were either cultivating land that they did not own or renting out land, compared to 15 percent in 1994. The 1997 level of participation in rental markets is similar to that found in other countries in Latin America.2021 72. While most of the growth in rental activity happened between 1990 and 1994, this expansion persisted between 1994 and 1997, although at lower rates. Because the reform of Article 27 had made land rental transactions legal beginning in 1992 (even if they were practiced before), it is not surprising to observe an initial boom in rental activity followed by lower rates of growth. Also, while household Table 12 Priiainin Land Rental Markets participation in rental markets between 1990 and 1994 increased for by Farm Size and Region in 1997 both large and small farmers (with more or less than 5 ha NRS), (in %) Not Rented between 1994 and 1997, an increase in participation was only owned to others observed in the small farm sector (Vol. II, Olinto, Table 2). ALL 9.9 7.8 <5 7.5 9.8 73. Given that by 1994, none of the ejidos in the sample had gone >=5 12.1 6.0 through PROCEDE, the observed increase in rental market activity North 11.6 8.5 <5 3.5 17.3 between 1990 and 1994 is unlikely to have been caused by land >=5 17.5 2.1 titling per se. Similarly, the fact that participation rates experienced NorthP. 4.3 18.8 similar increases between 1994 and 1997 among ejidatarios with and <s 0.s 37.5 >5 6.5 6.3 without PROCEDE titles suggests that PROCEDE did not Center 13.0 4.9 necessarily promote the development of rental markets in the sector. <5 5.2 >=5 16.1 4.4 74. The share of households participating in rental markets in Gulf 10.1 9.] 1997 reached almost 22 percent in for ejidatarios with PROCEDE <5 12.7 2.6 titles, and was 13 percent for ejidatarios without PROCEDE titles >=5 9.2 11.3 (Vol. II, Olinto). However, both of these participation rates South P. 4.1 5.1 <5 2.1 5.9 represent a 22 percent increase over the respective levels in 1994. >=5 6.5 4.1 The higher rate of land transactions amongst PROCEDE ejidatarios Source: 1997 Ejido Study, Vol. II, Olinto. 20 In Nicaragua, 21 percent of the households engaged in rentals, in Paraguay 26 percent and in Colombia 23 percent. Given that these surveys selected all farms and the ejido survey only interviewed land owners, one would expect a higher share of participation in rental contracts in the other surveys (Deininger, 1998; Davis, 1998; and Olinto, 1997) 21 The share of households involved in land exchange and ownership transactions presented in this section are not significantly different across the years and therefore are not representative of the entire ejido sector and are limited to the panel of surveyed households. Moreover, the fact that land rentals were illegal prior to 1992 may have led to the underreporting of such transactions. 24 may reflect the fact that during the first four years, PROCEDE was most prevalent in ejidos that were within easy access of transportation, had electricity, and did not have boundary or missing documentation problems, i.e., ejidos that were less isolated and more economically diversified and likely to engage in land transactions (para. 19). 75. Participation in land rental markets varies by region and farm size. Both large and small farmers (with less than 5 NRE hectares) participate in land rental transactions, however, large farmers are more likely to cultivate land they do not own, while small farmers are more likely to rent out land (Table 12). While this trend remained constant in the two periods, small farmers, experienced a greater increase in both types of rental transactions compared to larger farmers. Overall, ejido land rental markets appear to be most active in the Pacific North, especially with respect to renting out and least active in the South Pacific.22 76. Ownership Transactions. As expected, changes in ownership occur less frequently than rentals and sharecropping, although the number of transactions per year grew steadily between 1992 and 1997. There were a total of 186 land ownership transfers between 1992 and 1997 (Table 13)." Of these, 35 percent were purchases and 30 percent were sales, the remainder were bequests, grants and inheritances. The relatively modest increase in land transactions since 1992 reflects the fact that there was relatively little pent-up demand for such exchanges (due to the earlier non-compliance with the restrictions - para 12) and/or the fact that agricultural land transactions generally are rare compared to rentals and other temporary transactions. While ejidatarios mainly bought land from other ejido members (92%), sales were more evenly distributed between members of the ejido (41 percent) and non-members (59 percent). Slightly under half of the parcels transacted between 1992 and 1997 were less than 3 hectares (Vol. II, Olinto). Table 13 Breakdown of Ownership Transactions by Category and Year (# of transactions) '' - .92 9,3 -94 95 90 97 "* TOta Purchases 12 9 8 8 15 13 65 Bequested 1 2 1 2 6 Granted to others 2 1 5 4 5 1 9 27 Inherited 5 3 2 10 Received as grant 1 6 5 3 7 22 Sales 1 10 9 13 14 9 56 Total 21 11 31 30 38 37 18 186 ***For the 18 observations, it was not possible to ascertain the nature of the transaction, since these were the households not found in 1997. Source: 1997 Ejido Study, Vol. II, Olinto 77. Lower Land Market Transaction Costs. Prior to the rewriting of Article 27 of the Constitution, there were high transaction costs on land rentals imposed by the various restrictions. High transaction costs make it difficult for households to participate in land market transactions A study by the Procuraduria Agraria (PA) indicates renting out land was more common than renting in land, a trend which Table 11 above confirms. While in 76 % of the cases, ejidatarios rented in from other ejidatarios, in 50% of the cases ejidatarios rented out to non- ejidatarios. (Procuraduria Agraria, 1998). 23 Despite the inclusion of 18 ejidatarios who had sold their land and could not be found in 1997, the proportion of sales is likely to be underestimated because of non-random attrition. That is, it is likely that several of the missing households for which no information about their fortune was obtained in 1997 may also had sold their holdings. 25 and may also be indicative of an uncompetitive market. Table 14 presents the results of the estimation of a "friction" model for net rentals, where desired net land rentals are not always realized because of friction in the market, which may reflect high transaction costs. 78. In the model, the amount of land rented is a function of time, regional and household characteristics plus a friction parameter. The latter is specified as a function of a 1997 dummy. The main result is that, compared to 1994, transaction costs were lower in 1997, as indicated by the significant and negative coefficient for the 1997 dummy in the friction term (the composition of the transaction costs and possible reasons for decline are areas for further research). The model also reveals that: (a) households with permanent U.S. migration assets tend to have a higher demand for rented land, perhaps because they can use remittances to finance additional land rentals; (b) households residing in the North Pacific region have a lower demand for rented land, compared to the excluded North region; (c) households with higher levels of land assets have a lower net demand for rented land (ceteris paribus);and (d) larger households tend to have a significantly higher demand for rented land, as they are more likely to have excess household labor. 79. Informality of the Ejido Land Market. Despite the new legal and institutional framework designed to support land transactions in the ejido sector, the majority of transactions are informal and are not always fully recorded at the ejido level, or when appropriate, with the National Table 14 Maximum Likelihood estimation of the "friction" model for the land rental markets Estite ,Std Error t-ratio Constant -1.04 1.61 -0.65 1997 time dummy -0.08 1.01 -0.08 Land owned (NRE ha) -0.20** 0.05 -4.1 Land owned squared 0.08 0.05 1.57 Household Size 0.47** 0.16 2.89 Dependency ratio -0.70 1.60 -0.44 Max. education of a HH member 0.17 0.13 1.31 Procede dummy 0.06 1.08 0.06 North Pacific Region dummy -6.24** 1.27 -4.91 Center Region dummy 0.80 1.16 0.69 Gulf Region dummy -0.87 1.31 -0.67 South Pacific Region dummy -0.69 1.42 -0.48 Temp. HH migration assets -Mexico -0.75 1.17 -0.65 Temp. HH migration assets - USA 0.88 0.98 0.90 Perm. HH migration assets - USA 0.53 0.33 1.61 Friction Term: Constant 24.83** 1.28 19.35 1997 dummy -1.83* 1.02 -1.79 o 14.75** 0.51 28.40 Number of Observations 2580 Log-Likelihood -2484 *Significantly different from zero at the 10% level **Significantly different from zero at the 5% level Source: 1997 Ejido Study, Vol. II, Olinto 26 Agrarian Registry (RAN).24 Although the role of the new institutions appears to be increasing.25 One possible motivation for the high level of informality in the ejido land market is that PROCAMPO payments go to the user, not the owner of the land. For this reason, many ejidatarios prefer informal rental/loan agreements to formal contracts, which could place their PROCAMPO payments at risk (Case Study of Ejido La Misa, Vol. II, Katz). 80. The informality of the transactions may: (a) render the PROCEDE cadastre obsolete in a few years, thereby raising transaction costs associated with land rental and ownership transactions; and (b) suggests that the new institutions have not yet had time to penetrate the ejido sector or that it is not considered important to formalize land transactions. The comments below highlight the informality of many of the current land transactions in the ejido sector. * Exchange of titled land. The majority of the transactions between 1992 and 1997 did not involve the exchange of land with a PROCEDE title, although the share of transactions involving land with a PROCEDE title did increase during the period (Vol. II, Olinto) (about one quarter of all land transactions between 1994 and 1997 involved land with a PROCEDE title or dominio pleno). * Documentation of the transaction. Thirty percent of the land transactions involving ownership changes were by word of mouth and sixty percent were recorded in writing (Vol. II, Olinto)26. The fact that a transaction was recorded in writing did not necessarily imply that it was documented in a central record of the ejido, given the fact that most ejidos do not maintain written records. Only 13 percent of the ejidos in Mexico have a register and 16 percent maintain regular written records (INDA, 1996). * Informing the RAN and ejido authorities. According to the law, all sale transactions to non ejido members should be recorded with the RAN. However, despite the fact that 59 percent of all land sales were to non-ejido members, only 28 percent of the sales transactions were recorded between 1994 and 1997. (Vol. II, Olinto). In contrast, ejido authorities were notified for 43 percent of all land transactions (i.e, for 61 transactions). The relatively high proportion of land transactions approved by the ejido officials suggests that the traditional authorities still influence land transactions Dynamic Ejidos Facilitate Adjustment 81. The economic and institutional reforms of the 1990s have reduced the official role of the ejido in the productive activities of its members. Ejido authorities no longer control, dejure, the distribution of land within the ejido nor the distribution of BANRURAL credit (which often was linked to the practice of specified cropping technologies and patterns), although they still are responsible for the management of communal resources (forests, pastures, etc.). Ejidatarios 24A high degree of informality for both -tals and ownership transaction was also noted the 1997 survey of the PA. In temporary contracts, almost 80 percent were verbal, while less than 60 percent of the ownership transactions were registered before any authority (ejido or RAN). (Procuraduria Agraria, 1998). 25 The RAN was established by the reform of Article 27 in 1992. It serves as the registry for all ejido and agrarian community land and is the counterpart agency to the national registries that document transfers for all other forms of land tenure in Mexico. 26 For the remaining 10% (18 observations) it was not possible to ascertain the formality of the transactions since these were the households not found in 1997. 27 relate, more frequently than before the reforms, directly with the Government via programs such as Alianza para el Campo and PROCAMPO. 