Report No. 51902-MX MEXICO Agriculture and Rural Development Public Expenditure Review December 2009 Agriculture and Rural Development Unit Sustainable Development Department Latin America and the Caribbean Region The World Bank Document of the World Bank © The World Bank Agriculture and Rural Development Unit Latin America and the Caribbean Region 1818 H Street, NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.wordlbank.org All rights reserved Cover design: Michelle Friedman, World Bank Photos: The World Bank Photo Collection The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions: The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax: 202-522-2422, email: pubrights@worldbank.org. Document of the World Bank Report No. 51902-MX Contents Report Information ..................................................................................................... iii Preface ..........................................................................................................................v EXECUTIVE SUMMARY ....................................................................................... vii 1. Introduction ............................................................................................................1 Purpose of the Study.....................................................................................................1 Overview of The Mexican ARD Policy Context .........................................................3 Performance of the ARD Sector and of Rural Poverty ................................................7 2. Diagnostics of the Federal Rural Budget ...........................................................11 Mexico's Rural Development Budget: The Programa Especial Concurrente (PEC) .11 Classification of ARD Programs ................................................................................14 The PEC Structure ......................................................................................................18 Changes in the 2008 PEC ...........................................................................................24 Other Fiscal Costs Outside PEC .................................................................................26 Relative Size of ARD expenditures............................................................................29 3. Efficiency Assessment ..........................................................................................32 Comparison with Latin American Countries..............................................................32 Comparison with OECD Countries ............................................................................36 APE and Agricultural performance at the State Level ...............................................40 APE and the Performance of Different Crops Over Time .........................................46 4. Equity Assessment ...............................................................................................52 State and Municipal Level Equity Assessment ..........................................................52 Producer and Household Level Equity Assessment ...................................................59 5. M&E and Institutional Aspects of ARD Programs ..........................................66 Monitoring and Evaluation .........................................................................................66 Institutional Issues ......................................................................................................69 6. Options for Moving Forward ..............................................................................73 Key Features of the ARD Program in Mexico ...........................................................73 Policy Options for Reform .........................................................................................75 1) Improving the ARD Planning System ...................................................................75 2) Rationalizing the Overall System for Farmer Support...........................................77 3) Improving the Support System for Small Producers through gradual decentralization ..........................................................................................................84 4) Improving the M&E and Institutional Aspects of ARD Programs ........................86 5) Increasing the Positive Environmental Externalities of ARD Programs and Supporting the National Climate Change Strategy ....................................................88 Appendix to Chapter 6: Technical Notes ...................................................................94 Technical Note 1: New Zealand's Successful Agriculture Policy Reform Experience ..................................................................................................................94 Technical Note 2: INDAP: Technology Transfer Targeting Small Farmers in Chile.... ...................................................................................................96 Technical Note 3: Payments for Environmental Services in Natural Resource Sectors ........................................................................................................................98 7. References ...........................................................................................................102 i List of Boxes Box 2-1: Dimensions in the Decentralization of Rural Development Programs in Mexico .......................................................................................................................... 17 Box 2-2: Efficiency gains from switching from private to public goods ..................... 20 Box 2-3: The role of agricultural subsidies in Mexico's water crisis .......................... 27 Box 6-1: OECD Analysis on Inter-Institutional Horizontal Co-ordination ................. 76 Box 6-2: Challenges Facing the Rural Finance Sector in Mexico ............................... 79 Box 6-3: Proposal by Winters and Davis (2007) on PROCAMPO Reform ................ 80 Box 6-4: Proposal by Sumner and Balagtas (2007) on Ingreso Objetivo Reform ....... 82 Box 6-5: Applying a Territorial Approach to Rural Development in a Decentralization Option ........................................................................................................................... 86 Box 6-6: Rural Proofing in the UK and Canada .......................................................... 87 ii Report Information CURRENCY EQUIVALENTS (Exchange rate effective as of 12/1/2009) Currency Unit: Mexican Peso (MxP) US$ 1= MxP 12.88 FISCAL YEAR January 1 ­ December 31 Abbreviations and Acronyms AOI Agricultural Orientation Index APE Agricultural Public Expenditure ARD Agriculture and Rural Development ASERCA Apoyos y Servicios a la Comercialización Agraria CABSA Program to Develop Environmental Services Markets for Carbon Capture and Biodiversity and to Establish and Improve Agroforestry Systems CC Concentration Coefficient CIDRS Comisión Intersectorial para el Desarrollo Sustentable CLCs Cuenta por Liquidar Certificada CONAFOR Comisión Nacional Forestal CONAGUA Comisión Nacional de Agua CONAPO Consejo Nacional de Población CONEVAL Consejo Nacional de Evaluación ENIGH Encuesta Nacional de Ingresos y Gastos de los Hogares ENNVIH Encuesta Nacional sobre los Niveles de Vida de los Hogares EU European Union FAIS Fondo de Aportaciones para Infraestructura Social FAO Food and Agriculture Organization FAPPA Fondo para el Apoyo a Proyectos Productivos FIRA Fideicomisos Instituidos en Relación con la Agricultura FIRCO Fideicomiso de Riesgo Compartido FISM Fondo para la Infraestructura Social Municipal FONAES Fondo Nacional de Apoyos para las Empresas Sociales GATT General Agreement on Tariffs and Trade GDP Gross Domestic Product GOM Government of Mexico GSSE General Service Support Estimate IDB Inter-American Development Bank INEGI Instituto Nacional de Estadística y Geografía LAC Latin America and Caribbean LDRS Ley de Desarrollo Rural Sustentable M&E Monitoring and Evaluation iii MPS Market Price Support NAFTA North American Free Trade Agreement NRM Natural Resources Management OECD Organization for Economic Co-operation and Development PDR Programa de Desarrollo Rural PEC Programa Especial Concurrente para el Desarrollo Rural Sustentable PER Public Expenditure Review PES Payment for Environmental Services PROCAMPO Programa de Apoyos Directos al Campo PROGAN Programa de Estímulos a la Productividad Ganadera PRONASOL Programa Nacional de Solidaridad PSAB Pago por Servicios Ambientales del Bosque PSAH Pago por Servicios Hidrológicos Ambientales PSE Producer Support Estimate RDE Rural Development Expenditure SAGARPA Secretaria de Agricultura, Ganadería, Desarrollo Rural, Pesca y Alimentación SCT Secretaria de Comunicación y Transportes SECON Secretaria de Economía SEDESOL Secretaria de Desarrollo Social SEMARNAT Secretaria de Medio Ambiente y Recursos Naturales SFP Secretaria de Función Publica SHCP Secretaria de Hacienda y Crédito Publico SMA Salarios Mínimos Agropecuarios SRA Secretaria de Reforma Agraria TFP Total Factor Productivity TSE Total Support Estimate USA United States of America iv Preface At the request of the Secretaria de Hacienda y Credito Publico (SHCP), the World Bank has produced this ARD PER that analyzes Mexico's Rural Development federal budget, known as the Programa Especial Concurrente (PEC). The analysis focuses on the categorization of PEC, efficiency aspects using international comparison as well as state level comparisons, and equity aspects at the state and household levels. The core team for this report comprised John Nash (Lead Economist, LCSSD), Jose Maria Caballero (FAO consultant), Christian Borja-Vega (Consultant), Jozef Draaisma (Senior Country Economist) and Yurie Tanimichi Hoberg (Task Team Leader). Dianelva Montas and Erika Salamanca (LCSAR) handled logistics in Washington DC, Alejandra Gonzales and Rosa Maria Hernandez-Fernandez handled mission logistics in Mexico City, Maria Claudia Pachon and Diana Cubas edited the Spanish translation and Michelle Friedman assisted finalization and publication of the report. The team benefitted from the detailed comments by Steven Webb (peer reviewer), Dina Umali-Deininger (peer reviewer), Mona Sur (peer reviewer), Evelyne Rodriguez (external peer reviewer), Ethel Sennhauser (Sector Manager, LCSAR), David Rosenblatt (Sector Leader/Lead Economist, LCSPR), Gustavo Saltiel (Sector Leader), John Scott (Professor, CIDE), Stefano Pagiola (Senior Environmental Economist, ENV) and Susana Sanchez (Senior Financial Economist, AFTSN). Technical background papers for the report were prepared by Luis Gomez Oliver (Consultant), John Scott (Consultant/Professor, CIDE), and Alberto Valdes (Consultant). In addition, Stefano Pagiola (Senior Environmental Economist, ENV), Ricardo Hernandez (Senior Environmental Specialist, LCSEN), Robert Davis (Senior Forestry Specialist, LCSAR), Gerardo Segura (Senior Rural Development Specialist, LCSAR), and Harideep Singh (Senior Rural Development Specialist, LCSAR) also provided technical input into the report. The team worked under the overall guidance of Mr. Carlos Treviño (Subsecretario de Egresos, SHCP) and received technical guidance from SAGARPA through Mr. José Luis López Díaz Barriga (Official Mayor). v vi MEXICO Agriculture and Rural Development Public Expenditure Review EXECUTIVE SUMMARY =============================================================== 1. Mexico's large ARD program represents a very significant fiscal effort by the Government of Mexico (GOM) on behalf of the rural population. Average public expenditure per capita is now similar in the urban and rural sectors, unlike that in many Latin American countries, where there is still an "urban bias" in the allocation of public expenditure. The present day Agriculture and Rural Development (ARD) program represents the outcome of the far-reaching reform efforts which began in the late 1980s to modernize the sector and introduce a more efficient, equitable and less distortionary policy environment. 2. Mexico's ARD policy reforms included many novel features signaling a clear departure from past policies. They resulted in some international best practices such as the success of the Oportunidades program in rural poverty reduction, the decoupled design1 of PROCAMPO, and the strong global leadership of the GOM in mainstreaming climate change in ARD issues. This report offers options and alternatives for the GOM to move forward on further improving its ARD public expenditure program. 3. There are four reasons why the study is timely. First, 2008 marked the end of the North American Free Trade Agreement (NAFTA) transition period and signaled the full integration of the North America markets. Trade liberalization and NAFTA were powerful influences on recent Mexican Agriculture and Rural Development (ARD) policies, since NAFTA was conceived and negotiated in an environment in which there was some fear that reduction of trade barriers would create pressure on local producers, and in particular would cause a collapse in domestic production of corn, a product deeply engrained in Mexican culture. Thus, the end of the transition is an appropriate time to reflect on these policies. Second, the current global food price crisis and the long-term prospect of high prices for agricultural goods provides heightened social and political attention to reassessing the main tenets of Mexican ARD policy and revisiting the main programs.2 Third, the growing concern of the GOM for the sustainable use of natural resources, in particular water resources, and for the impacts of climate change calls for analysis of the budgetary implications of measures to improve 1 Decoupled payments are defined as those whose amount is not dependent on current production, input use or prices, and thus is expected to have minimal impact on production decisions. 2 While commodity prices have fallen since their peaks early in 2008, they remain high relative to levels of recent years, and in general this seems to be the most probable scenario for the future. Recent events are a reminder that global commodity markets are -- and are likely to remain ­ volatile, underscoring the importance of policies that help producers to increase their competitiveness under a variety of conditions and allow them flexibility to respond to changing circumstances. vii natural resources use and adjustment to climate change. This is particularly important since recent forecasts show that Mexico will be disproportionately affected by climate change, in particular by hurricanes, changes in temperature and precipitation, and increased frequency and severity of floods and droughts.3 GOM adopted a National Climate Change Strategy in 2007 in recognition of this important issue. A detailed analysis of ARD expenditures should help the GOM and other interested parties in all three of these areas. Finally, the recent global financial crisis highlights the importance of efficiency in public expenditure programs as the role of the government is expanding in attempting to stave off a recession. Key Findings of the Report 1. Overall Strategy 4. Total rural expenditure (including universal social services) is estimated to be MxP 382 billion, or a quarter of total public spending and 4 percent of national GDP. This represents a very significant fiscal effort on behalf of the rural population. Given the current share of the rural sector in the national population (24 percent), average public expenditure per capita is now similar in the urban and rural sectors. There is, hence, no "urban bias" in the allocation of public expenditures and subsidies. 5. The current structure of ARD expenditures in Mexico is largely the result of past decisions to subsidize particular farming sectors, regions, crops, and other aspects of farming. Most subsidies were introduced independently of others because of circumstances at the time; and once introduced, the subsidies have proved difficult to phase out. The resulting system is complex and poorly oriented toward a clear long-term strategy or apparent socioeconomic rationale. 6. The Programa Especial Concurrente has not worked as a coordinating mechanism of ARD programs as intended. The Sustainable Rural Development Law of 2001 established a coordinating framework for ARD public expenditures through the Comisión Intersectorial para el Desarrollo Sustentable (CIDRS), headed by SAGARPA. CIDRS was expected to serve as a forum to coordinate GOM's ARD activities. But a strategy has not been developed, and CIDRS has not oriented or coordinated federal resource allocations to ARD programs. The 3 Third National Communication to the UNFCCC (November 2006). Increases in temperature: by 2020 projected temperature is expected to increase 0­2.5ºC in the winter (December­February) and 0.9­2.2ºC in the summer (June­August). Reduction in precipitation: Rainfall is expected to decrease by up to 15 percent in the central part of Mexico and by less than 5 percent near the Gulf of Mexico, mainly between January and May; by 2020 projected precipitation fluctuations will be in the range of ­7 to +12 percent (December­ February) and ­8 to +12 percent (June­August). Increased frequency and intensity of extreme weather events: the number of severe storms and the intensity of severe drought are also expected to increase; sea water temperature is expected to increase 1­2ºC, leading to stronger and more intense tropical hurricanes in the Caribbean Sea, the Gulf of Mexico, and the Pacific Ocean near Mexico, with a 6 percent expected increase in wind intensity and a 16 percent expected increase in precipitation within 100 kilometers from the center of a hurricane; cold fronts may become less frequent. viii main instrument for CIDRS to coordinate ARD public expenses and activities is the Programa Especial Concurrente (PEC). But PEC has never functioned as a planning tool; rather it operates as an annual budget exercise for tabulating the budget lines of all rural development programs together. PEC appears as an annex of the annual federal budget. 7. Some current ARD programs are incompatible with the far-reaching and ambitious climate change goals in Mexico's 2007 National Climate Change Strategy. 4 Mexico is the only developing country to have submitted three National Communications to the United Nations Framework Convention on Climate Change (UNFCC), demonstrating a strong commitment and leadership role in the international climate change agenda. However, certain major ARD programs counteract these goals--for example, the electricity subsidy for farmers to pump groundwater (tarifa 9), which has led to the overexploitation of many aquifers in water-scarce regions, and the Ingreso Objetivo program, which encourages farmers to produce low-value crops and even water-intensive crops (fodder) using scarce water resources. 2. Rural vs. Agricultural vs. Social Objectives in the Budget Productive public goods programs (such as infrastructure, research and development, and the like) are underfunded. Two-thirds of productive programs are private goods programs (subsidies to individuals and families, cash transfers, and so on), and less than one third of spending in public goods goes to productive programs. While nearly all social programs are targeted to the poor (including small farmers and vulnerable groups), very few agriculture programs are. More than 95 percent of social program expenditures are targeted to the poor, but only 8 percent of agricultural programs are. In contrast to the importance of the rural non-farm sector, PEC reflects an agro-centered view of rural development. Some 70 percent of spending in productive programs is specifically for agricultural activities. Productive programs not targeted to agriculture receive only part of the remaining 30 percent, since most of these programs support agricultural activities as well. Agricultural programs are skewed toward private goods, while nonagricultural rural programs emphasize public goods. More than 80 percent of agricultural programs support private goods, while less than 25 percent of nonagricultural programs do. Certain programs, such as tarifa 9, have high environmental costs that generate huge negative externalities. These externalities are expected to 4 The National Climate Change Strategy was formulated in 2007 and identifies opportunities to reduce emissions on a voluntary basis as well as measures for national and local capacity for response and adaptation. The strategy proposes concrete adaptation and mitigation measure for all sectors, including agriculture, covering all the main aspects of climate change policy. Climate change strategies and action plans have also been developed at the subnational level for Mexico City and the states of Veracruz and Nuevo Leon (World Bank 2008b). ix become more serious as natural resources are increasingly exposed to climate change. 3. Efficiency Assessment The allocation of agricultural spending by states is correlated to the size of their agricultural sectors; when normalized by the size of the sector, it is not related to the growth of their agricultural sectors. That growth per peso spent differs substantially across states suggests that reorienting agricultural public expenditure to states where it has the highest impact would increase the growth of the sector countrywide for a given level of public spending. Reallocating spending from private to public goods would allow for further increase in efficiency in agricultural spending. Regression results show that a 10 percent increase in agricultural public expenditure on private goods as a percentage of the value of agricultural production is associated with a 2.6 percent reduction in agricultural GDP growth. On the other hand, spending on public agricultural goods shows a positive, although not statistically significant, impact on agricultural GDP growth.5 4. Equity Assessment ARD spending (which comprises agricultural spending and rural development spending) is inequitable. Simulations show that the distribution of ARD expenditures is flat for the poorest 70 percent of rural households, which receive less than MxP 500 per capita per month in ARD subsidies. However, the richest 10 percent receive on average more than MxP 3,000 per capita per month. Taken alone, rural development spending is progressively distributed, with the poorest 20 percent of rural households receiving 33 percent of benefits. In contrast, agricultural spending is extremely regressive, with more than half of spending concentrated in the richest decile. Agricultural spending is so regressive that it cancels out about half the redistributive impact of rural development spending. Rural development programs decrease the Gini coefficient (reduce inequality) by about 14.2 percent, while agricultural spending increases the Gini coefficient (raises inequality) by about 6.7 percent. 5 Using a 1985­2001 dataset on rural expenditure for LAC, Lopez and Galinato (2007) also show that reducing the share of subsidies to private goods (or, equivalently, increasing the share of public goods) in the government's budget has, ceteris paribus, a large and significant positive impact on rural per capita income. x 5. M&E and Institutional Aspects of ARD Programs M&E of ARD programs has greatly improved recently, but important challenges remain. These include carrying out evaluations not only of single programs but also of the entire PEC, developing and integrating databases of ARD program beneficiaries, conducting deeper program evaluations at longer intervals instead of the current annual cycle, and systematically using evaluation results for managerial and budgetary decision-making. Implementation matters--and it is particularly complex in production- oriented programs. How programs are implemented is almost as important as how they are designed. Key aspects to take into consideration for implementation include the three-year, no reelection electoral system of municipal authorities, which hinders their involvement in medium- to long- term endeavors like ARD; the year-based budgetary system in Mexico, which does not favor the continuity of programs and views required by ARD and does not offer security to beneficiaries regarding program stability; administrative and budgetary norms that make it difficult to disburse program funds and limit disbursement to several months of the year, thus placing big pressure on those executing the program;6 the scarcity of recurrent funds for program operation; and the poor economic compensation and weak morale for bottom-level program operators, which is inconsistent with the importance of their function. Options for Moving Forward 8. To make well informed decisions, the GOM will need to weigh the tradeoffs of policy options based on considerations such as cost, sequential importance, technical difficulty, risks, and impact. While some options promise greater benefits in the longer run, others could be targeted for quick action. Policy options are reviewed in the following five areas: 1) Improving the ARD planning system. 2) Rationalizing the overall farmer support system. 3) Improving the support system for small producers through gradual decentralization. 4) Improving the M&E and institutional aspects of ARD programs. 5) Increasing the positive environmental externalities of ARD programs and supporting the national climate change strategy. 