82. This section examines how the ejido is functioning today, particularly with respect to ejido based social capital. Social capital refers to the network of formal and informal organizations and community institutions that regulate and manage decision making and collective action. In this study, it is defined as those ejidos with regular assemblies, norms and practices to manage common resources (e.g., pasture and forests) and links to higher level umbrella organizations in the social sector. 83. The principal conclusions of the analysis are: (a) compared to 1994, ejidatarios are placing less importance on the regular convening of the assembly and on linkages with other regional or national ejido organizations, and more importance on the management of common resources; and (b) ejidos with higher levels of social capital are associated with greater levels of income, education, social service access and private sector contracts, and lower levels of participation in off-farm activities. In addition, ejidatarios residing in ejidos with higher levels of social capital are more likely to view PROCEDE favorably (Vol II, Cord, Gacitua-Mario and Wodon). 84. Ejido Functioning. To assess how the ejido is functioning today, the analysis focuses on the operation of the ejido assembly and the linkages between the ejido and higher level organizations (e.g., regional or national organizations that regroup smaller community based ejido organizations). The ejido assembly is the highest authority in an ejido. It is comprised of all the ejidatarios and is responsible, inter alia, for setting the norms for contributions to public works, the use of common property and for land management (i.e., accepting or rejecting the ejido's participation in PROCEDE, resolving boundary conflicts, and approving dominio pleno and land sales to outsiders). The assembly has some enforcement capabilities, but relies mainly on consensus and peer pressure to achieve compliance. 85. Compared to 1994, ejidatarios are placing less importance on the regular convening of the assembly and on linkages with other higher level ejido organizations. On the other hand, the ejido assemblies are focusing more on the management of the common ejido resources, although their role in this area remains limited: * Although more than half of the ejidos had monthly assemblies in 1997, the number of ejidos not having assemblies at all increased from 4 percent in 1994 to 14 percent in 1997. Participation in the assemblies tends to be high with more than 80 percent of the households responding that they always attend the assemblies. Less that the majority of ejidos, 41 percent, indicated that they had updated internal regulations for the ejido, which had been a requirement of the 1992 legislation related to the modification of the Constitution. * With the exception of requirements to repair fences, which existed in half of the ejidos in 1997, ejido requirements governing grazing limits, weed control and pasture rotation were relatively infrequent (i.e., only 13 percent of the ejidos registered such norms). Nonetheless, compared to 1994, the share of ejidos claiming they had regulations in these areas was higher. The study did not reveal the extent to which these norms and regulations were actually implemented by the community. 28 * The share of ejidos participating in higher level ejido based organizations declined from 32 percent in 1994 to 24 percent in 1997. 86. Implications ofEjido Functioning. To analyze the implications of ejido social capital on ejidatarios and their communities, an ejido social capital index was created clustered around three groups: (i) ejido level organizational variables (frequency of meetings of the ejido assembly, average quorum in assemblies, existence of written norms and regulations, update of the regulations following the 1992 reform); (ii) multi-ejido level variables (whether the ejido belongs to one or more Union de Ejidos or Asociaci6n Rural de Interds Colectivo, or to other organizations with social, economic or political benefits) and: (iii) ejido level variables on the management of common resources (forest cooperation, grazing cooperation, existence of common pastures, rules for the repair of fences, limits to heads of cattle that can be grazed, weed control, rotation between grazing areas, grazing fees, penalties for non-compliers, existence of common forests, etc.). 87. Ejidos with high levels of social capital are viewed as dynamic communities. It should be noted at the outset that our social capital index reflects mostly ejido-related indicators. Due to the limits of the survey, dimensions of social capital linked to other traditional and informal organizations are less well represented in the index than the pure ejido-related variables. 88. The ejido's ranking in this index was then compared to several household and community characteristics. The main results are the following: * ejidatarios with higher rates of participation in off-farm activities (including migration) and low incomes (in the bottom 40 percentile) tend to reside in ejidos with low levels of social capital; * ejidatarios with higher educational levels and who are indigenous tend to reside in ejidos that rank higher on the ejido-based social capital index; * communities with higher levels of ejido based social capital tend to have higher access to social services and private sector contracts and have fewer problems with water scarcity; * access to government assistance or problems of deforestation and erosion on communal lands are not associated in any significant way with the index; and * ejidatarios with high levels of social capital were more likely to be in favor of PROCEDE, implying that the adoption and implementation of PROCEDE would be higher in communities with high levels of social capital (Vol. II, Cord, Gacitua-Mario, Wodon). 89. These results do not exhibit causality in any form. In other words, higher levels of ejido- based social capital are merely associated with higher income and educational levels, but do not necessarily cause them. Nevertheless, five preliminary conclusions can be drawn that should be confirmed by future research. First, the role of the ejido in the productive processes is becoming increasingly heterogeneous. In ejidos where agriculture, particularly livestock or forestry, offers significant income-generating opportunities, these ejidos are likely to score higher on the social capital index and will continue to be involved in the management of common resources, such as pasture and forests. In contrast, ejidatarios who are increasingly searching outside the ejido for their livelihood (shown by high rates of participation in migration and off-farm activities) will 29 focus less on the management of their common resources and local community, and their ejido is likely to score lower on the index developed above. Box 3 Participation in non-Ejido Organizations Several researchers have noted that the new institutional framework has stimulated the organization of intra-ejidal groups that have more narrowly defined economic objectives, generally in the areas of agricultural production, marketing, commercialization and agricultural financing. These rural economic organizations have become more dynamic and extensive as small producers face a different economic scenario with new rules for resolving conflict and accessing resources and services (Fox 1997, Rello et. al, 1990). Producers may prefer these smaller, more flexible and focused organizations over the large sectoral organizations, as the latter may be less effective in managing small projects and resolving productive problems at the low level. Almost 40 percent of ejido households participate in formal and informal rural organizations. Approximately 4 percent of the households participate in each of the following formal organizations: irrigation groups, rural producer societies, cattle rancher associations, forestry associations, women's agricultural production groups (UAIMs), social solidarity firms (SSSs) and ejido enterprises. Approximately 1 percent of the households participated in producers associations, cooperatives, savings groups, credit unions and marketing groups. Approximately half of the ejidatarios participated in informal associations. Membership in the latter was measured by involvement in traditional collective activities such as the tequio ayuda muta and faena. The principal household contribution in all types of associations, according to the ejido level survey, is family labor. Monetary contributions were only important in some of the formal producer associations, such as cattle associations. The lack of monetary investment in the smaller organizations could represent a drawback for their professionalization, as most of them do not have the regular staff and/or resources to develop their organizational and management skills. 90. Second, ejido-based social capital is important to gain access to basic social services and infrastructure. In this manner, well functioning ejidos may come to resemble the traditional community councils seen in many other countries. In other cases, for instance in ejidos with high levels of out-migration, the role of the community in accessing basic infrastructure and services is likely to decline. 91. Third, the low association between the functioning of the ejido and access to government services is to be expected given the fact that the ejido assembly has less control over access to government resources, given such programs as PROCAMPO and Alianza. Fourth, it would appear that private sector organizations prefer to deal with the ejido as a whole entity and not with individual ejidatarios. This is perhaps because they are more confident in the implementation of a contract signed by ejido authorities or perhaps they are just exploiting economies of scale in dealing with the entire community as a single partner in a contract. Fifth, given the apparent decline in the role of the ejido and the state in productive decisions and activities, it is likely that ejidatarios will increasingly seek membership in non-ejido based productive (Box 3). PROCAMPO and ALIANZA -A New Generation of Government Programs 92. In the mid-1990s, the Government initiated two new programs, PROCAMPO and Alianza, to support the adjustment of the agriculture sector to an open market economy. These programs offer a new generation of government assistance that aim to promote investment and raise the productivity of agriculture without distorting market incentives. This section examines the overall objectives and outreach of both programs. The principal conclusions are: 30 * PROCAMPO is widely received in the ejido sector, with a participation rate of 84 percent. Despite the broad outreach of the program, there are some concerns about its overall transparency, particularly concerning the eligibility requirements and the annual certification process, which may have undermined the program's impact and participation rates. * PROCAMPO provides an average of 8 percent of household income, but this may rise to as high as 40 percent for low income families. As expected with an income transfer program, participation in PROCAMPO is associated with lower poverty (defined as those households in the bottom 40' income percentile). In addition, preliminary econometric evidence suggests that PROCAMPO may have a multiplier effect in that a transfer of one Peso generates in the end an additional income of two Pesos for a household. While more research is needed, there are three potential explanations for this multiplier effect. * First, the income effect of the PROCAMPO payment may lead to a net increase in household investment, encouraging the producers to purchase additional inputs and to obtain more agricultural income than they otherwise would. * Second, the PROCAMPO payment, by providing a constant stream of income to the household, could encourage the household to invest in riskier and higher yielding investments and/or could again raise the household's propensity to invest since it is assured a constant stream of income. * Third, there may also be a Keynesian multiplier effect whereby the increased PROCAMPO income generates additional local consumption demand, which translates to increased product sales and work opportunities for households. The strength of the Keynesian multiplier effect depends on how much leakage there is for external products. * In the case of Alianza, the program reached about ten percent of the households in 1997. The program's low outreach may be explained by its recent introduction, not enough program dissemination in the ejido sector, and the low level of available counterpart resources on the part of ejido households * The initial analysis suggests that there was no significant impact of participation in Alianza on poverty. This may reflect the fact that: the program was introduced relatively recently and that the investments have not yet had time to pay off; the value of the Alianza transfer to the household is relatively small; and the case studies noted some government procurement of inputs and requirements for group membership in some Alianza subprograms, which may reduce their impact. 93. PROCAMPO. The Program of Direct Payments to the Countryside, known as PROCAMPO, is an income support program for agricultural producers that began during the Fall/Winter 1993/94 agricultural cycle. The program was initiated to ease the removal of the guaranteed prices for corn and beans and provides agricultural producers (those with the legal usufruct rights over the land) with a fixed payment per hectare based on historical acreage cultivated with nine crops. 94. It is a transitional program expected to last for 15 years and terminate in 2008. During the first ten years, the amount of payment is to remain constant in real terms and then is to be phased 31 out. However, between 1994 and 1997, the amount of payment per hectare in real terms declined an average of 5 percent per year, due to budgetary constraints (Vol. II, Cord and Wodon). 95. PROCAMPO covers almost 14 million hectares, or 90 percent of Mexico's cultivated area. In the Fall-Winter cycle of 1997, almost three million agricultural producers received PROCAMPO payments for a total of 7.5 billion pesos (US$ 984 million). Of the area covered, roughly 80 percent was rainfed land and about 45 percent belonged to producers with 5 hectares or less. The regional distribution of PROCAMPO was fairly similar to the breakdown of agricultural GDP for Mexico (Table 15). Table 15 National Distribution of PROCAMPO by Region and Size of Producer (1997) Disfributionu of. Dstribution Distribution of Distribution of North 21% 30% 7% 23% North Pacific 17% 12% 2% 10% Center 36% 34% 19% 15% Gulf 12% 8% 5% 2% Pacific South 14% 16% 11% 4% Total 100% 100% 45% 55% Source: SAGAR 96. PROCAMPO and the Ejido Sector. The principal aim of PROCAMPO is to provide income support to producers of basic crops in Mexico. In this respect, Government has been successful in delivering the income support to a broad number of producers in the ejido sector. The large majority of the ejidatarios (84 percent) in the survey participated in PROCAMPO receiving a payment for 5.2 hectares on average. The participation rates were lowest for ejidatarios renting out all of their land (probably because the PROCAMPO benefit was given to the renter) and for subsistence producers with less than two hectares. Similar to the national data, the share of ejidatarios with five hectares or more benefiting from PROCAMPO was 49 percent. Table 16 Perceived Land Use Requirements for PROCAMPO There are no cultivation requirements, the land may remain fallow and still qualify 14% The PROCAMPO land must be cultivated, but with any crop 52% PROCAMPO land must be cultivated with one or more of the PROCAMPO crops 49% There must be a PROCEDE title for the PROCAMPO land 2% The producer must have a Banrural account to receive his payment 2% 1. More than one answer was possible per household. Source: 1997 Ejido Study, Vol. II, Cord and Wodon. 97. The principal reason why 16 percent of the ejidatarios did not participate was because they did not comply with the requirements (60 percent) or because they did not know about the program (22 percent). Many of the producers who are not participating in PROCAMPO may in fact be eligible. While the PROCAMPO program is widely known and has a positive image in the ejido sector, there was some uncertainty over the program's management at the local level. In 32 particular, the ejidatarios provided a wide variety of responses on the cultivation requirements (Table 17). The lack of understanding of these requirements may reflect the fact that the cultivation requirements have changed three times since the program's inception in 1994. In addition, the selection process to identify the qualified beneficiaries is not always clear to some ejidatarios. During the case studies, the ejidatarios indicated that they were not sure how the certification process takes place. In some cases, the list of eligible producers was updated yearly and in other cases it was a closed document. 98. What are some of the key determinants of participation in PROCAMPO? A probit on participation in PROCAMPO revealed that indigenous households were more likely to participate in PROCAMPO than non-indigenous households (despite the fact that there was no significant difference between the participation rates for indigenous and non-indigenous ejido households) (Vol II, Cord and Wodon). Larger households were also more likely to receive PROCAMPO. While participation rates were lower for households owning a high share of irrigated land, they were actually higher for households who cultivated a high share of irrigated land, probably reflecting the fact that the payment went to the household cultivating the land, and not to the owner. Residence in the North Pacific, South Pacific and Center regions had a negative impact on PROCAMPO participation, compared to the North and Gulf regions, holding all other factors constant. Finally, the overall availability of PROCAMPO in the given state had a positive impact on household participation in the program, although this relationship was non-linear and grew smaller in areas where program participation was very high. 99. Impact of PROCAMPO. On average, PROCAMPO payments Figure 1 provide about 8 percent of ejidatario Contribution of PROCAMPO and other Income Sources to Household Income By household income across all income deciles, and its contribution may reach as high as 40 percent for low- income families (Figure 1). As 45 expected with an income transfer 30 program, participation in 12 Othr mmme PROCAMPO reduces the 110 likelihood a household being below Ejdo I coe . - the poverty line (defined as being in the lower 40t' income percentile) by about 42 percent (Vol. II, Cord and Source: 1997 Ejido Study. Ejdio income indicates income from community owned Wodon). resources, such as forestry enterprises. 100. What explains the significant impact of PROCAMPO on poverty in the ejido sector? Does PROCAMPO have a multiplier effect on hou:,-.hold income, whereby one Peso provided by the program results in a more than one peso increase in household income? Three explanations are proposed to explain this potential multiplier effect. First, the income effect of the PROCAMPO payment leads to a net increase in household investment, for example, encouraging the producers to purchase additional inputs than they otherwise would (i.e., it would be more than an infra- marginal increase). The expanded use of inputs then leads to a higher net return from the beneficiaries' agricultural activities. This explanation is supported by the fact that about 70 33 percent of the program beneficiaries claimed to use at least part of their investment to purchase inputs. 101. Second, the PROCAMPO payment may provide a constant stream of income to the beneficiary households, which encourages the household to invest in riskier and higher yielding investments. It has been shown that given their limited ability to handle risk, poor rural households in Mexico allocate their limited labor, land and capital in such a way as to avoid risk and income fluctuations (World Bank, 1995). In such a context, the assurance of a basic stream of income may reduce the risk aversion of the poor ejidatarios, making them more likely to shift out of maize and beans into higher yielding and riskier asparagus and strawberries. Third, there may also be a Keynesian multiplier effect whereby the increased income generates additional local consumption demand, which translates to increased product sales and work opportunities for households. The strength of the Keynesian multiplier effect depends on how much leakage there is for external products. 102. The impact of participation in PROCAMPO on household income was analyzed using a three stage least squares model to avoid problems of endogeneity between household income and the amount of the PROCAMPO payment (Vol. II, Cord and Wodon). While more research is needed to confirm this finding, the model suggests a PROCAMPO multiplier effect and portrays that a Peso transferred through PROCAMPO creates 2.2 Pesos of household income. Still, given the lack of information on the program's administrative costs and alternatives, no firm conclusions can be drawn at this stage about the program's overall effectiveness. That is, more work is needed to provide a full cost-benefit analysis. 103. Moreover, despite the positive impact of PROCAMPO on household income, certain structural features may limit the program's overall contribution to household income. In almost all cases (95 percent), the PROCAMPO payment arrived after planting, which may have diminished incentives to use PROCAMPO to purchase inputs, thereby reducing the program's multiplier impact. In addition, the declining and variable amount of the PROCAMPO payment may reduce the certainty associated with the 15 year program and undermine some of its risk reducing aspects. 104. Also, in very few cases do ejidatarios use their PROCAMPO payment to obtain credit from a trader, which could encourage further farm or off-farm investment. In less than 10 percent of the cases, producers signed over their PROCAMPO check to an input supplier or deposited it in a bank. The apparent hesitancy of ejidatarios to use their payment as collateral for credit may reflect their risk management strategy, which encourages the ejidatarios to use their PROCAMPO resources for both investment and consumption purposes (Vol. II, Cord and Wodon). As currently designed, producers must sign over their full PROCAMPO payment to a bank or input supplier, which denies them the flexibility to use it for a variety of purposes. In addition, banks and traders may be unwilling to accept an expected PROCAMPO payment as collateral, given the variable amount of the payment and its timing, as well as uncertainty as to whether or not the local authorities will certify the ejidatario as eligible to receive the payment for a particular year. 34 105. Alianza Para el Campo. In mid-1996, the Government launched a far-reaching national rural development program, known as Alianza para el Campo (Alliance for the Countryside) to foster agricultural productivity through productive investments (under a matching-grants scheme) and the provision of support services (research, extension, information, training). Alianza offers a menu of 24 federal and at least 10 state and regional programs. The major programs of ALIANZA are ferti-irrigation, mechanization, rural equipment, pasture improvement, and kilo for kilo. These programs account for just over half of total federal expenditures on Alianza. 106. The cornerstone of the Alianza program is its decentralized approach and the delegation of administration and decision-making to the States. Almost all Alianza programs require co- financing by state governments and especially producers. 27 In 1997, Alianza accounted for just under ten percent of the Ministry's programmed expenditures and was the third largest program managed by SAGAR (following PROCAMPO and CONASUPO). The distribution of Alianza resources by region is fairly similar, with the exception of the Center region, which received almost double the share of the other regions. At a national level, the program intended to target ejidatarios in 1997 who were scheduled to comprise 72% of the 1.4 million expected beneficiaries. In actuality, comuneros (those producers residing in agrarian communities) and private sector producers comprised the majority of the 1.1 million beneficiaries (Vol. II, Cord and Wodon). 107. Alianza in the Ejido Sector. The participation of ejidatarios in Alianza hovers at just over ten percent. The probit on ejido household participation in Alianza suggests that ejidatarios with more land, higher incomes and residing in the Gulf region and in more developed and less isolated ejidos (as defined by the percentage of ejidatarios in the community with lighting and access to paved roads) are more likely to participate in Alianza (Vol. II, Cord and Wodon). The higher participation of households in the Gulf region is further underscored by the fact that their participation rates were more than three times higher than the other regions. 