1) Improving the ARD Planning System 6 Many agricultural public expenditure programs are very sensitive to the timeliness with which resources reach farmers, especially in areas where the rainy season is short. Operational and budgetary norms often stand in the way of timelines, resulting in the late arrival of subsidies. xi 9. A more effective planning system for ARD spending could be put in place. The guidelines for allocating budgetary resources and the authority and planning capacity of the CIDRS would have to be strengthened for PEC to be effective. A recent OECD review of rural development issues in Mexico (OECD 2007c) highlights the challenge faced by Mexico where SAGARPA, a sectoral ministry without sufficient authority over peer sectoral ministries is mandated to plan the ARD expenditures as chair of CIDRS. The OECD proposal to move CIDRS from SAGARPA to the President's office would better reflect the multisectoral character of rural development, and strengthen the Commission's planning capacity. Another option is to place CIDRS under SHCP because SHCP is primus inter pares as a ministry, has a multisectoral mandate and view, and is responsible for looking into the effectiveness and quality of public expenditure and preparing the federal budget. In addition to strengthening CIDRS, a strong nucleus of agricultural policy analysts, staffed with experts of recognized academic excellence could be formed to operate autonomously under the SAGARPA. 10. The planning system for ARD spending should be based on a comprehensive national ARD strategy. Without a comprehensive ARD strategy and planning system for ARD expenditures it is difficult to achieve coherence among ARD programs and their objectives. A comprehensive ARD strategy with consistent, explicit, and selective objectives as well as quantifiable results and performance indicators would be an important step toward better ARD policy. The strategy could serve as a framework for individual programs, which would be assessed by their contribution to its objectives. The strategy would also serve as a framework for the M&E of individual programs. Along with the strategy, an ARD budgetary planning system would be required to ensure consistency between the strategy and budgetary allocations. PEC has led to substantial progress in this direction. But so far PEC provides only the framework--not the content--of such a system. 2) Rationalizing the Overall System for Farmer Support 11. Federal public expenditures in productive programs should be reoriented in three ways. Gradually discontinuing market price support (MPS) to inputs and outputs, as well as compensatory cash transfers. Substantially increasing allocation to public goods such as irrigation improvement and management; drainage and transport infrastructure; other rural infrastructure (such as that provided by FAIS); research, extension, and training; sanitary and phyto-sanitary services; market intelligence (including collecting weather information through weather stations and agricultural statistics); emergency programs; and other public goods. Increasing support to small producers. xii 12. Non-targeted production-oriented programs benefiting mostly commercial farmers should be separated from productive and natural resource management (NRM) programs targeting small producers, vulnerable groups, and the poor which should be decentralized to the states. Resources for non- targeted production-oriented programs could be maintained at the federal level, but funds now spent by different ministries for targeted programs could be pooled. The pooled funds could be distributed to state governments as block grants according to an objective formula to allow state governments to create their own programs that promote small producers. In addition, given the current imbalance in favor of non-targeted programs, some resources could be moved from non-targeted programs to the pooled fund. These decentralized funds could also support programs that assist small producers in adapting to climate change and increasing positive externalities in NRM programs. 13. Federal non-targeted programs that support on-farm and value chain investments should be rationalized. These programs include PROGAN, Apoyos a la Competitividad de las Ramas Productivas/Adquisición de Activos Productivos, and the non-targeted programs of Alianza. One option is to merge these types of program into a new non-targeted support program for on-farm and agriculture value chains investments. Another is to maintain different programs but redesign them and perhaps eliminating some to avoid duplications, sharpen their objectives, and make them more effective and results oriented. 3) Improving the Support System for Small Producers through gradual decentralization 14. Gradual decentralization to state governments of the mandate and resources to support small producers would help address challenges. There are several reasons for this: Decentralization would facilitate a territorial approach for rural development, which would help clustering investments. The number, dispersion, and heterogeneity of small producers favor responsibilities at the subnational level, where constraints and opportunities can best be identified. The subnational level is also more appropriate for attaining synergy between productive programs and other investments and services (such as NRM, infrastructure or education) favoring small producers. Resources could be more rationally used at the subnational level if state governments apply them according to state-level strategic plans for the development of the small producers. 15. After the transition period, decentralization would imply transferring to state governments, in addition to the present FAIS to finance small rural infrastructure, block grants for the decentralized development of small rural producers (in productive and NRM activities). This, in turn, implies (i) phasing out corresponding federal ARD programs and allowing state governments to xiii replace them with their own programs, so that they can exert strong ownership of the rural development agenda for the development of small producers in their states; (ii) pushing ahead with administrative federalization (federalización administrativa) to transfer to state governments the assets, staff, and other operational resources of federal ministries required for state governments to properly operate the newly decentralized funds; and (iii) strengthening the implementation capacity of state governments to be able to takeover these responsibilities. Decentralization of funds to support small producers would facilitate the use of a territorial approach to rural development. 4) Improving the M&E and Institutional Aspects of ARD Programs 16. Alternatives exist to enhance the M&E system of ARD expenditures. There have been recent important advances in the M&E of rural programs. There are, however, still remaining important areas for improvement. These include carrying out evaluations not only of single programs but also of the entire PEC, the development and integration of databases of the beneficiaries of ARD programs, deeper program evaluations at longer intervals instead of the current annual cycle, and systematic use of evaluation results for managerial and budgetary decisions based on action agendas agreed upon by the evaluators, the implementing agency, and a third party. Adherence to these agendas should be monitored closely. As a logical follow on to the Ley de Desarrollo Rural Sustentable, Mexico could also start a "rural proofing" system to anticipate and monitor national policies' impact on the rural sector. Canada and the United Kingdom both found success with rural proofing systems (called "rural lens" in Canada). 17. Alternatives exist to improve program implementation. Implementation challenges are complex and varied. Many are institutional limitations that take time to resolve. Simplifying the number and type of programs would certainly help respond to the implementation challenge, as would a comprehensive ARD strategy. To help identify and address these issues, CONEVAL's process evaluations could be widely used to maintain the focus on implementation challenges. Issues that could be addressed include: Revising the budgetary and administrative regulations under which ARD programs are implemented to ensure that they provide sufficient flexibility and do not unnecessarily impair or slow down implementation. Paying due attention to the recurrent/operational cost needs of programs, which are particularly large for production-oriented programs. Improving dissemination of programs and the accountability of operators vis- à-vis beneficiaries and using "client satisfaction" criteria to assess performance. Improving the economic incentives and morale of program operators and ensuring that incentives are well aligned with program objectives. xiv 5) Increasing the Positive Environmental Externalities of ARD Programs and Supporting the National Climate Change Strategy 18. Strategies for supporting agriculture and for dealing with climate change need to be mutually reinforcing. For the past two decades over 80 percent of economic losses from weather-related disasters occurred in the agriculture sector. Agricultural policies and climate change policies thus need to be mutually reinforcing by incorporating two principles. First, farmers need to respond to localized changes in climate depending on their individual conditions and constraints. Policies should expand their options rather than encourage them to make choices that do not correspond well to individual circumstances. Second, farmers should not be encouraged to "maladapt" by overusing scarce resources or increasing production in high-risk areas. Both principles have implications for agricultural policies that are congruent with the National Climate Change strategy. 19. Expanding payments for environmental services (PES) programs beyond forestry could create positive synergies between agricultural production and the environment. Specific investment programs to reduce emissions that are identified by the National Climate Change Strategy and that could be scaled up include programs to increase the use of biomass, e.g. through high-efficiency wood burning stoves in rural communities and renewable energy sources for rural areas as supported by the Fideicomiso de Riesgo Compartido (FIRCO), livestock programs to rehabilitate degraded rangelands, and hydro-metrological hazard risk management and water resources management. Scaling up these programs could be accomplished either through national programs or through decentralized programs. Such scaling up should be accompanied by appropriate monitoring measures to ensure that the programs are responding to their environmental objectives. In particular, there is considerable scope to replace distortionary programs such as Ingreso Objetivo with targeted environmentally-friendly PES programs, which would reduce distortions and provide income support to small farmers while also providing valuable national and global externalities. Summary Matrix of Options 20. In order to make well-informed going forward, the GOM will need to weigh the trade-offs among the various criteria of policy options such as cost, cost, sequential importance, technical difficulty, risks, and impact. The following table summarizes the policy options presented in the study, based on some of these criteria. The matrix will make clear that while there are options which promise greater benefits in the longer run, there are also areas that the government could target for quick action. xv Table: Summary Matrix of Policy Actions Policy Options Implement Likely Technical Fiscal Cost or ation Impact Difficulty/ Savings Risks 1. Improving the ARD Planning (1) Prepare and approve a national Short High Medium Low Cost ARD strategy (2) Reposition CIDRS within the federal government to strengthen its coordination capacity, and build an Short Medium High No Cost or Savings effective ARD budgetary planning system (3) Create a nucleus of high-level agriculture policy analysts under Short Medium Low Low Cost SAGARPA umbrella. 