108. What is the impact of participation in Alianza on the ejido household? In general, the value of the Alianza transactions were small as ejidatarios tended to select programs that required little counterpart funding and where the total investment is also low (such as kilo por kilo and rural equipment). The majority of the Alianza beneficiaries in the survey received less than 1000 pesos (US$117) and contributed less than 1000 pesos (65 percent and 76 percent respectively). The value of the transactions (both the amount paid in and the producer's contribution) rise dramatically with the amount of land used per producer. The value of the transactions of indigenous producers was less than for non-indigenous producers, suggesting that while their level of participation equals that of ejidatarios, they receive a proportionately smaller share of the program's financial benefits than the ejidatarios (Vol. II, Cord and Wodon). 109. The results of the poverty probits suggest that Alianza has no significant effect on a household's likelihood of being poor (Vol. II, Cord and Wodon). This may reflect the fact that at the time of the survey in 1997, participants had benefited from Alianza for a maximum of two crop cycles and that it vill take more than one year for such an investment program to have a beneficial impact on income. Alternatively, the insignificant income impact may reflect the fact 27 In 1996, producers contributed an average of 50 percent, the federal government 32 percent and the state governments 18 percent. The producer's contribution varies, but may goes as high as 75 percent for programs that support the purchase of large scale machinery and is 0 percent for the popular improved seed programs. 35 that the overall contribution of the program to household income is very small or negligible - either due to the amount of real resources transferred or due structural aspects which undermine the effectiveness of program resources. Some of these structural features are reviewed below. 110. First, the requirement for some Alianza subprograms that ejidatarios form a group to access the Alianza resources may undermine the program's potential impact for several reasons. During the case studies, some ejidatarios noted that previous group efforts with producers' associations carried with them high transaction costs. In addition, even in communities with strong existing organizations, some ejidatarios noted that they felt uneasy with group managed programs because they perceive themselves and their organizations as not having the sufficient skills to adequately manage the process. Another concern raised over group participation is that some ejidatarios may face the choice of joining a group to request assistance that is not their top priority or not participating at all (which may lower the overall level of participation in the program). 111. Second, during the case studies, some ejidatarios noted that Government officials, under certain subprograms procure the goods and supplies instead of the producers themselves. As a result, the inputs procured may not meet the ejidatarios specifications, which may undermine their usefulness. Third, while not directly related to the program's immediate impact, the fact that the government tends to procure the inputs under Alianza undermines the development of local retail markets and the overall sustainability of the program. If producers are to purchase many of the inputs supported by Alianza in future years with their own resources, it is important that they strengthen their ties with local suppliers and that the latter are aware of the producer's needs so that they will be more likely to stock the relevant items in following years. Regional Forces ofAdjustment 112. This section examines household strategies in each of the five regions and suggests regional forces are an important determinant of household options and adjustment strategies.21 While most households are increasingly involved in both the off-farm and agricultural sectors, the relative focus of the ejido households' on these two activities may vary by region. 113. Households in the North Pacific, with the highest average income levels, have retained a strong focus on agriculture and have actually decreased their participation in off-farm activities (Table 17). Their strong agricultural involvement is noted by their significantly above average land holdings and use of improved seeds and technical assistance. 28 Region 1: North (Coahuila, Chihuahua, Durango, Nuevo Le6n, San Luis Potosi, Tamaulipas, and Zacatecas), 21% (compared to 24% for national population of ejidatarios); Region 2: North Pacific (Baja California, Baja Califomia Sur, Nayarit, Sinaloa and Sonora), 9% (compared to 9.5% for national ejido population); Region 3: Center (Aguascalientes, Distrito Federal, Guanajuato, Hidalgo, Jalisco, M6xico, Michoacin, Morelos, Puebla, Quer6taro, and Tlaxcala), 34% (compared to 31.5% for national population);Region 4: Gulf (Campeche, Quitana Roo, Tabasco, Veracruz, and Yucatan), 20% (compared to 19.1% for national population)Region 5: South Pacific (Colima, Guerrero, Chiapas and Oaxaca), 15% (compared to 16.1% for national population). 36 114. Households in the North and Gulf have significantly rising participation in off-farm activities and migration, as well as a strong involvement in agriculture and livestock. Although average household income levels are similar, households in the Gulf appear to be investing more in agriculture than households in the North. Average land holdings, livestock herds, participation in credit and land markets, and use of chemicals have risen significantly for households in the Gulf, while either declining or experiencing no significant change for households in the North. Table 17 Household Characteristics by Region Z enteremnaaray caenrat resl sAlmtF ans Fas is name -1 NbEtht. -;Ts Jest .. I akip d id- e . - ) 1)..i.: ( ..- '() (2) (1.2 Q aI l3". units 1994 1997 1994 1997'W, .1994 1997 - 1994 1997 - 1994 1997 Niof neighied ,i.,2 I - - 4 2 Lan Ase mp.Ee 6.0 I A 2 + ++I Canton r.er i.ifca.le - fr e.f sihae 111 rde ':: '# 4 HiI ! V 1 Human Capalj Edcn:r head z: - Tru or .-.- -, Ika.u.r*nc rnmt-e V offfmal "re I 1- 4 ag lal Technol i dc4 5L I i I + ++r1 .11grafttan23 '*: 2" K . I share .-I ifti-DSe -IL, here it a i -ocn.e++ rnemter c ent I I . N~tl _ CaPLal4sseLs~1 LandmItrisers V - Organtialnif 9 v '+ Tchmr 'I >"V; 4;&~ V Fne. Ii - I r, 1'. , ouc:1997r, Ejido Study. SPIuie Fecholog j4~tr I. ~ 9Ainzr'ne I'. . I, i. '. .r I Ge n programs 2-~n Mc a.-iZ vi cl II I .. ccjmp.:. Household -' .44L Ar 1 ln.:onw I,c,NfI ti, . --ij H Toul.P. qF., , -' - , --. -1 I I,J t n." L,: Land Ilarke[s I C -7'4 1 w. ' M. -r MI emr .. mre ':-u' 1. The first test column indicates whether Or not there was significant change for the vaniable between 1994 and 1997, '+ is statistically significant at 90% level, while '++'I is statistically significant at 95% level. 2. The second test column indicates whether or not the region's score for the vaniable is significantly different from the survey's average, '--' is statistically significant at 90% level, while -H-' is statistically significant at 95% level. Source: 1997 Ejido Study. 37 average household income levels are similar, households in the Gulf appear to be investing more in agriculture than households in the North. Average land holdings, livestock herds, participation in credit and land markets, and use of chemicals have risen significantly for households in the Gulf, while either declining or experiencing no significant change for households in the North. The lower involvement in wage labor activities in the Gulf may reflect the region's lower degree of urbanization (45 percent of the population resides in communities of 2,500 inhabitants or more compared to 69 percent for the North). 115. Households in the Center and the Pacific South have also significantly expanded their participation in migration and off-farm activities, while continuing to practice traditional agriculture. These households also have the lowest average income levels, which is suggestive of the presence of few adjustment options. Households in the Center appear more oriented towards migration and especially off-farm activities, while households in the Pacific South have more agricultural assets (mainly land). In both cases, there are no signs of significant investment in agriculture, as the use of technology, capital assets, access to Government programs remains significantly below average in both regions. It is in these regions that the Government needs to concentrate efforts to explore and support alternative development strategies for the farm and non- farm sectors. Implications for Rural Poverty Factors Associated with Poverty in the Ejido Sector 116. Poverty in rural Mexico is pervasive and severe. In 1994, 47 percent of the ejido households were below the poverty line and 34 percent were in extreme poverty.29 At this time, the evolution of poverty in the ejido sector since 1994 is not evaluated because the income measures used in the two survey periods are sufficiently different so as to make comparisons across time unreliable (Vol. II, Lanjouw). Consequently, the analysis: (a) first, examines characteristics associated with poverty in each period (defining poverty as the population in the bottom 40 percent in per capita terms for each year); and (b) second, evaluates factors associated with the emergence from poverty (i.e., households who were able to rise from the bottom 40 percent in 1994 to join the top 60 percent in 1997). 117. The main conclusions are that lower poverty is increasingly associated with three strategies: (a) participation in off-farm activities, particularly commerce related activities; (b) an intensification of agricultural activities, and in particular specialization in fruits and vegetables; and (c) residence in a dynamic ejido (i.e., a community with a high level of ejido based social capital). 118. Probit regressions hre used to measure the association between poverty and a range of household, ejido and regional-level characteristics in 1994 and 1997. The probits indicate where explanatory variables are significantly associated with poverty, but they do not indicate causality between the explanatory variables and poverty. In particular, the probits examine the relationship between poverty and off-farm income, agriculture, participation in government programs, the level of social capital within an ejido, credit, region and ethnicity, and the previous poverty " De Janvry, Gordillo and Sadoulet, 1997. These rates were calculated using the 1989 adjusted ECLA poverty lines and are similar to ECLA's overall estimate of rural poverty in Mexico for 1992 as well as the World Bank estimates for 1992 (World Bank, 1996). 38 ranking of a household. The principal results are presented below (Table 18 and Vol. II, Lanjouw). 119. Off-farm income. Off-farm income has traditionally been associated with lower poverty, but this relationship appears to be growing stronger over time in the ejido sector: * Off-farm income and migration assets have traditionally been associated with lower poverty. In 1994, the poverty rate for households with low labor market participation was 65 percent, compared to the overall rate of 47 percent. The probits provide further evidence that off-farm income and migration assets were both associated with lower poverty in 1994. In the case of off-farm income, the relationship with poverty grew even stronger in 1997 compared to 1994. * However, the association between off-farm income and poverty is not indiscriminate, as some off-fann activities were more associated with lower poverty than others. Households engaged in commerce related wage activities were less likely to be poor than households engaged in casual non-agricultural or manufacturing labor. * International migration assets were associated with lower poverty in both periods, while Mexican migration assets were not significant in either period. * Higher education and fewer dependents were also associated with lower poverty, providing further evidence that non-agricultural income is associated with lower poverty; as both education and few dependents facilitate access to non-agricultural income (para. 39). Both higher education and lower income are known to be associated with lower fertility. 120. Agriculture. The level of land assets is no longer as closely associated to poverty as is how the land assets are used. In 1994, the more land cultivated by a household the smaller the likelihood that it would be poor (as indicated by the negative coefficients on per capita land assets and on leasing-in land). However, in 1997, lower poverty was more related with how land is used (e.g., cultivated with fruits and vegetables) and less related to the size of the plot cultivated. The share of land allocated to fruits and vegetables was the largest significant variable in the equation, while the coefficient for leasing-in was insignificant and leasing-out was positively associated with lower poverty. 121. Government productivity programs. Participation in PROCAMPO is significantly associated with lower poverty, which is to be expected as the income transfers provided under the program raise household income and reduce poverty. On the other hand, neither PROCEDE nor Alianza were significantly associated with lower poverty, suggesting that these programs have not been targeted to either high or low income households or have not been effective, as yet, in reducing poverty (para. 109). 39 Table 18 Correlates of Poverty in 1997 Probit Estimates Number of obs = 1039 chi2(38) = 173.79 Prob > chi2 0.00 Log Likelihood = -522.94 Pseudo R2 = 0.24 Coefficient P> z (dF/dx) HH Poor 1994 0.27 0.00 HH Education of most educated HH member, 1997 -0.01 0.04 Household (HH) size, 1997 0.03 0.02 # dependents, 1997 0.02 0.16 # non-family member relations in Mexico, 1997 -0.00 0.70 # non-family member relations in US, 1997 -0.06 0.00 # workers in casual non-ag wage labor, 1997 -0.08 0.17 # of workers in commerce related wage labor, 1997 -0.30 0.02 #.of workers in manuf. Related wage labor, 1997 -0.03 0.66 # of workers in other wage labor in 1997 -0.15 0.00 # of workers in commerce related self employment in 1997 -0.12 0.00 # of workers in other self employment in 1997 -0.08 0.04 %Indigenous dummy 0.07 0.20 % of land in corn, 1997 -0.03 0.68 % of land in basic grains, 1997 -0.10 0.44 % of land in fruits and vegetables, 1997 -0.38 0.00 HH borrowed from BANRURAL, 1997 -0.15 0.07 HH borrowed from PRONASOL, 1997 -0.03 0.79 HH borrowed from other government source, 1997 -0.06 0.68 HH borrowed from other formal source, 1997 -0.21 0.11 HH borrowed from friends and relations, 1997 -0.01 0.83 HH borrowed from moneylender in 1997 0.21 0.06 HH borrowed from other informal source in 1997 -0.14 0.30 Per capita land owned in 1997 -0.03 0.01 HH leased in land in 1997 0.03 0.59 HH leased out land in 1997 -0.19 0.02 Ejido social capital index in 1997 -0.07 0.00 % ejido land irrigated in 1997 -0.17 0.11 Kms to nearest urban center in 1997 -0.00 0.61 Distance squared to urban center -0.00 0.94 % of road to closest urban center paved in 1997 0.00 0.93 HH PROCEDE beneficiary in 1997 -0.04 0.43 HH PROCAMPO beneficiary in 1997 -0.10 0.08 HH ALIANZA beneficiary -0.07 0.26 North pacific 0.04 0.68 Center 0.03 0.60 Gulf -0.11 0.11 South pacific -0.13 0.04 ------------------------------------------------------------------------- obs. P] 0.38 pred. P | 0.34 (at x-bar) (*) dF/dx is for discrete change of dummy variable from 0 to I z and P>|zj are the test of the underlying coefficient being 0 Source:1997 Ejido Study, Vol. II, Lanjouw 40 122. Credit. Borrowing from Banrural is associated with lower poverty, while borrowing from a moneylender is associated with higher poverty. This observation is consistent with credit market segmentation and rationing of the poor from low cost credit sources. 123. Ejido social capitaL Residence in a cohesive, functioning ejido also appears to be increasingly associated with lower poverty as shown by a negative coefficient for ejido social capital index (see para. 86 and Vol. II, Cord, Gacitua-Mario and Wodon). This equivalent index was not significant in 1994, perhaps suggesting that residence in communities with increasing cohesiveness and dynamism was more important in 1997 than in 1994. While this finding is consistent with the argument that social capital and strong communities perform an important function in lifting people out of poverty (Narayan, 1997), it is important to recognize that the correlation could just as easily be driven in the opposite direction - that richer households are more likely to be able to afford to spend the time attending ejido meetings and participating in other activities underlying this social capital index. 124. Poverty lag. Earlier work on rural poverty and rural financial markets in Mexico suggests the existence of a rural poverty trap in Mexico (World Bank, 1995 and 1996). While the 1997 poverty probit does offer some evidence of stickiness in emerging from poverty, it does also suggest that the poverty trap may not be as closed as earlier perceived. The probability of being poor in 1997, taking all other explanatory variables at their means, is 27 percentage points higher if the household was poor in 1994. The next section analyzes more closely the characteristics of households that have been successful in emerging out of poverty. 125. Ethnicity, Region and Poverty. Traditionally, ethnicity and residence in the Pacific South and the Gulf have been associated with higher poverty. In 1994, the poverty rate in the ejido sector was 65 and 57 percent in the Gulf and the South Pacific (excluding Chiapas) and capped 70 percent in indigenous communities (de Janvry, Gordillo, and Sadoulet). 126. However, the poverty probits suggest that region and ethnicity in and of themselves do not have a clear relationship with poverty. The probit revealed no significant relationship between ethnicity and poverty in 1997. Similarly, households residing in the Pacific South or the Gulf (weakly significant) were less likely to be poor than households residing in other regions in 1997. Although further research is necessary, these results suggests that households residing in the Pacific South or indigenous households do not have an inherent set of characteristics that keeps them poor; but instead that they lack the key assets, skills and infrastructure necessary to allow them to participate fully in economic development (Box 4). Another contributing factor to the high poverty rates among households in indigenous communities and the South may be discrimination in labor markets, which precludes them from participating in the more lucrative segments. This issue should also be explored further. 41 Box 4 Indigenous Households in Mexico's Ejido Sector The importance of assets in overcoming regional and ethnic barriers to economic development is highlighted by differences between indigenous ejido households residing in the Gulf, the Center and the South Pacific. Between 1994 and 1997, indigenous households in the Gulf significantly increased their use of chemical inputs and expanded their participation in off-farm activities. Also during this period, they acquired significantly more land and benefited from significantly higher levels of government support than other indigenous households. In contrast, indigenous households in the Pacific South and Center, did not see their land assets increase nor were able to modernize their agriculture, despite an increased participation in off-farm activities. Incomes among indigenous households in the Center and South are more than 50 percent smaller than for indigenous households in the Gulf. The causes behind the different adjustment path require further research, but a preliminary assessment suggests that the higher levels of initial land assets, PROCAMPO support and access to credit and wage labor led to the higher levels of investment in agriculture as well as to the higher incomes. Household Assets and Characteristics by Indigenous Category Indigenous in Gulf Indigenous in South Pacific Non-Indigenous HHs and Center Test Test Test Test Test Test units 1994 1997 1994 1997 1994 1997 # ofweighted obs. 80 80 135 135 1088 1088 Land Owned NRE has .8.26 15.87' ++ ++ 3.81 4.55 -- 8.64 10.35 ++ + Cattle (#) # .3.28 6.67 ++ 2.54 3.52 - 6.29 7.07 ++ Education years 4.47 4.06 -- 4.66 ++ Share of households with at least one member working offfarm % 24 73 ++ ++ 49 61 + 44 60-++ wagelabor % 21 66 ++ ++ 36 41 39 46 ++ selfemployed % 6 22 ++ 15 24 8 24 ++ Total Credit % 32 55 ++ ++ 8 13 -- 32 31 Agricultural Technology HYV % 11 19 7 8 -- 20 26 ++ 4-+ Chemicals % 69 85 ++ ++ 23 26 -- 46 50 Government programs Alianza para el campo % 20 + 15 12 - Procampo % 92 ++ 79 84 Household Income Pesos 25309 12433 - 27187 ++ 1. The first test column indicates whether or not there was significant change for the variable between 1994 and 1997, '+' is statistically significant at 90% level, while '++' is statistically significant at 95% level. 2. The second test column indicates whether or not the region's score for the variable is significantly different from the survey's average, '+' is statistically significant at 90% level, while '++' is statistically significant at 95% level. * 39% of these households reside in the center and 57% reside in the south Source: 1997 Ejido Study. Emerging from Poverty 127. Who were the households that were poor in 1994, but were no longer poor in 1997? To answer this question, several household and regional characteristics from 1994, as well as the differences between these variables in 1994 and 1997 were regressed on the change in a 42 household's poverty status across the three years. The regression includes two types of variables: first, level variables corresponding to the base year (1994); and second, differences in these variables between 1994 and 1997 (Table 19). * The results confirm the importance of an increase in off-farm income and again highlight the heterogeneous nature of the returns from different off-farm activities. An increase in the average level of participation in commerce and other "modem" non-agricultural occupations among the sampled households in the ejido increases the likelihood that a given household will emerge from poverty." In contrast, an increase in the average level of participation in manufacturing ejidos had no significant effect on escaping poverty. * The role played by dynamic communities in emerging from poverty is again underscored by this analysis. In particular, residents of ejidos where since 1994 the members became more dynamic as a community (as defined by an increase in the ejido social capital index) were more likely to exit from poverty in this period. * The results also suggest that it was possible for households with little non-agricultural experience in 1994 or education to escape poverty. Neither the average 1994 level of involvement in non-agricultural activities nor the household's 1994 educational level were significantly related to a reduction in poverty in 1997. * The analysis also provides further support that the area of land cultivated and changing crop patterns were not significantly associated with a move out of poverty between 1994 and 1997. * Larger households or those that reduced the number of dependents were more likely to emerge from poverty between 1994 and 1997. These results suggest the presence of a labor constraint and the need for households to be actively involved in non-agricultural as well as agricultural activities to escape poverty. * Finally, the results suggest that residence in the Gulf was positively associated with emerging from poverty, while ethnicity was not associated with a greater or smaller likelihood in emerging from poverty. These results highlight the importance of the Gulf effect and also suggest that ethnicity is not necessary correlated with rural poverty in Mexico (Box 4). 30 To avoid problems of endogeneity, a proxy based on the involvement of all the households surveyed from an ejido in non-agricultural activities is used to represent the role of such activities in emerging from poverty. 43 Table 19 Regression of Changes in Household Income Ranking VanabeNam Coeffcient P /t// Household (HH) size, 1994 0.02 0.00 ~ HH Size 0.00 0.54 # Dependents, 1994 -0.01 0.66 -# Dependents -0.03 0.01 Education of most educated HH member, 1994 -0.00 0.41 -Education 0.00 0.84 Indigenous dummy -0.04 0.16 Ejido social capital index, 1994 0.00 0.70 -ejido social capital index 0.02 0.05 North Pacific dummy -0.02 0.72 Center dummy 0.00 0.89 Gulf dummy 0.21 0.00 South Pacific dummy 0.05 0.16 Sum of casual non-agricultural workers among HH sampled in ejido, 1994 (1) 0.00 0.89 -Change in sum of casual non-ag. workcrs ~ 0.21 Sum of commerce and other "modern" non ag. workers among HH sampled in -0.00 0.87 ejido, 1994 (1) -Change in sum of commerce and other "modem" non ag. Workers 0.01 0.01 Sum of casual non-agricultural manufacturing workers among HH sampled in -0.00 0.75 ejido, 1994 (1) -change in sum of manuf. workers 0.01 0.25 Avg. land under com in HHs sampled in the ejido, 1994 (1) 0.07 0.15 -Change in has of land under corn 0.02 0.74 Avg. land under basic grains in HHs sampled in the ejido, 1994 (1) 0.06 0.61 -Change in has under basic grains 0.14 0.35 Avg. land under fruits and veg. in HHs sampled, 1994 (1) -0.07 0.50 -Change in area under fruits and veg. -0.14 0.24 Constant -0.24 0.00 Number of Observations: 1046 1. These variables are constructed at the ejido level, rather than the household level, so as to avoid problems of endogeneity. The variable summarize the explanatory variable at the ejido level, where for each household the summarizing calculation ignores the information for the specific household in question. - change Source: 1997 Ejido Study, Vol. II, Lanjouw Signs ofIncreasing Integration 128. The economic and institutional reforms have supported the increasing integration of the ejido with the non-farm economy. Over time, these growing connections are expected to reduce differential returns to labor, capital and even land within the ejido and non-ejido farm sectors. This ultimately should lead to a reduction in rural poverty in Mexico. This section highlights the key signs of the integration of the ejido with the non-farm economy and non-ejido sector and then briefly reviews several implications of the integration process. 