2. Rationalizing the overall system for farmer support (1) Increase expenditure allocation to public goods: irrigation improvement and management, drainage and transport infrastructure; other rural Short, Medium infrastructure (such as that provided and Long High Medium High Cost by FAIS); research, extension and training; sanitary and phyto-sanitary services; market intelligence; emergency programs; and others (2) Modify PROCAMPO Short Medium Medium No cost or saving (3) Discontinue PROCAMPO Medium High High High Saving (4) Modify Ingreso Objetivo Short Medium Medium No cost or saving (5) Discontinue Ingreso Objetivo Medium High High High Saving (6) Modify input support programs Short Medium Medium No cost or saving (tarifa 9, agrodiesel, others) (7) Discontinue input support High Saving programs with or without temporary Short High High compensatory payments (tarifa 9, agrodiesel, others) (8) Rationalize federal non-targeted programs oriented to support on-farm and value chain investments (mainly PROGAN, Apoyos a la Short, Medium Competitividad de las Ramas High High Low Cost/No Cost or Productivas/Adquisición de Activos Savings Productivos, and the non-targeted programs of Alianza). (9) Reduce the allocation to non- Short, Medium targeted production oriented programs and Long High Low Medium to allow for an increase in expenditure Saving to support small producers (see 3 (1))* (10) Increase support to the rural non- Short, Medium Medium Cost farm sector and Long High High (11) Strengthen the rural finance Short, Medium Cost sector to increase service access Medium, and High High xvi Long 3. Improving the support system to small producers (1) Increase expenditure allocations to Short, Medium support small producers* and Long High Low Medium Cost (2) Pool resources from targeted production/NRM oriented programs of Medium, Long different federal entities, decentralize those resources to state governments High High No Cost or Saving according to an agreed distribution formula and decentralization protocol, and discontinue the corresponding federal programs (3) Accelerate the decentralization of Short, ARD federal offices (federalización Medium, and Medium High Low Cost administrativa) Long (4) Agree with state governments on a protocol containing guidelines for decentralization of funds to support small producers: (i) national No Cost or Saving benchmarks, (ii) fund distribution Short, Medium Medium Medium formula, (iii) M&E system, (iv) implementation support system, and (v) state government responsibilities 4. Improving the M&E and the implementation of programs (1) Adjust budgetary and administrative regulations of ARD No Cost or Saving programs to ensure flexibility and Short Medium High timely execution (2) Increase recurrent cost allocations Short, Medium for production oriented programs and Long Medium Medium Medium Cost (3) Improve the dissemination of ARD Short, Medium programs and Long Medium Low Low Cost (4) Improve the accountability of Short, Medium program operators at all levels and and Long High Medium Low Cost introduce "client satisfaction" criteria to assess performance (5) Improve the economic and moral Short, Medium incentives of medium and bottom level and Long High Medium Medium Cost program operators (6) Carry out evaluations of the entire Short, Medium Medium Low/Medium Low Cost PEC and Long (7) Develop and integrate databases of Short, Medium Medium Medium Low Cost ARD program beneficiaries and Long (8) Carry out in-depth program Short, Medium Medium Low Low Cost evaluations at less frequent intervals and Long (9) Organize a system to agree on and monitor action agendas Short Medium Low Low Cost deriving from evaluation results 5. Increasing environmental externalities of ARD programs and supporting the national climate change strategy (1) Improve SAGARPA norms to regulate slash and burn agriculture in Short Medium Low No Cost or Saving light of increasing forest fires xvii (2) Increase research allocation to research and technology transfer of reducing emissions from agriculture Short, Medium Medium Medium Cost such as no-till technology, reducing and Long High GHG gas from livestock, reforestation, efficient use of inputs (3) Revise regulations to incentivize farmers to sell electricity generated by Medium Medium High Medium Saving biomass to the grid^ (4) Improve targeting of geographical area and payment levels of payment for Short, Medium High Medium High Cost environmental services programs, and and Long scaling up its operations (5) Replace distortionary support Short, Medium programs with payments for High Medium No Cost or Saving and Long environmental services Implementation term: Short (< 2 years), generally associated with changes in laws or regulations; Medium (3­5 years), requiring some institutional changes; Long (> 5 years). Effects term: Short (< 2 years); Moderate (3­5 years); Long (> 5 years). Impact: High, medium, low (relative to other policy options proposed in these notes, not relative to other more general reform options). Technical difficulty/risks of undertaking the options suggested: High, moderate, low. Fiscal cost to government: High (large public investment programs), moderate (some public program expenditure required), low (little public expenditure). Fiscal saving: high, moderate, low. Note: * = The suggested proposal is to shift resources from non-targeted production oriented programs to targeted ones (eventually decentralized to state governments) to support small farmers. Thus, the net effect should be budget neutral, i.e. the overall expenditure should not increase or decrease. ^ = The suggested proposal is expected to increase investments by farmers to generate biomass energy. Although, this proposal per se is budget neutral, in the long-run this should result in fiscal cost saving since as farmers begin to generate biomass energy, they would buy less subsidized electricity from the grid. xviii 1. INTRODUCTION PURPOSE OF THE STUDY 1. This study examines agricultural and rural development (ARD) public expenditures in Mexico. The study is based on federal public expenditures. 7 Included in ARD spending are all the public programs considered as such by the Secretaria de Hacienda y Crédito Publico (SHCP), i.e. compiled in the Programa Especial Concurrente para el Desarrollo Rural Sustentable (PEC). 8 For completeness, some policies and programs benefiting the rural sector that do not result in budgetary expenditures, but in forgone government income, namely tax exemptions and discounts to farmers on electricity bills (tarifa 9) are also considered. However, the study does not address trade policies, another important government intervention that changes relative border and domestic prices.9 2. In recent years, a growing set of evaluations has become available for the principal ARD programs in Mexico, and a number of broader evaluations of ARD policies have been produced (OECD 2006a, 2007d, WB 2005a, 2005b, 2006a, and 2008, Taylor et. al. 2007). The World Bank has also produced a number of public expenditure reviews (PER) and sub-sector studies in Mexico: Guanajuato state-level PER [2002], Veracruz state-level PER [2003], Mexico Country PER [2004a], Infrastructure PER [2005c], Water PER [2007a], Access to Finance for Farmers and Agricultural Enterprises [2004b], Mexico's Competitiveness: Reaching its Potential [2006b], Policy Note on Sensitive Crops (Maize, Sugar and Beans) [2007b]). The present assessment builds on this rich set of results. It aims to provide a comprehensive appraisal of the overall ARD public spending strategy implemented in Mexico, in terms of both the scope of the programs considered and criteria of evaluation. Given the wide range and dispersion of ARD spending in many, largely uncoordinated programs, as reported in the budgetary analysis, this assessment focuses on the principal strategic programs and groupings of programs, and their overall impact on ARD. 7 The study team made an effort to collect ARD expenditure data from state governments through a questionnaire, but the number and quality of questionnaire returns did not allow the compilation of suitable estimates of public expenditure in ARD from state governments' own budgets. Since ARD spending from state governments' own budgets are known to be rather limited, it is unlikely that their omission affects the main findings of this study. 8 The PEC does not necessarily include the budget of all activities pertaining to rural development, but this classification will be used for this study since this is what the Government of Mexico has identified as the federal budget for rural development. 9 For a review of distortions caused by public agricultural policy including trade policies, see World Bank (2008c). Also, OECD (2006a) provides a general overview of both the trade and public expenditure policies of Mexico. 1 3. There are at least four reasons why the study is timely. First, year 2008 marked the end of the North American Free Trade Agreement (NAFTA) transition period and the full integration of the North American markets. Trade liberalization and the signing of NAFTA was a powerful influence in shaping recent Mexican ARD policies, since NAFTA was conceived and negotiated in an environment in which there was some fear that reduction of trade barriers NAFTA would create pressure on local producers, and in particular would cause a collapse in domestic production of corn, a product deeply engrained in Mexican culture. Thus, the end of the transition period is an appropriate moment to reflect on these policies. Second, the current global food price crisis and the long term prospect of a high price scenario for agricultural goods provides heightened social and political attention to reassessing some of the main tenets of current Mexican ARD policy and revisit some of the main programs. 10 Third, the growing concern of the Government of Mexico (GOM) for the sustainable use of natural resources, in particular water resources, and for the impacts of climate change calls for an analysis of the possible budgetary implications of measures aimed at improving natural resources use and adjusting to climate change. This is particularly important since recent forecasts show that Mexico will be disproportionately affected by climate change, in particular by hurricanes, changes in temperature and precipitation, and increased frequency and severity of floods and droughts.11 GOM has adopted a National Climate Change Strategy in 2007 in recognition of the importance of this issue. A detailed analysis/assessment of ARD expenditures should help the GOM and other interested parties in all three of these areas. Finally, the recent global financial crisis highlights the importance of efficiency in public expenditure programs as the role of the government is expanding in attempting to stave off a recession. 4. The study is structured in six parts as follows: The first part presents the Mexican ARD context in terms of policy and performance. The second part dissects the ARD public budget, classifying expenditure programs in various ways so as to provide an overview of the scope and composition of ARD spending. The third 10 While commodity prices have fallen since their peaks early in 2008, they remain high relative to levels of recent years, and in general this seems to be the most probable scenario for the future. Recent events are a reminder that global commodity markets are -- and are likely to remain ­ volatile, underscoring the importance of policies that help producers to increase their competitiveness under a variety of conditions and allow them flexibility to respond to changing circumstances. 11 Third National Communication to the United Nations Framework for Climate Change (November, 2006). Increases in temperature: by 2020 projected temperature is expected to increase in the winter (December- February) between 0 and 2.5 C and in the summer (June ­ August) in the range of 0.9 and 2.2 C 11 . Reduction in precipitation: the rainfall is expected to decrease by up to 15% in the Central part and by less than 5% in the area of the Gulf of Mexico, mainly between January and May; by 2020 projected precipitation fluctuations will be in the range of -7 to +12% (December-February) and -8 to +12% (June- August). Increased frequency and intensity of extreme weather events: the number of severe storms and the intensity of periods of severe drought is also expected to increase; the sea water temperature is expected to increase between 1 and 2 C leading to stronger and more intense tropical hurricanes in the Caribbean Sea, the Gulf of Mexico and the Mexican portion of the Pacific Ocean, with an increase of 6% in wind intensity and an expected increase in precipitation of 16% within a radius of 100km from the center of the hurricane; the cold fronts may become less frequent. 2 part analyses efficiency issues by means of international and state-level comparisons of agriculture public expenditures (APE) and agriculture performance indicators. The main focus of the assessment is on APE, but these are evaluated within the broader context of rural development policies, considering the degree of consistency and complementarity between the two sets of policies. 12 The fourth part investigates equity issues emerging from ARD expenditures. Equity is evaluated at the level of geographic units (states and municipalities) and of individuals and households. In addition to assessing the equity impacts of the main rural programs, an effort is made to evaluate the overall equity implications of the entire APE and of rural development expenditure (RDE). The fifth part contains a brief analysis of the Monitoring and Evaluation (M&E) system of ARD programs and of selected institutional aspects, and highlights some issues. Finally, the sixth part offers reflections on possible directions for moving forwards both for the overall evolution of ARD expenditures and for some of the main programs. 