129. The three key signs of integration of the ejido sector with the non-farm economy and non- ejido farm sector are the: (a) growing participation of ejido households in off-farm activities and the importance of off-farm income to the ejido household; (b) declining agricultural focus of some 44 ejido households as shown by rising land market transactions; and (c) increased frequency of land transactions in the ejido sector. These three signs are reviewed below. 130. Since 1994, the share of households involved in off-farm activities rose by 50 percent, and reached 60 percent of all households in 1997. Households participating in off-farm activities are motivated either by the higher return of these activities for their household (and over time are likely to exit completely from agriculture), or by a desire to diversify in order to raise and stabilize their income. 131. The increased participation of ejido households in off-farm Fure 2 Sore fIncome activities is also highlighted by the importance of off-farm income to the ejido household. 45 In 1997, off-farm income provided an average of 44 30 culture percent of total household income in the sector. The 15 .. . Other contribution of off-farm income remained relatively constant 2 6 9 10 across all income deciles, which underlines: (a) the extent to Source: 1997 Ejido Study. which the ejido sector has become thoroughly integrated with the non-farm sector; (b) that returns to income in off-farm and agricultural activities are highly variable and (c) that a diversified structure of income, although important as suggested above, is not sufficient to emerge from poverty (Figure 2). 132. The increasing integration of the ejido sector with the non-farm sector has reduced the importance of land assets in determining household income. As expected, the importance of off- farm income is higher for households with few land assets: households using two or less hectares NRE received 60 to 70 percent of their income from off-farm sources, compared to 25 to 35 percent of income from off-farmi sources for households with 10 or more hectares NRE (Table ). The importance of off-farm income to subsistence producers suggests that land assets are no longer the main determinant of income. 133. Another sign of integration, particularly with the non-ejido farm sector, is the increasing frequency of land transactions. The frequency of land transactions (excluding changes in ownership) increased by 22 percent between 1994 and 1997 and it is likely that the increase would be even more marked if comparisons were made with 1990 levels. While most transactions occur within the ejido, it is expected that over time there will be more transactions with non-ejido members. 45 134. The third sign of integration with the non-farm sector is the fact that about 10 to 15 percent of the ejido households appear to be reducing their focus on agriculture, as indicated by a decrease in the use of inputs and land area allocated to corn (Table 10). It is likely that these households prefer to allocate their labor and capital resources to non-farm activities, as they offer them a higher return. Table 20 Sources of Income by Farm Size (in percent) All ha e-2lha -5 haj5-O0ha 10-18 ha 18+-ha Total Farm 44.2 11.7 13.4 34.3 45.2 56.6 63.8 Off-fa rm Wages 24.6 35.4 36.8 31.9 28.6 18.2 9.2 Self-employment 12.0 22.3 25.9 11.0 9.4 10.1 9.0 Remittances 7.5 4.0 6.9 7.9 7.8 7.6 7.9 Other PROCAMPO 7.4 3.8 8.1 8.2 7.2 6.6 8.6 Alianza 0.5 0.0 0.0 0.2 0.6 -0.2 1.8 Land Transactions 1.5 17.3 2.2 0.8 -0.1 0.8 -0.7 Misc. (1) 2.5 18.8 6.5 5.6 1.2 0.2 0.5 (1) Misc. includes income from garden plots, wood sales, and collective activities. Source: 1997 Ejido Study. Gains from Integration 135. The increased integration of the ejido sector with the non-farm sector has led to more flows of labor, capital (wages and remittances) and land resources in and out of the ejido sector. These increased flows have several implications. First, they have placed pressure on farming in the ejido to be competitive to alternative uses of capital and labor both outside of the farm sector and outside of the ejido sector (i.e., the private farm sector). Ultimately, the gap between the marginal productivity of capital and labor between the ejido and non-farm economy will be reduced, thereby leading to higher incomes and a reduction in poverty. Second, more than before, a reduction in the returns to farming will lead to an exit of resources from farming in the ejido sector. However, it is important to note that thus far there has been no massive outmigration from the sector nor a proliferation of distress sales, despite the adverse economic environment for farming in the 1990s. However, as the children of the current ejidatarios inherit their land rights, they may be more sensitive to relative changes in incentives for farming as they are less tied to the land than their parents. 136. Third, as ejido households become more involved in non-farm activities, their attachment to the ejido may decline and the role of the ejido in the productive decisions of its members is likely to decrease. In some cases this may lead to a weakening of the ejido and a loss in its ability to attract key social services and infrastructure to the community. In other cases, the ejido may evolve into a relatively important community council - with little involvement in productive activities, but an important role in attracting and managing common community services and infrastructure. Ejidos with valuable agricultural and natural resources are more likely to remain 46 strong as members seek rely on the ejido structure to ensure that they continue to benefit from and conserve these resources. The role of the ejido in managing its natural resource base should be further explored. 137. The continued integration of the ejido and non-farm sectors and the realization of the benefits of increased income and lower poverty will depend upon the level of development of factor markets and upon the extent to which all ejido households have access to these markets. The shallowness of rural capital markets in Mexico and the suggestion that there may be discrimination against women in labor and capital markets in Mexico highlight the risk that poorly performing factor markets may delay the integration process and undermine its hoped for benefits. Moreover, education and easy access to urban centers (via paved roads) have been shown to important determinants of access to non-agricultural income and it is important that access to these services grows in the ejido sector. Easing the Transition Process: What is the Government's Role? 138. This section addresses what should be the role of Government in supporting the continued adjustment to the economic and institutional reforms of the 1990s. The overarching objectives of the recommendations discussed below are to: (a) support the increasing integration of the ejido with the non-farm economy; and (b) promote the intensification of agriculture amongst low income ejido households with potential. Both of these strategies offer avenues for ejido households to emerge from poverty and could be deepened, where possible, with targeted governmental support in five areas. The aim of this section is not to present a rural development strategy for the social sector, but rather to present broad several lines for future action that flow directly from the analysis presented above. 139. Increasing participation in non-farm activities and migration. Participation in non-farm labor market activities has expanded sharply in the ejido sector since 1994 for both men and women (Table 4). In 1997, 60 percent of the households participated in off-farm activities and they provided (including remittances) an average of 44 percent of household income. The contribution of off-farm income was relatively constant across all income deciles; illustrating that both the poor and the non-poor are involved in the off-farm sector (Figure 2). 140. The returns to different types of non-farm labor vary significantly. This suggests non-poor households participate in activities with the highest returns, such as self-employment and commerce related activities, while the poor participate in activities with the lowest returns such as casual wage labor and manufacturing activities. While women participate less than men do in non-farm activities, their participation is increasing. Preliminary evidence suggests that women may earn less than similarly qualified men and that it is more difficult for a woman to acquire off- farm employment. 141. Despite the variability of the returns to non-farm activities, they are clearly associated with lower poverty and this relationship grew stronger in the ejido in 1997 compared to 1994 (Table 18 and Table 19). Moreover residents of ejidos whe: the members increased their involvement in non-farm activities since 1994 were more likely to emerge from poverty. 47 142. International migration to the U.S. has also increased, while domestic migration appears on the decline (Table 6). The former is important, as family members who migrate to the U.S. are more likely to send remittances than Mexico migrants. International migration assets were associated with lower poverty in both 1994 and 1997. On the negative side, households (often headed by women) left behind in ejidos where there is a lot of migration may have less access to social services and basic infrastructure than other households. 143. Government could undertake several initiatives to enhance prospects for ejido households to participate in off-farm employment and migration. The main priority areas for action are: * ensuring the availability of quality education to all ejidatarios and their families, as education is strongly correlated with access to non-agricultural employment and income, as well as with emerging from poverty -in addition to having an intrinsic value as a tool of empowerment; * providing key infrastructure packages (roads, communications facilities, etc) to improve the quality of access to urban centers and to encourage firms, banks and other facilities to locate in rural areas (Binswanger, Khnadker and Rosenzweig, 1986); * offering tariff rate subsidies to private providers of infrastructure and services (roads, communications, etc.) to encourage them to deliver services and invest in low income areas where the high unit costs and low ability to pay often lead to under-investment; these subsidies should be well-targeted towards low income households so as to limit their fiscal burden; * evaluating the pilot programs that provide information on regional employment possibilities in low income isolated areas; and * expanding efforts to provide migrants with social services (mobile education and health care, child care, etc.) to ensure that the next generation has the appropriate skills to participate fully in labor markets, especially given the tendency of women to migrate for longer periods of time and with their children. 144. Developing Rural Financial Markets. Formal and informal rural financial markets in Mexico continue to remain shallow. Only 20 percent of the households received credit of any kind in 1997 compared to 30 percent in 1994 (Table 7). Households with a greater involvement in agriculture were more likely to receive credit. The low rates of participation in financial markets may, in part, be explained by the fact that PROCEDE titles cannot be used as collateral to obtain credit. A PROCEDE title does not give the ejidatario the right to mortgage his or her land or sell out it outside of the ejido, without the approval of the ejido assembly. 