5. The current structure of ARD expenditures in Mexico is largely the result of an accumulation of past decisions of subsidizing particular farming sectors, regions, crops or aspects of farming. Each subsidy program was largely decided independently of others because of circumstances at the time; and once introduced, subsidies have proved difficult to phase out. The resulting system is complex and does not seem to respond to any clear long-term strategy or socioeconomic rationale. Thus, the aim of the study is to assist in the assessment of the current overall ARD expenditure structure. There is a rich discussion in Mexico concerning how specific ARD programs could be reformed and improved. Although the study enters that discussion in the final chapter on policy options, its main contribution is in the analysis of the global structure of ARD expenditures and their efficiency and equity, and in providing guiding principles and specific suggestions for a comprehensive reassessment. OVERVIEW OF THE MEXICAN ARD POLICY CONTEXT 6. The principal ARD policies currently implemented in Mexico originated in the context of a broad, market-orientated reform effort to modernize the agricultural sector in the late eighties and early and mid-nineties. This included accession to the General Agreement on Tariffs and Trade (GATT) in August 1986, constitutional reform of the land tenure system designed to liberate agricultural land markets in 1992, progressive deregulation of state-trading in agriculture products and inputs throughout the 1990s and the dismantling of the 12 The study distinguishes between three types of public expenditures affecting rural areas: agriculture and rural development (ARD) expenditures; agriculture public expenditures (APE); and rural development expenditures (RDE). APE refers to those programs that more directly affect agricultural production. RDE are expenditures in rural programs that do not affect agricultural production or only in an indirect way. ARD expenditure is the sum of both. Classification of programs is often difficult because of the gray area between APE and RDE, particularly when the impact on agriculture is indirect, and because many programs include expenditures of both types. The classification of the main programs is presented in the text. 3 government trading company, the Compañía Nacional de Subsistencias Populares, and the opening up of agricultural commodity markets under NAFTA, initiated in 1994, with a long transition period, culminating in the full liberation of maize, beans, sugar and dairy products in 2008.13 7. These reforms aimed to introduce efficient and equitable ARD programs. The "second agrarian reform" (Gordillo et al. 1999), as this ambitious reform effort has been labeled, was accompanied by extensive changes in ARD policies, seeking to introduce more efficient and equitable and less distortionary policy instruments. The long-drawn "first" agrarian reform was accompanied, from the Cardenas administration in the 1940s until its formal ending by President Salinas de Gortary in 1992, by two principal forms of agricultural support: input support (irrigation, fertilizers, stockholding) and market price support (MPS). By design, these support policies where both distortionary and inequitable, failing to reach small farmers who were intended to be the principal beneficiaries of the agrarian reform. 8. The three principal support programs. Farmers were partly compensated for the gradual reduction of MPS through three principal support programs: (i) the Programa de Apoyos a la Comercialización (hereafter Apoyos)14, an output-based subsidy program introduced in 1991, (ii) the Programa de Apoyos Directos al Campo (hereafter PROCAMPO), a per hectare direct transfer program decoupled from production and prices, introduced in 1994, and (iii) Alianza para el Campo (hereafter Alianza), a family of investment support programs offering matching grants and support services, introduced in 1996. The expectation was that these programs would not only play a compensatory role in the face of growing external competition but, in the case of PROCAMPO and Alianza, would also provide the necessary support for farmers to modernize production and switch to higher value crops in the context of the newly liberalized land and product markets. 9. The reform included many novel features signaling a clear departure from past policies. In the context of previous support policies, the decoupled design of PROCAMPO made this program revolutionary in terms of efficiency and equity. By decoupling transfers from production and prices and setting entitlements according to historical area, the program was not only expected to minimize distortions in productive decisions, but also to directly transfer resources to farmers, including subsistence ones, for the first time in Mexico's history. Not 13 The end of the transition period was not the traumatic event that had been feared 15 years earlier, for two reasons. One was that throughout the implementation period, the government had allowed importation at low or zero tariffs of larger quantities than had been required under the agreement. Over the 15 year period, imports of corn (mostly yellow) had grown substantially, but somewhat surprisingly, so had domestic corn production (mostly white). Thus, the formal elimination of the final tariffs did not much change the reality on the ground. And secondly, the final phase occurred in an environment in which global commodity prices ­ particularly corn -- were at historically high levels. This brought high profits for Mexican producers, while provoking a food price crisis for consumers. 14 The Programa de Apoyos a la Comercialización and PROCAMPO are both managed by Apoyos y Servicios a la Comercialización Agraria (ASERCA), an agency within SAGARPA. Apoyos includes a number of sub-programs, the most important of which is Ingreso Objetivo. 4 the whole reform package, however, shared such novelties. In particular, Ingreso Objetivo, the main sub-program of Apoyos, is a MPS mechanism for basic crops such as rice, corn, wheat, sorghum and others, and is essentially no different from the policies of the pre-reform period. 10. The agriculture support programs were accompanied by rural development and poverty reduction programs. The reform in agricultural support policies was accompanied by another reform in rural development and anti-poverty policies. The principal anti-poverty program introduced was the Programa de Educación, Salud y Alimentación (Progresa, in 1997; renamed Oportunidades in 2001), offering direct cash transfers to poor rural households conditional on human capital investment (attending basic education and using health services).15 Three other important targeted rural development programs were introduced in this period: (i) the Fondo de Aportaciones para Infraestructura Social (FAIS, in 1996), a large decentralized fund for basic infrastructural investment replacing the Programa Nacional de Solidaridad (PRONASOL) of the Salinas administration (1988-1994), (ii) the Programa de Empleo Temporal (1995), a multi-agency, self- targeted temporary employment program16 and (iii) the Programa de Desarrollo Rural (PDR, in 1996), the principal Alianza program targeted to small producers. 11. The Rural Development Law of 2001 created a framework to coordinate rural development activities and expenditures in Mexico. Reforms were also accompanied by an important institutional innovation, the passing of an umbrella law for rural development, the Ley de Desarrollo Rural Sustentable (2001). This law introduced a new and comprehensive institutional architecture for agriculture and rural development in Mexico. At the national level, it created a national council for rural development with governmental and civil society participants, and an inter-ministerial rural development committee, the Comisión Intersecre- tarial para el Desarrollo Rural Sustentable (CIDRS), chaired by the Secretaria de Agricultura, Ganadería, Desarrollo Rural, Pesca y Alimentación (SAGARPA), with the participation of various federal government ministries. In addition, mixed government-civil society committees were created at the state, district and municipal levels, and also for the main agricultural product chains and the main ancillary services, although their functions and authority is still irregular and often limited. The rural development law also mandated the creation of a coordinating framework for ARD expenditures, the Programa Especial Concurrente para el Desarrollo Rural Sustentable (PEC). However, beyond offering a budgetary classification scheme to organize ARD expenditures, PEC has not had much impact on the allocation of ARD resources. Coordination of the numerous 15 In 2001 the program was extended to urban areas and upper-secondary education and renamed Oportunidades. 16 Originally, Programa de Empleo Temporal (PET) included the participation of Secretaria de Desarrollo Social (SEDESOL), Secretaria de Medio Ambiente y Recursos Naturales (SEMARNAT), Secretaria de Comunicacion y Transportes (SCT), and SAGARPA, but the SAGARPA component has recently been discontinued. 5 Mexican ARD programs, executed by various ministries and agencies, remains a challenge. 12. ARD spending has doubled since 2000 showing an increase in the most distortionary programs. Since 2000, ARD spending has almost doubled in real terms, reaching a federal ARD budget of MxP 204 billion for 2008 (see Figure 2- 1). This expansion took place in the context of the liberalization of most agricultural products in 2003 and the liberalization of "sensitive" products in 2008. The expansion of APE was in large part allocated to existing and new distortionary instruments such as the agricultural diesel subsidy, which was started in 2001, partly reversing rather than deepening the effect of the market- oriented reforms. 13. Specifically, the use of the most distortionary instruments fell in the 1990s, while the use of the least distortionary instruments rose, but this positive trend was reversed in recent years. Overall, the reforms led to a sharp reduction in the participation of the most distortionary instruments (MPS, output and variable input payments) with the combined share of the latter two in APE declining from 50 to 20 percent between 1990 and 1996 (see Figure 1-1).17 On the other hand, the share of the least distortionary instruments (public goods and payments based on historical entitlements) increased from 30 to 70 percent in the course of the decade. Since 2000, however, these trends have been reversed, with the more distortionary instruments gradually gaining ground and the least distortionary losing it. Figure 1-1: Least and Most Distortionary APE (% Total APE) Least distortionary APE 80% 70% 60% 50% 40% 30% 20% 10% 0% 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Payments based on noncurrent A/An/R/I Public goods (GSSE) 17 The sharp fall in payments based on variable inputs in 1997-2000 may be due to inconsistencies in the OECD data. The fall in 1997 corresponds to a drop in the electricity subsidy for groundwater pumping for irrigation, while the sharp increase in 2001 corresponds to the introduction of large agricultural diesel subsidies. The new output-based payments in 2001 correspond to the introduction of Ingreso Objetivo (ASERCA), with large subsidies for grains and other crops in 2001. 6 Most distortionary APE 60% 50% 40% 30% 20% 10% 0% 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Payment based on variable inputs Payments based on output Note: "Non-current A/An/R/I" refers to transfers linked to historical area planted/animal numbers/receipts/income. For further details see OECD (2007a,b) Source: OECD (2007) as cited in Scott (2008) 14. Some of the current ARD programs are incompatible with Mexico's far- reaching and ambitious climate change goals as represented in the National Climate Change Strategy adopted in 2007.18 Mexico is the only developing country to have submitted three National Communications to the United Nations Framework Convention on Climate Change, demonstrating a strong commitment and leadership role in the international climate change agenda. However, certain major ARD programs counteract these goals, for example the electricity subsidy for farmers to pump groundwater (tarifa 9), which has led to overexploitation in many of the country's aquifers in water scarce regions, as well as the Ingreso Objetivo program, which encourages farmers to produce low value crops using scarce water resources and even crops that are water intensive, as is the case with fodder. PERFORMANCE OF THE ARD SECTOR AND OF RURAL POVERTY 15. Overall, agricultural growth has been weak but has shown signs of acceleration since 2000. Between 1980 and 2007 agricultural GDP grew 1.6 percent a year on average, lagging behind total GDP growth of 2.7 percent (see Figure 1-2). This trend is unsurprising, reflecting the expected contraction of the share of the primary sector in a developing economy. The relative difference between the two growth rates, however, has narrowed rather than widened in 18 The National Climate Change Strategy was formulated in 2007 and identifies opportunities for emissions reductions on a voluntary basis, as well as measures for the development of necessary national and local capacity for response and adaptation. The strategy proposes concrete adaptation and mitigation measure for all sectors, including agriculture, covering all the main aspects of climate change policy. Climate change strategies and action plans have also been developed at the sub-national level for Mexico City, the states of Veracruz and Nuevo Leon (World Bank, 2008b). 