145. While the increasing integration of the ejido economy with the non-farm sector has provided alternative capital flows to the ejido sector, weak financial markets remain a constraint for the household to move into more lucrative non-farm areas such as self-employment activities and modernize agriculture. In other countries, non-formal financial institutions have often filled the gap with services more suited to the needs of rural entrepreneurs. These services are not limited to agricultural activities and offer savings as well as credit instruments. 48 To deepen access to credit and savings instruments in rural Mexico, constraints to the development of non-formal financial markets (trader credit, savings and loans, credit unions, non-government organizations, etc.) should be explored and addressed. 146. Agriculture: Diversification and Specialization. Overall, agricultural strategies experienced little change in the ejido sector since 1994, with three exceptions: (a) an increase in input use, with a recovery to the levels of 1990; (b) an increase in irrigated corn, and to a lesser extent, basic grains; and (c) an increase in land holdings, reflecting mainly (i) the accumulation of large tracts of land by a few households in the Gulf; (ii) the transfer of some communal land to individual households; and (iii) land quality improvements. 147. These relatively mild aggregate changes in agricultural strategies mask underlying shifts of resources in, out and within the sector. These changes, largely regional specific, illustrate an increasingly heterogeneous agricultural sector within the ejido, as: (a) some ejidos modernize their production -mainly in the Gulf and North Pacific; (b) others continue to cultivate, but at a reduced intensity and look increasingly outside of agriculture for their income - mainly in the North and Center; and (c) yet another, larger group scattered throughout the country, but concentrated in the Pacific South, remain as relatively small producers heavily dependent upon subsistence agriculture. 148. To support the overall agricultural adjustment process, Government must recognize the sector's increasing diversity. In particular, priority objectives for the Government to consider are: (a) ensuring that government policies do not impede the adjustment process and the shift of some households out of agriculture; (b) supporting low income households with productive potential in agriculture to invest in and raise their agricultural productivity; and (c) improving the efficiency and competitiveness of Mexico's agricultural markets. 149. To implement these objectives, three sets of recommendations are offered on PROCAMPO, Alianza and product markets. PROCAMPO is a well-conceived program. It provides producers with income support that is non-distortionary and allows households to decide whether to intensify or reduce their involvement in agriculture. The outreach of PROCAMPO in the ejido sector is high (84 percent of all households receive PROCAMPO), with most households using the income support for both consumption and investment purposes. On average, PROCAMPO payments provide about 8 percent of household income in the sector, although the program's contribution may reach as high as 40 percent for low-income families (Figure 1). As is to be expected with an income transfer program, the analysis reveals that PROCAMPO has a strong positive impact on poverty reduction in the ejido sector. Moreover, the analysis suggests a multiplier effect (in that a dollar transferred creates more than a dollar increase in household income). This multiplier effect implies that the program has a positive impact on the household's investment behavior as well as employment and income generating opportunities in the community. 150. To facilitate the use of PROCAMPO for investment and consumption purposes and thereby enhance both the productive and risk management aspects of the program, the following modifications should be considered: 49 * distributing the payment earlier in the crop cycle or at least announcing the amount of payment prior to planting (even if the actual payment is made after planting, this will add some certainty to the producers income stream); * allowing ejidatarios to designate half beneficiaries - either banks or traders that could receive half of their PROCAMPO payment, while they retain the other half for consumption purposes or emergency use; * eliminating planting and land use requirements, which would lower administrative costs, add certainty to the producers' income stream (they will not have to requalify each year) and allow for earlier payment of PROCAMPO. If this does not seem viable, the planting and land use requirements could be disseminated more widely in the rural sector to increase the likelihood that banks and local traders would accept PROCAMPO payment as collateral; and * make the list of eligible producers open to the public at the local level, which will also add certainty to households stream of PROCAMPO payments and facilitate long-term investments. 151. Alianza Para el Campo aims to promote agricultural investment and modernization, which as noted above is an important avenue of poverty reduction. The program began in 1996 and by the time of the survey in 1997 the target population only had access to the program for one or at most two crop seasons. In 1997, 11 percent of ejido households were participating in the program. The relatively low participation rates may be explained by: (a) the relatively recent introduction of the program; (b) not enough program dissemination in the ejido sector; (c) counterpart resource constraints on the part of the ejido households; and (d) two structural aspects which may undermine the impact of the program and hence its demand among ejidatarios. First, some programs rely on group solicitations for support, which raises the transaction costs of receiving Alianza resources and may undermine the likelihood that producers are able to participate in the programs of most interest to them. Second, the fact under some subprograms government officials are responsible for procurement may have led to some input purchases that do not fully meet the specifications of the ejidatarios as well as concerns that the program is not encouraging the development of local distribution channels. 152. While many improvements have already been introduced to Alianza since the survey was undertaken, several modifications in program design are suggested below that could raise the benefits ejido households derive from Alianza: * allow ejidatarios to purchase their own inputs directly from local distributors, in order to facilitate the development of local private sector retail outlets; * provide program beneficiaries with vouchers that they could redeem at any participating input supplier for the input of their choice. Voucher programs are widely supported by farmers in Eastern Europe because they are more flexible and allow the producer more freedom in deciding when and how to use the subsidy; (Box 5) * increase the dissemination of information on the program in the ejido sector so that a larger proportion of ejidatarios and ejidatarias are aware of the program, understand its objectives and are clear on how to access the resources; 50 * evaluate the requirements for group participation to access Alianza resources to ensure that they are effective and do not make it difficult for individual ejidatarios to select the program most suited to their needs; and * increase the sensitivity of the program staff and design to ensure that women are offered the opportunity to participate in and benefit from the program. Box 5 Voucher Program in Romania In 1997, the Romanian Government implemented a voucher program to support the development of private wholesale channels and retail markets for agricultural inputs and technology. The voucher program is a provisional program which will be phased out in two to three years. In 1997, vouchers were distributed to all landowners commensurate with the size of their holdings, the threshold size being 0.5 ha and the maximum size 5 ha. A farmer could thus receive up to 6 vouchers. Vouchers were distributed in the spring and autumn season at the Post Offices. They could be used to buy seeds and planting material, fertilizers, pesticides, fuel oil, and mechanical services. Input suppliers could redeem the vouchers at selected banks which were later reimbursed by the Ministry of Agriculture. In 1998, a survey, including only rural households, was conducted to evaluate the program. During 1997, 98.6 % of all vouchers had been distributed. Only 1% of the vouchers were sold, about 7.6% were used as gifts for friends and relatives. The average size of land owned by households receiving vouchers was between 2 and 3 ha. In the autumn season 1997, vouchers accounted for about 30% of the total farmers' costs. Vouchers were given a variety of uses: Large farmers tended to use vouchers mostly to buy seeds and fertilizer, while small farmers, who raise mostly cattle or sheep, used them mostly for buying fuel. This shows that vouchers are more likely to meet the needs of small farmers than traditional agricultural input subsidies (e.g., subsidies on fertilizers and seeds). Farmers' reaction to the voucher program was very positive: 98% of the surveyed farmers were satisfied with the program and 96% favored its extension. Farmers, however complained that the value of the vouchers was too low, the distribution was late, and the applicability of the vouchers limited (veterinary services and livestock equipment were excluded). 153. The analysis of the incentive framework revealed high levels of implicit taxation for agricultural products (Table 2). The significant price wedge between Mexican producer prices and equivalent world market prices revealed by the NPR analysis is highly suggestive of uncompetitive agricultural markets Mexico. The lower producer prices, brought on in part by the market inefficiencies, undermine incentives for agriculture. Improving the competitiveness of agricultural markets is a key area for action, and in particular Government should consider: * Exploring the causes behind the noncompetitive markets. Do they reflect the influence of a monopolistic market structure for farm markets? Do the costs reflect high transport costs due to local monopolies and/or regulation, or do they refilct inefficiencies and distortions along the marketing channel. What is the role played by expensive storage costs due to 51 the high interest rates, a scarcity of rural and wholesale infrastructure, and the poor regulatory and financial framework for storage? 154. Improving the Sustainability and Growth ofEjido Land Markets. Since 1994, the frequency of land market transactions in the ejido sector has increased. About 20 percent of the households participated in rental transaction in 1997, a 22 percent increase over 1994 (Table 11). While the share of households engaging in ownership transactions remains small (less than 5 percent of all households in 1997), the number of transactions has risen gradually since 1992 (Table 13). Moreover, preliminary evidence suggests that transaction costs for land rentals have declined. Both the increasing number of transactions and the decline in transaction costs indicate that the ejido land market has grown. This deepening of ejido land markets has facilitated the increasing integration of the ejido with the non-farm and non-ejido economies and is suggestive of a more efficient allocation of land both within and outside of the sector. 155. The impact of the PROCEDE ejido land-titling program on the growth of land markets remains unclear. Ejidatarios with PROCEDE titles in 1997 engaged in more rentals in both 1994 and 1997 than households without titles. However, both households with and without PROCEDE experienced similar percent increases in the frequency of land rentals between 1994 and 1997 (Table 11). The higher frequency of rentals in 1994 and 1997 among ejidatarios who had a PROCEDE title in 1997 probably reflects the fact that during its first four years, the program was more prevalent in ejidos that were likely to have a greater demand for land transactions (i.e., the ejidos tended to be within easy access of transportation, had electricity and did not have severe boundary or documentation problems). 156. Government must address three challenges in this area during the next five years. First, it must complete PROCEDE in the remaining 41 percent of the ejidos, which may pose more challenges than the first half of the titled ejidos, as the larger and more isolated ejidos and those with pending land issues were left for the end of the titling program. Second, the informal nature of most land ownership transactions could render the PROCEDE cadastre out of date in a few years time and could threaten the continued reduction in transaction costs (as boundaries and ownership rights are not properly recorded). Third, ensure that the program treats female and males equally, particularly in light of the growing role of women in agriculture. To address these three challenges, Government could consider: * assisting communities so that they can properly document intra-ejidal and external land transactions at the ejido level and then transmit the relevant records to the RAN; moreover, the information on land transactions should be made public locally and nationally to ensure the transparency of the emerging ejido land market; * modifying certain aspects of PROCEDE to facilitate its implementation in the more remote regions and indigenous communities (South Pacific and North) and analyzing more closely the impact of PROCEDE on the development of ejido land markets to ensure the program is delivering its expected benefits; and * initiating a comprehensive gender-training program for PROCEDE officials and beneficiaries, which would cover inheritance issues and the rights of female ejidatarios. 157. Learning the Lessons ofDynamic Ejidos and Regions. Although the direction of causality is not clear, the analysis indicates that dynamic ejidos are associated with lower levels of poverty 52 and a higher share of the members emerging from poverty. Dynamic ejidos with high levels of ejido social capital are also more likely to have higher access to basic social services, infrastructure and private sector contracts. 158. The analysis also highlights that indigenous and non-indigenous households in the Gulf region have increased their participation in off-farm activities, modernized their agriculture and been more successful than other households in emerging from poverty. 159. To understand the role of social capital and the factors driving the development process in the Gulf region since 1994, the following recommendations are provided: * Provide training and technical assistance to interested communities to assist them to improve the management of common resources as well as the delivery of social services, while avoiding the temptation to provide them with direct financial resources, which often creates distorted organizational incentives. * Analyze the role of ejido social capital in the development process as well as the effects of government programs such as PROCAMPO, PROCEDE, Alianza, and Solidarity Municipal Funds on ejido social capital. The analysis suggests that participatory programs involving the full ejido membership and involvement, such as PROCEDE, may be conductive to the development of social capital. This research should distinguish whether or not gender differences in participation in government programs has a differential impact on social capital. * Analyze the recent development spurt for indigenous households in the Gulf region to better understand what were the key factors behind the increases in labor market participation and the modernization of agriculture experienced by many households in the region (Box 4). The study also reveals that this region had significantly above average levels of access to credit, Alianza and PROCAMPO and the role of these programs in the development process should be clarified. 53 Bibliography Binswanger, H., S. Khnadker and M. Rosenzweig. 1986. "How Infrastructure and Financial Institutions Affect Agricultural Output and Investment in India," World Bank Working Paper 163, The World Bank. Burfisher, M., S. Robinson and K. Thierfelder. 1994. NAFTA, Mexican Agricultural Policy Reform, and Labor Mobility. Paper presented at the winter meetings of the International Agricultural Trade Consortium. Davis, B. 1997. Economic Reform and Determinants of Income among Agricultural Households in Mexico and Nicaragua. Unpublished Doctoral Dissertation, University of California, Berkeley. Davis, B. 1999. The Adjustment Strategies of Mexican Ejidatarios in the Face of Neoliberal Reform. Working Draft. Deininger, K. 1998, "Can Negotiated Land Reform Enhance Efficiency and Equity? Evidence From Colombia" , paper presented at the American Agricultural Economics Association summer meetings in Salt Lake City, 1998. de Janvry, A., G. Gordillo and E. Sadoulet. 1997. Mexico's Second Agrarian Reform: Household and Community Responses, 1990-1994 (San Diego: Center for U.S.-Mexican Studies). de Janvry et al. "Reformas del Sector Agricola y el Campesino en Mexico. (San Jose: IFAD), 1995. Dovring, F. 1969. "Land Reform and Productivity: The Mexican Case - Analysis of Census Data," LTC Paper 63. University of Wisconsin, Madison. Fox, J. 1997. "How does Civil Society Thicken? The Political Construction of Social Capital in Rural Mexico," State-Society Synergy. Government and Social Capital in Development pp. 119-149. Research Series #24. Berkeley CA: University of California, Berkeley. GAM/AMEC. "Mexico's Grain Sector: the Case of Chiapas." May 1998. Health, J. 1990. "Mexico: Enhancing the Contribution of the Land Reform Sector to Mexican Agricultural Development," Report No. 8310-ME, Latin America and the Caribbean Regional Office, The World Bank. Instituto Nacional de Desarrollo Agrario, INDA. 1996. Principales Resultados de la Encuesta Sobre Sociedades de Solidaridad Social. Direcci6n de Investigaci6n y Diagn6stico, INDA: M6xico, D.F. INEGI, 1991. Censo Agricola, Garodero, y Ejidal, 1991. Mexico. 54 Levy, S. and S. van Wijnbergen. 1994. "Labor Markets, Migration and Welfare Agriculture in the North-American Free Trade Agreement," Journal of Development Economics, Volume 43, No.2., pp. 263-278. Narayan, D. 1997. "Voices of the Poor, Poverty and Social Capital in Tanzania," Environmentally and Socially Sustainable Development Studies and Monographs Series #20. The World Bank. Olinto, Pedro; 1997. "Land Tenure Insecurity, Credit Rationing, and Household Asset Accumulation: Panel Data Evidence from Rural Paraguay." Ph.D. Dissertation, University of Wisconsin-Madison. Pagan, J. A. and S.M. Sdnchez. 1998. "Gender Differences in Labor Market Decisions: Evidence from Rural Mexico." Mimeo. Procuraduria Agraria. Los tratos agrarios en ejidos certificados. Procuraduria Agraria, Mexico, 1998. Reardon, T., E. Taylor, K. Stamoulis, P. Lanjouw and A. Balisacan. 1998. "Effects of Non-farm Employment on Rural Income Inequality in Developing Countries: an Investment Perspective," paper presented at the Agricultural Economics Society Annual Conference, March 1998, University of Reading, United Kingdom. Rello, F. 1998. "Organizaci6n Rural y Capital Social en M6xico." Consultant Report. Rello, F. et al. 1990. Laws Organizaciones Productores Rurales en M6xico. Facultad de Economia, Universidad Aut6nma de M6xico. M6xico, D.F: UNAM. Stephen, Lynn. 1994. Viva Zapata!: Generation, Gender, and Historical Consciousness in the Reception of Ejido Reform in Oaxaca. Transformation of Rural Mexico No. 6, Ejido Reform Research Project, Center for U.S.-Mexican Studies, University of California at San Diego. Thiesenhusen, W. 1996. "Mexican Land Reform, 1934-1991: Success or Failure," in Randall, L. ed., Reforming Mexico's Agrarian Reform, (Armonk, New York: M.E. Sharpe), pp. 35-48. World Bank, 1992. "Mexico Sector Review: Adjustment in Agriculture to the Free Trade Agreement," Report No. 10851-ME, Latin America and the Caribbean Regional Office. World Bank, 1995. "Mexico Rural Financial Markets," Report No. 14599-ME, Latin America and the Caribbean Rt .onal Office. World Bank, 1996. "Mexico Rural Poverty," Report No. 15058-ME, Latin America and the Caribbean Regional Office. Yates, P.L. 1981. Mexico's Agricultural Dilemma. Tucson: University of Arizona Press. 55 Annex 1: Chronology Jan-97 Feb-97 Mar-97 Apr-97 May-97 Jun-97 Jul-97 Aug-97 Sep-97 Oct-97 Initial contact with Government Concept paper Fieldwork Data entryl-IF- - Case studies Data cleanup and management Preparation of synthesis Preparation of Green Cover Nov-97 Dec-97 Jan-98 Feb-98 Mar-98 Apr-98 May-98 Jun-98 Jul-98 Aug-98 Initial contact with Government Concept paper Fieldwork Data entry Case studies Data cleanup and management Data analysis and drafting of background reports Preparation of synthesis Preparation of Green Cover Sep-98 Oct-98 Nov-98 Dec-98 Jan-99 Feb-99 Mar-99 Apr-99 May-99 Initial contact with Government Concept paper Fieldwork Data entry Case studies Data cleanup and management Data analysis and drafting of background reports Preparation of synthesis Preparation of Green Cover 56 Annex 2: Data Sources 1. The analysis draws upon three ejido household surveys, which were carried out in 1990, 1994 and 1997, two community-based surveys and several case studies. In addition, the analysis of the incentive framework draws upon agricultural price data from SAGAR and other national and international sources. 2. The first household survey, which was organized by the Secretaria de Agricultura y Recursos Hydraulicos (SARH) and the Economic Commission for Latin America (ECLA), covered 35,090 ejidatarios residing in more than 5,007 ejidos across the country. The sample frame was based on the 1988 ejido census and was prepared by INEGI. It was a two stage stratified sample with random sampling at both stages. The variable used for the stratification of ejido was the average area of agricultural land per household in the ejido. The second stage involved the random selection of ejidatarios within the ejido and captured 6 % of the ejidatarios in each ejido, and at least 5 ejidatarios per ejido. The 1990 survey had been designed to be representative at the district level, with a precision of ten percent at a 95 percent confidence level on the per-household agricultural area. Weights were applied to ensure that final analysis reflected the national distribution of ejidos. The 1990 questionnaire focused on basic household characteristics as well as agricultural production trends and land and livestock ownership 3. The second household survey, which was fielded by the Secretaria de Reforma Agraria (SRA) in close collaboration with ECLA and researchers from the University of California, Berkeley (UCB), included 1,342 ejido households from 275 ejidos." The families interviewed in 1994 were not necessarily the same households interviewed in 1990, although they were from a subset of randomly selected ejidos from the 1990 survey. This survey was representative at the regional level. The 1994 questionnaire was broadened to include a more detailed accounting of participation in off-farm activities, including migration and self- employment activities, as well as participation in land and credit markets 4. The third survey was fielded in 1997 again by SRA and its agency, the Instituto Nacional de Desarrollo Agrario (INDA) with support from the World Bank and UCB. The 1997 questionnaire included a more detailed treatment of net farm as well as off-farm sources of income. It covers 1,665 households from 286 ejidos, of which some 1,287 unweighted observations can be linked to the 1994 sample. The panel approach used in the survey allows for a high-quality analysis of the adjustment process within the ejido sector, as it controls for many unobservable variables that are not adequately reflected in cross-sectional analyses. 5. As with the 1994 survey, the 1997 data is -esentative at the regional level. Moreover, as all three surveys are based on the same sari.e frame prepared using 1988 census data, they are representative of the ejido sector at that time. However, the advantage of the panel "The 1994 data have received extensive analysis to date (see de Jan -v. Gordillo and Sadoulet, 1997 and Davis, 1997). 57 approach is that it allows the tracking of households over time to analyze how they adjusted to the reforms. 6. Out of the 55 ejidatarios that were not resurveyed in 1997, 18 sold their land and departed the ejido, while approximately 20 could not be located for questioning and the remainder were not interested in participating in the second survey round. The 1997 sample is larger than the 1994 sample reflecting the fact that at least five ejidatarios were surveyed per ejido and the inclusion of ejidatarios from Chiapas, who did not participate in the 1994 fieldwork. Patricia Eyenga \StreetTalk\Projects@Files@LAC\PROJECTS\!MX\NLS\EJIDO\JUNE\MAIN603.DOC 06/03/99 3:04 PM  IBRD 29701 U N IT E D STATES O F A M E R I C A BAJA - ~ cA ORN? RA C Ñ l H U A H, U A BA7 -RNE-AKDON LIFOR NUEVO ON BQUNIARIES SR-V~~ARBOJDIB N )O0 T H DUPANGO Æ -< "-' iTRt4fQA.BUDRB 0 T~MULPAS ® AAZACATECASLP NAR 0 000 ~ 09 ooioo$ PrCZACATECAS L F G U E R R EHBEUZ Jc --P A lFlc pop was procd by h. M.p Daign Un,b o T h. Warid Ban. i boodaries. colors, denonations and nyovher gal ormgatiaofn wh his mp do no' imply. an the par' of T he Wodd Bak Grooup,any' / GUATEMALÅ n. hsa.s. r n nd, HONDURAS AUGUST 19