7 recent years, after 2000. In 2001 and 2003, the growth rate of agriculture GDP was actually above that of total GDP (3.5 percent and 3.1 percent vs. -0.2 percent and 1.3 percent, respectively). This, together with the stability of basic food prices and Oportunidades transfers, is widely credited for the significant reduction in rural poverty achieved during the recessionary 2000-2002 years (Székely and Rascon, 2004) and the overall 2000-2006 period. Figure 1-2: Average Annual Growth Rates of Total GDP and Agriculture GDP (%, 1980-2007) 4.5% 4.0% 4.0% 3.7% 3.6% 3.5% 3.2% 3.0% 2.6% 2.6% 2.5% 2.0% 2.0% 1.6% 1.7% 1.5% 1.0% 0.8% 0.5% 0.0% 19801989 19901993 19941999 20002004 20052007 GDP Agric. GDP Source: INEGI as cited in Scott (2008) 16. The rural poverty rate has decreased but the rural share in total poverty remains unchanged since 1992. The last decade has witnessed a significant and sustained decline in rural poverty rates. The incidence of extreme poverty (pobreza alimentaria) declined from 53 to 24 percent between 1996 and 2006 (Figure 1-3). Most of this decline is a recovery from the dramatic increase in poverty following the 1995 crisis, since the 1992-2002 decade was fully lost in terms of rural poverty reduction. More recently (2004-2006), the rate of decline of rural poverty has decelerated. In 2006, more than half of the rural population was moderately poor (pobreza patrimonial), and a quarter was extremely poor. Rural areas still account for two thirds of the extreme poor in Mexico, just as they did in 1992. Figure 1-3: Rural Poverty Rates and Rural Share in Total Poverty (%, 1992-2006) 8 90 80% 80 70% 70 60% Rural share in poverty (bar) 60 Poverty rates (line) 50% 50 40% 40 30% 30 20% 20 10 10% 0 0% 1992 1994 1996 1998 2000 2002 2004 2005 2006 Rural share in total poverty (alimentaria) Extreme Poverty rate (alimentaria ) Poverty rate (patrimonial) Source: CONEVAL as cited in Scott (2008) 17. To what extent does agriculture offer an opportunity for rural households to escape poverty? The poverty reduction potential of agriculture is a major theme of the Bank's World Development Report 2008 (World Bank, 2008a). This potential seems to have been sharply diminished in Mexico over the last decade, for independent (non-wage) farm incomes have collapsed from 28.7 to 9.1 percent of rural household income between 1992 and 2004, while total farm income contracted from close to 40 percent to just 17 percent (Figure 1-4). As shown in Table 1-1, however, agriculture is proportionally much more important for the poor, as a source of both independent income and wage labor employment. Figure 1-4: Sources of Rural Household Income (%) 9 100 90 9.1 80 28.7 Farm Income 8.2 70 60 18.5 9.0 50 15.5 40 Non Farm 36.3 30 Income 22.8 20 3.7 10 1.6 4.1 0.1 2.0 Transfers 4.2 8.6 4.6 0 1992 2004 Domestic inter-household transfers Oportunidades and Procampo transfers International transfers Pensions Non-Farm Wage Labor Indep. Non-Farm Activs. Agricultural Wage Labor Independent Farming Source: Ruiz-Castillo (2005) as cited in Scott (2008). Total does not add up to 100% because smaller or unspecified income sources were excluded. Table 1-1: Household Deciles Ordered by Per Capita Income (2006) Agricultural activities by household (hh) deciles ordered by income per capita (2006) hh with agricultural workers hh with independent farming income Annual income hh Households % Decile Households % Decile Deciles million MP MP/hh 1 3,222,510 60% 705,977 26.6% 2,705 3,831 2 1,492,371 32% 249,587 9.4% 1,830 7,331 3 946,424 24% 190,263 7.2% 1,253 6,586 4 625,353 15% 119,835 4.5% 1,038 8,664 5 578,002 13% 103,074 3.9% 1,853 17,977 6 340,805 9% 86,394 3.3% 982 11,362 7 390,019 9% 68,100 2.6% 977 14,349 8 233,630 7% 63,465 2.4% 917 14,456 9 144,672 5% 30,022 1.1% 878 29,249 10 152,976 4% 39,521 1.5% 3,521 89,093 Total 8,126,762 18% 1,656,238 6.2% 15,954 9,633 Source: Scott (2008) based on data in ENIGH (2006) 10 2. DIAGNOSTICS OF THE FEDERAL RURAL BUDGET MEXICO'S RURAL DEVELOPMENT BUDGET: THE PROGRAMA ESPECIAL CONCURRENTE (PEC) 18. PEC has not worked as a coordinating mechanism of ARD programs as intended. The Sustainable Rural Development Law of 2001 established a coordinating framework for ARD public expenditures through the Comisión Intersectorial para el Desarrollo Sustentable (CIDRS), headed by SAGARPA. The expectation was that CIDRS would develop a rural development strategy and serve as a forum to coordinate GOM's ARD activities. A strategy, however, has not been developed, and CIDRS has not yet actually managed to orient and coordinate the allocation of federal resource to ARD programs. Thus, the PEC, annually prepared by CIDRS, has, to date, not functioned as a planning tool; it operates as an annual budget exercise where the budget lines of all rural development programs are tabulated together. PEC appears as an Annex of the annual federal budget.19 19. The PEC budget has been steadily increasing since its first compilation in 2001, almost doubling between 2001 and 2008. The only year in which it declined was 2004. During 2001-2008, it has averaged an annual growth rate of 9.2 percent in constant prices (Figure 2-1). Figure 2-1: PEC Budget 2001-2008 (2007 Constant MxP) 19 The federal budget has, since 1996, explicitly classified the rural development budget even before the formulation of the PEC in 2001. Before 2001, the rural development budget was compiled under the category "gasto en desarrollo rural (rural development costs)"in the budget decree. Since many expenditure programs operate in rural and urban areas, a difficult task of SHCP when constructing PEC is to estimate what part of these programs is spent in rural and urban areas. The criteria used for the apportioning, which in some cases have been challenged by experts, should be made transparent and disclosed jointly with PEC figures. 11 200,000 180,000 MxP million in 2007 value 160,000 140,000 120,000 100,000 80,000 60,000 40,000 20,000 0 2001 2002 2003 2004 2005 2006 2007 2008 Year Source: Centro de Estudios para el Desarrollo Rural Sustentable y la Soberanía Alimentaria as cited in Oliver and Santillanes (2008). 20. PEC reflects GOM's strong commitment to rural development. PEC amounted to 12 percent of the federal budget in 2006 and included more than 60 programs with 85 subprograms and 94 components to be implemented by 13 different secretarías plus sub-national governments. The actual executed 2007 PEC budget was MxP 177 billion (Table 2-1). The largest programs in PEC are PROCAMPO (executed by SAGARPA), Alianza para el Campo (SAGARPA), Apoyos a la Comercializacion (SAGARPA) and Oportunidades (SEDESOL). Approximately one third of the total PEC budget is executed by SAGARPA. Table 2-1: Mexico's Rural Development Budget (PEC) 2006-2008 12 MxP million Percent (%) Allocation Current prices Constant prices (2007) 2006 (a) 2007(b) 2008 (c) 2006 2007 2008 2006 2007 2008 Total PEC 152,934.31 176,794.3 204,000.0 160,944.6 176,794.3 198,598.4 100.00 100.00 100 125,882.21 144,637.2 171,900.3 134,187.2 146,262.0 167,348.6 82.31 82.57 84.26 Total by Ministries Ramo 04. 298.5 300.00 400.0 310.1 300.0 389.4 0.20 0.17 0.20 Ramo 05. SRE 109.2 71.40 69.7 76.0 71.4 67.9 0.07 0.04 0.03 Ramo 06. SHCP 6,180.60 8,291.90 10,795.3 6,781.2 8,291.9 10,509.5 4.04 4.73 5.29 Ramo 08. SAGARPA 48,899.41 58,384.71 65,341.4 53,055.4 58,536.9 63,611.3 31.97 33.33 32.03 Ramo 09. SCT 2,857.80 7,545.7 2,602.3 2,857.8 7,345.9 1.72 1.63 3.70 Ramo 10. Economía 654.5 838.50 918.2 923.3 838.5 893.9 0.43 0.48 0.45 Ramo 11. SEP 23,161.75 24,694.9 25,695.1 23,686.8 24,041.0 16.08 13.22 12.11 Ramo 12. Salud 9,790.80 15,056.45 12,643.8 12,745.9 15,940.5 12,309.0 6.40 8.60 6.20 Ramo 14. STyPS 74.10 66.95 111.0 78.9 67.0 108.1 0.05 0.04 0.05 Ramo 15. Reforma Agraria 4,342.0 4,772.30 5,435.4 4,618.4 4,779.8 5,291.5 2.84 2.72 2.66 Ramo 16. 16,525.10 17,519.2 9,379.0 14,289.8 17,055.3 7.15 9.43 8.59 SEMARNAT 10,930.4 Ramo 20. SEDESOL 17,372.30 14,233.85 26,290.7 17,892.2 16,525.1 25,594.6 11.36 8.13 12.89 Ramo 21. Turismo 1.5 76.50 135.0 1.7 76.5 131.4 0.00 0.04 0.07 Total by other 27,052.10 30,532.50 32,099.7 26,757.4 30,532.3 31,249.7 17.69 17.43 15.74 categories Ramo 19. 5,151.90 Aportaciones a 5,512.90 6,300.0 5,070.0 5,512.7 6,133.2 3.37 3.15 3.09 Seguridad Social Ramo 31. Tribunales 584.2 250.00 790.2 601.2 694.0 769.3 0.38 0.14 0.39 Agrarios Ramo 33. Aportaciones Federales a Estados y 21,316.0 694.00 24,758.5 21,086.2 24,075.6 24,102.9 13.94 0.40 12.14 Municipios Ramo 23. Previsiones Salariales y Económic. 24,075.60 250.0 250.0 243.4 13.74 0.12 Source: Oliver and Santillanes (2008). a. 2006: Executed Budget, data from Public Accounts, SHCP b.2007: Budget incorporated in the Proceso Integral de Programación y Presupuesto (PIPP), SHCP. c. 2008: Programmed budget as presented in the Presupuesto de Egresos de la Federación 2008. 21. Difficulty in coordinating the PEC budget is due to several factors. First of all, there is the magnitude and sheer number of programs and agencies involved, each with their own objectives, beneficiaries and rules.20 Secondly, it is difficult for SAGARPA, a line ministry that chairs the CIDRS, to dictate to other line ministries which ARD programs they should or should not have, and how they should implement them. The situation would be different if CIDRS and the annual PEC formulating exercise were under the mandate of Presidencia or the SHCP. A third reason is the absence of a comprehensive national ARD strategy with measurable objectives and outcomes and a clear indication of how different programs would serve to achieve them. Partly because of this, many programs that were initiated as temporary interventions responding to transitory situations became permanent over time. Also, due to the absence of a national ARD 20 PEC programs are implemented by 13 out of the 18 federal ministries, as well as state governments for decentralized programs, other entities of the federal government and the legislative branch of government. In recent years PEC has included close to 100 different programs, although the consolidation of the 2008 PEC resulted in a reduction to 73 programs, 85 sub-programs and 94 components. 13 strategy, the majority of programs are formulated to serve objectives and criteria of the originating agency, or even of specific offices within those agencies. As a result, the PEC budget encompasses numerous programs, sub-programs and components with duplicating, or at times even contradicting activities, as well as with noticeable gaps. The introduction of PEC as an accounting instrument was a valuable step forward in bringing transparency to ARD budgetary allocations, and is a first step towards coordination and consistency in ARD expenditures. PEC, however, is not yet fulfilling its role as the mechanism to promote the efficient allocation and use of public resources to promote ARD. CLASSIFICATION OF ARD PROGRAMS 22. This section classifies the PEC budget (2006) into various categories: public- private, social-productive, targeted to the poor-untargeted, etc., in order to better understand the structure of Mexican ARD expenditures. 21 The classification exercise is complex and leaves room for subjective judgment, partly because the categories themselves are not always clear-cut, and partly because many programs have sub-programs and components falling under different categories. 22 Programs have been disaggregated to the lowest level possible to enable proper classification. Whenever available, data from SHCP accounts of the executed budget were used. When these data were only available at the aggregate level, additional information was used to allow for disaggregation.23 23. Public goods programs. Included here are programs typically provided by governments in market economies, generally because they supply goods or services whose consumption is neither excludable nor rivalrous. 24 These programs usually include investments like rural infrastructure (excluding on-farm structures), and general social or productive services such as rural education and health, telecommunications, information, training and research systems, technology transfer to small producers, sanitary systems, natural resource conservation and environmental programs, emergency programs, and operational budgets of institutions. Table 2-2 presents the breakdown of these programs in PEC. 21 The classification was conducted for 2006 since this was the latest year for which a completed executed budget was available at the level of disaggregation necessary for the analysis. 22 Because of this, some programs appear under more than one category. The amounts appearing in the different categories correspond to that spent in the sub-programs or components falling under the category. This is the case, for instance, of Fondo para la Infraestructura Social Municipal (FISM), part of which is classified as "social" and part as "indirectly productive", and of Alianza, partly classified as a "public goods" and partly as a "private goods" program. 23 Additional sources include implementing agency data and external evaluation reports. 24 "Excludable" means that persons can be prevented from consuming the good or service, and therefore it is feasible to charge for its consumption. "Rivalrous" means that consumption of a good or service by one person reduces the availability for another person to consume. Economic principles dictate that it is most efficient for the public sector to provide goods and services that are neither excludable nor rivalrous; therefore they are called "public goods". Also include here are "quasi-public goods" falling between public and private goods, for which beneficiaries are only partially excludable from the consumption of the good or service, and the cost of providing the good or service is less than proportional to the number of consumers, hence are not fully rivalrous, e.g. education often has the characteristics of a quasi-public good. 14 Table2-2: Major Public Goods Programs (categorized into social, productive, and indirectly productive) 2006 MxP Percent Million (%) PUBLIC GOODS PROGRAMS 79,120.68 100.00 A. Social Programs 40,186.94 50.79 1. Fondo para la Infraestructura Social Municipal (FISM) dentro del Ramo 33. Aportaciones a estados y municipios 8,681.40 10.97 2. SPSS (Sistema de Protección Social en Salud) 8,653.10 10.94 3. Aportaciones federales dentro del Ramo 33. Aportaciones a estados y municipios 7,939.40 10.03 4. Programa IMSS-Oportunidades (sin Salud Indígena y Progresa) 5,151.90 6.51 5. Comisión Nacional de Desarrollo de los Pueblos Indígenas (CONADEPI) 4,790.70 6.05 6. Enciclomedia 1,801.60 2.28 7. Programas Alimentarios de DICONSA y LICONSA 1,593.00 2.01 8. Forestal (incluye Proárbol) 761.94 0.96 9. Áreas naturales protegidas 465.50 0.59 10. Programa de Atención a Jornaleros Agrícolas 137.90 0.17 11. Otros 210.50 0.27 B. Productive Programs 24,043.74 30.39 1 Programa de desarrollo de capacidades (SEP) 4,975.50 6.29 2. Fondo para la Infraestructura Social Municipal (FISM) dentro del Ramo 33. Aportaciones a estados y municipios 4,695.20 5.93 3. Infraestructura hidroagrícola 3,297.10 4.17 4. Actividades recurrentes de apoyo productivo con presupuesto regular de SAGARPA* 2,943.10 3.72 5. Caminos rurales 1,823.20 2.30 6. Alianza para el Campo** 1,529.14 1.93 7. Forestal (incluye Proárbol) 913.50 1.15 8. Programa Normal de Sanidades 801.40 1.01 9. Sistema Financiero Rural 677.80 0.86 10. Microrregiones (incluye FONAES) 638.00 0.81 11. Otros 1,749.80 2.21 C. Programas de Fomento Productivo Indirecto (Indirectly Productive) 14,890.00 18.82 1. Gasto operativo (netamente administrativo) de la SAGARPA 4,485.80 5.67 2. Educación agropecuaria (SEP) 3,843.30 4.86 3. Instituciones de Educación Agropecuaria dependientes de SAGARPA 1,910.70 2.41 4. Conflictos y Conciliación Agraria 824.40 1.04 5. Gasto operativo de los Tribunales Agrarios 584.20 0.74 6. Procuraduría Agraria 583.70 0.74 7. Gasto operativo de la SRA 545.20 0.69 8. Programa de Certificación de Derechos Ejidales y Titulación de Solares (PROCEDE) 487.90 0.62 9. INEGI 363.50 0.46 10. Registro Agrario Nacional (RAN) 361.70 0.46 11. Otros 899.60 1.14 Source: Oliver and Santillanes (2008) 24. Private goods programs. Included here are programs that are usually provided (or could be provided) by the private sector in market economies (mostly because they supply goods or services whose consumption is excludable and rivalrous), and also government programs that provide measurable subsidies to individuals or 15 families usually under some targeting criteria.25 Included in this category are, inter alia, subsidies to on-farm productive infrastructure and equipment, subsidies to rural credit and agricultural insurance, input support, marketing support, and compensatory cash transfers. It should not be assumed that government funding of private goods is necessarily undesirable. There are private goods that are productivity enhancing and may make sense for government to invest in them for policy purposes, even if they consist of measurable subsidies to individuals or families. This is for instance the case with Oportunidades, which enhances human capital, and with many Alianza subsidies, which enhance the productive potential of farms. Table 2-3 presents the breakdown of these programs. Table 2-3: Major Private Goods Programs (categorized into public, private, and indirect private programs) 2006 MxP Million Percent (%) PRIVATE GOODS PROGRAMS 73,812.70 100.00 A. Social Programs 29,726.00 40.27 1. Oportunidades (Educación ­ SEP) 13,941.50 46.90 2. Oportunidades (SEDESOL) 11,793.90 39.68 3. Programa de Vivienda Rural 2,207.50 7.43 4. Oportunidades (Salud ­ SSA) 1,108.60 3.73 5. Fondo para pago de adeudos a braceros rurales del 42 al 64 298.50 1.00 6. Adultos Mayores en Zonas Rurales 297.90 1.00 7. Apoyo para la repatriación de cadáveres a México 39.40 0.13 8. Asistencia jurídica urgente a mexicanos en EE.UU. y defensa de condenados a muerte 17.00 0.06 9. Migrantes en situación de probada indigencia 8.10 0.03 10. Otros 13.60 0.05 B. Productive Programs 43,825.90 59.37 1. PROCAMPO 15,024.50 34.28 2. Ingreso Objetivo (maíz, frijol, sorgo, algodón, arroz, trigo, soya...)/Apoyos directos al productor por excedentes de comercialización 6,961.23 15.88 3. Alianza para el Campo 4,900.10 11.18 4. Programas Hidráulicos 3,807.40 8.69 5. Fondo de Compensación a Costos Energéticos Agrícolas 2,684.20 6.12 6. Fondo de Apoyo a la Competitividad de las Ramas Productivas 2,258.30 5.15 7. Programa Ganadero (PROGAN) 2,070.00 4.72 8. Programa de Empleo Temporal (SCT)* 791.90 1.81 9. FOMAGRO (FIRCO) 746.70 1.70 10. Forestal (incluye Proárbol) 635.16 1.45 11. AGROASEMEX (Fondos de Aseguramiento y de Microseguros) 520.00 1.19 12. Otros 3,426.41 7.82 C. Indirect Productive Programs 260.80 0.35 45. Obligaciones Jurídicas Ineludibles 260.80 0.35 Source: Oliver and Santillanes (2008). 25. Productive programs: These programs are targeted at directly improving the productive capacity of agriculture or other rural economic activities. They 25 By including measurable subsidies to private individuals of families, this category includes cash transfer programs such as PROCAMPO and Oportunidades, and to "merit goods" programs (Musgrave, 1959) like Vivienda Rural. 16 include, inter alia, productive infrastructure, agricultural research and technology transfer, sanitary systems, subsidies to rural credit and agricultural insurance, improvements of productive resources, production subsides, marketing supports, and marketing information. A different category labeled indirect productive programs expenses captures those aimed at improving productive capacity but with less direct impact on production as well as ministerial expenses to operate the relevant programs. This category includes programs such as tertiary level agricultural education, agricultural institutional programs, and land registry programs, as well as ministerial operational budgets. 26. Social programs: These programs consist of expenditures aimed at improving the general living conditions of the rural population and are not directly oriented at expanding their productive capacity. Included here are, inter alia, the improvement of social infrastructure, cash transfers not tied to current or historical productive use, education, health and social welfare programs, and consumption support programs. Naturally, social programs have an indirect productive impact, but their primary aim is to improve the wellbeing of the beneficiaries not their productive capacity. 27. Targeted and untargeted programs: Programs can further be categorized into those that are targeted to the poor, small producers and vulnerable groups and those that are not. Included in targeted programs are those oriented to the indigenous populations, peasant women and casual workers, as well as other programs aimed at improving the conditions of small producers or the rural poor in which targeting is mandated in the operation rules, even if in practice a good part of the expenditures does not actually reach the poor. Non-targeted programs are those benefiting the rural population at large, not specifically aimed at small producers, vulnerable groups or the poor. 28. Program decentralization: Another potential distinction is between centralized and decentralized programs. This distinction, however, is difficult to establish in practice because decentralization has several dimensions and the situation of programs differs according to these dimensions (see Box 2-1). The design, normative aspects and most or all funding of programs included in PEC come for the federal level although there already is some level of decentralization in the so- called "decentralized" subprograms of Alianza. Box 2-1: Dimensions in the Decentralization of Rural Development Programs in Mexico The many rural sector programs that exist in Mexico operate under different institutional arrangements. Some devolve decision-making power to sub-national governments to various extents, while others keep most or all aspects under the control of central authorities. A simple centralization/ decentralization dichotomy is insufficient to describe how rural programs work, for program processes move along several dimensions, some under the control of central authorities, others under sub-national governments. a) Funding source. Funding of rural programs may come from one, two or all three levels of government. For most programs the federal level provides all or most of the resources. The programs, for instance, of SRA (PROMUSAG and FAPPA) and SECON (FONAES) do not have as a rule, state or municipal counterpart funding. When two government levels contribute, they are usually the federal and state governments. This is the case with the Agriculture and Livestock sub-programs of Alianza or the 17 PRODEFOR program of CONAFOR. A combination of federal and municipal funding, like in FIRCO's Microcuencas program, does exist but is uncommon. In hardly any program do all three levels of government contribute funds, with the exception of Alianza's Rural Development Program (PDR). Only in the case of states' own programs, like the fertilizer program of the Secretaría de Desarrollo Rural in Guerrero or the Núcleos de Desarrollo Comunitario program of the Secretaría de Desarrollo Humano in Jalisco, does funding come exclusively from state sources. Rural Development programs financed entirely by municipalities are rare. b) Regulatory authority. Program operational rules serve to establish objectives and set forth key implementation asepects, ranging from the eligibility criteria to the amount and type of benefits. Except for states' own programs, these rules are issued by the federal government. All rural development programs from federal ministries operate under centralized rules. When states and municipalities are involved in program funding, additional criteria may be agreed between the government levels involved, formalized as convenios and anexos técnicos, which usually include budgetary and programmatic provisions. Alianza´s decentralized sub-programs are an example of this. c) Allocation of funds to States. Most rural development programs operate on a demand basis, i.e. producers apply for and compete for funding. In the case of centralized programs, competition may be at the national level, in which case no allocation or indicative disbursement figure is given to the states. This is for instance the case with CONAFOR's PRODEPLAN. In other cases, like in SEDESOL's Oportunidaes Productivas and in FONAES, annual expenditure ceilings are established for each state, and proposals compete at the state-level within those ceilings. Funds, however, are kept at the federal level. In the case of decentralized programs, not only funds are allocated to the states, they are actually transferred to fideicomisos located in the states where they are mixed with local funds. d) Spending authority. Even when funds come from federal sources, prioritization and selection of proposals may involve sub-national governments. The most centralized scenario is when the selection process is carried out centrally by federal ministries. This can be done with the prior advice of a state-level joint committee, like in the SRA's programs, or without it, like in the Integración Productiva, Agencias de Desarrollo Local and Fondo de Financiamiento Social programs of SEDESOL's Oportunidades Productivas. In other cases, like in FONAES, proposals under a certain amount are approved by the state offices of the corresponding ministry. In more decentralized scenarios, joint state-federal committees may conduct preliminary selection of proposals with the final decision being left to the federal level. A more decentralized process is when, like in Alianza's decentralized programs or in CONAFOR's PRODEFOR program, state-level joint committees or even municipal-level committees make the final decision. Total decentralization of spending occurs when the use of resources is decided by sub-national authorities alone, but this only happens with state's own programs. e) Operational tasks. Rural development programs entail a number of operational tasks, such as promotion and dissemination, reception and processing of applications, technical formulation of proposals, and supervision of activities. Different government levels may take part in these tasks, which are usually supported by private providers of services. Source: World Bank 2006a THE PEC STRUCTURE 29. ARD expenditures are equally divided between public and private goods.26 In 2006, 52 percent of PEC (MxP 79 billion) went to public goods programs and 26 Lopez and Galinato (2007) analyze the rural public expenditure of LAC countries for 1985-2001using a FAO database which shows that 66 percent of Mexico's rural public expenditure is private, which is the fourth highest in the region after Brazil (87 percent), Dominican Republic (80 percent), and Guatemala (69 percent). Contrarily, Honduras (9 percent) and Uruguay (19 percent) have the lowest shares. The definition of rural expenditure used in this FAO study does not coincide with that of the PEC since universal social services such as education and health are also included in the FAO's definition of rural expenditure. 18 the remaining 48 percent (MxP 73.8 billion) to private goods programs (Table 2- 4). The implementation of public programs was led by SAGARPA (17 percent), followed by the Ministry of Education (14 percent), and the Ministry of Health (11 percent, spent mainly in seguro popular). In addition, budget transfers from the federal government to states and municipalities, used mainly for rural infrastructure, amounted to 27 percent of public goods programs. Expenditure in private goods concentrated in three ministries: SAGARPA (48 percent, mainly for PROCAMPO, Apoyos, Alianza and PROGAN), SEDESOL (20 percent, for Oportunidades and the rural housing program), and the Ministry of Education (19 percent, also for Oportunidades). Table 2-4: ARD Public and Private Goods Programs by Government Agency (2006) Total Public Goods Private Goods Allocation MxP MxP Million Percent (%) MxP Million Percent (%) Percent (%) Million Total 152,933.38 100.00 79,120.68 100.00 73,812.70 100.00 04. Gobernación 298.50 0.20 0.00 0.00 298.50 0.40 05. Relaciones Exteriores 109.20 0.07 31.10 0.04 78.10 0.11 06. Hacienda y Crédito Público 6,180.50 4.04 5,331.90 6.74 848.60 1.15 08. SAGARPA 48,898.69 31.97 13,538.94 17.11 35,359.75 47.90 09. SCT 2,634.00 1.72 1,842.10 2.33 791.90 1.07 10. Economía 654.39 0.43 39.10 0.05 615.29 0.83 11. Educación Pública 24,594.90 16.08 10,653.40 13.46 13,941.50 18.89 12. Salud 9,790.80 6.40 8,682.20 10.97 1,108.60 1.50 14. Trabajo y Previsión Social 74.10 0.05 74.10 0.09 0.00 0.00 15. Reforma Agraria 4,342.00 2.84 3,034.50 3.84 1,307.50 1.77 16. SEMARNAT 10,930.40 7.15 6,385.94 8.07 4,544.46 6.16 19. Aportaciones a Seguridad 5,151.90 3.37 5,151.90 6.51 0.00 0.00 20. Desarrollo Social 17,372.30 11.36 2,455.30 3.10 14,917.00 20.21 21. Turismo 1.50 0.00 0.00 0.00 1.50 0.00 31. Tribunales Agrarios 584.20 0.38 584.20 0.74 0.00 0.00 33. Aportaciones Federales a Estados y Municipios 21,316.00 13.94 21,316.00 26.94 0.00 0.00 Source: Oliver and Santillanes (2008) using data from SHCP 30. ARD expenditures are equally distributed between social and productive programs. Thus, 46 percent of PEC (MxP 70 billion) went to social programs, 44 percent (MxP 68 billion) to productive programs, and 10 percent (MxP 15 billion) to indirectly productive programs 27 (see Table 2-5). Social programs were mainly implemented by SEDESOL (Oportunidades, rural housing, and food 27 It should be noted that the even distribution of ARD expenditures between productive and social programs is influenced by the classification of PROCAMPO as a productive program. PROCAMPO could arguably be classified as a social program since its objective is to support farmers' incomes rather than to boost production, and there is no conditionality on the use of the entitlement. If PROCAMPO were classified as a social program, social programs would account for 55.5 percent of ARD expenditures, productive programs for 34.6 percent and indirectly productive programs for 9.9 percent. 19 programs), the Ministry of Education (Oportunidades and Enciclomedia), and the Ministry of Health (social protection system and Oportunidades). These three ministries accounted for nearly half (49 percent) of spending in rural social programs. Almost three quarters of the expenditure in productive programs was executed by SAGARPA (in particular PROCAMPO, Alianza, Apoyos, and PROGAN), followed by SEMARNAT, mainly for irrigation infrastructure, and the Ministry of Education for programs in agriculture technical training and plant breeding training. Indirect productive programs comprised mainly tertiary level agricultural education, and regular recurrent costs of agencies implementing PEC programs. These costs come mainly from three ministries, SAGARPA, the Ministry of Education and the Ministry of Land Reform, which were jointly responsible for 91 percent of this spending category. 31. Production-oriented public goods programs are underfunded despite their importance for ARD. Table shows that less than one third of spending in public goods programs goes to directly production-oriented programs. The largest spending in production-oriented programs of a public goods type was for productive infrastructure (10.1 percent of public goods), training programs by the Ministry of Education (6.3 percent), the productive part of FAIS, irrigation infrastructure, and the public goods sub-programs of Alianza (5.7 percent). The low level of spending by SAGARPA is noticeable given its mandate for the development of the rural productive sector. SAGARPA allocates only 9.1 percent of its total budget to production-oriented programs of a public goods type. Considering the documented importance for production activities of the provision of public goods28, the fact that only 16 percent of PEC is for production-oriented public goods programs signals overfunding of private goods, some of which could potentially be reallocated to the funding of public goods (see Box 2-2). Table 2-5: Cross-classification of Public/Private and Social/Productive ARD Programs, 2006 Total Cost Social Productive Indirect Productive MxP Percent MxP Percent MxP Percent MxP Percent (%) Million (%) Million (%) Million (%) Million Total Cost 152,933.38 100.00 69,912.94 100.00 67,869.64 100.00 15,150.80 100.00 Public 79,120.68 51.74 40,186.94 57.48 24,043.74 35.43 14,890.00 98.28 Private 73,812.70 48.26 29,726.00 42.52 43,825.90 64.57 260.80 1.72 Source: Oliver and Santillanes (2008) using data from SHCP Box 2-2: Efficiency gains from switching from private to public goods Recent econometric work finds that in many countries in the LAC region it is crucial to shift rural public expenditures from large subsidies going to specific groups of producers and towards the increased provision of public goods1. As an illustration of the importance of the mix of spending between private and public good, R. Lopez used the FAO LAC Regional Office database from several countries (1985-2001) to classify expenditures. Public goods included technology generation and transfer, soil conservation, sanitary and phytosanitary protection, communications and information services, rural infrastructure, and social services (education and health). Private goods included commodity-specific or focalized items, marketing assistance and promotion, subsidized credit, and irrigation. 28 For the importance the investment public goods see Lopez and Galinato (2007), World Bank (2005a), World Bank (2008). 20 The key message from the analysis is that while government expenditures do have a positive effect on agricultural per capita income, the composition of those expenditures is important. The long-run effects on per capita agricultural GDP of increasing the share of expenditures going to subsidies were large and negative (and highly statistically significant.) A reallocation of 10% points of public expenditures from subsidies to public goods would increase per capita agricultural income by about 2.3 % without increasing total expenditures. On the poverty side, the analysis found no statistically significant direct effect of the mix of expenditures on poverty, in part due to the poor state of rural poverty data. The main impact of the mix of public expenditures on poverty is indirect, through the effect on per capita agricultural income. Overall, the findings from the cross-country analysis of Lopez for Latin America are consistent with other analyses for India and China, where spending on rural roads and agricultural research has been found to have large poverty alleviation effects. The allocation criteria for expenditures are not complex in theory ­ public goods should be a priority, and coverage, targeting and the cost-effectiveness of transfers should enter any project evaluation. But designing public spending is complicated in practice, because the policy maker would also need to have empirical measures of the relevant "elasticities" reflecting how agricultural growth and poverty reduction responds to expenditures on various program categories. And from the national development perspective one should have information on how national welfare responds to agricultural development. Policy design is further complicated by the heterogeneity of farming conditions with respect to initial product mixes and levels of productivity, access to infrastructure, soil fertility, and access to markets. This heterogeneity stems in part from the natural resource base and geography, but also in part from the approach governments have taken in the past to expenditures on public goods. Past investments in infrastructure and institutions, such as telecommunications and transport, and the support network of finance, research and extension, will determine the ability of small farmers to take advantage of opportunities in both domestic and international markets. Finally, although the role of governments as the provider of public goods is well established, the performance of governments is often disheartening. 1 The work by Ramon Lopez of the University of Maryland was pioneering in this area and first developed for the WB regional report "Beyond the City, The Rural Contribution to Development" in 2005, and later published as Lopez and Galinato (2007). Source: Valdes (2008b) 32. While the majority of social programs are targeted to the poor (including small farmers and vulnerable groups), the majority of agriculture programs are not targeted. Although the definition of poor differs significantly by each program and has changed over time within programs, 96 percent of the expenditure in social programs in PEC is targeted to the poor (Table 2-6), while only 7.9 percent of the expenditure in agricultural programs is targeted (Table 2- 7). Because of leakages, part of the expenditure in some targeted programs may not reach the poor. Oportunidades, which is the largest targeted program, is particularly well targeted. Expenditures in other programs, particularly in those involving infrastructure and area- or community-based targeting, are less well targeted. On the other hand, part of the expenditure in non-targeted programs favors the poor. This is, for instance, the case with PROCAMPO payments to which small and very small farmers are eligible. Table 2-6: Targeting of Social Programs MxP Percent Million (%) SOCIAL PROGRAMS 69,912.94 100.00 A. Programs targeted at poor 66,865.00 95.64 21. Oportunidades (Educación ­ SEP) 13,941.50 19.94 21 2. Oportunidades (SEDESOL) 11,793.90 16.87 3. Fondo para la Infraestructura Social Municipal dentro del Ramo 33. Aportaciones a estados y municipios 8,681.40 12.42 4. SPSS (Sistema de Protección Social en Salud) 8,653.10 12.38 5. Aportaciones federales dentro del Ramo 33. Aportaciones a estados y municipios 7,939.40 11.36 6. Programa IMSS-Oportunidades (sin Salud Indíg. y Progresa) 5,151.90 7.37 7. Comisión Nacional de Desarrollo de los Pueblos Indígenas (CONADEPI) 4,790.70 6.85 8. Programa de Vivienda Rural 2,207.50 3.16 9. Programas Alimentarios 1,593.00 2.28 10. Oportunidades 1,108.60 1.59 11. Otros 1,004.00 1.44 B. Programs not targeted at poor 3,047.94 4.36 1. Enciclomedia 1,801.60 2.58 2. Forestal (incluye Proárbol) 761.94 1.09 3. Áreas naturales protegidas 465.50 0.67 4. Telefonía Rural 18.90 0.03 Source: Oliver and Santillanes (2008) using data from SHCP Table 2-7: Targeting of Agriculture Programs MxP Percent Million (%) AGRICULTURAL PROGRAMS 47,559.97 100.00 A. Programs targeted at poor 3,760.34 7.91 1. Alianza para el Campo (componente de desarollo rural) 2,083.55 4.38 2. FOMAGRO (FIRCO) 746.70 1.57 3. PIASRE 462.90 0.97 4. Fondo de Tierras 248.50 0.52 5. Fondo Nacional de Apoyos para Empresas en Solidaridad (FONAES) 173.59 0.36 6. Microcuencas (FIRCO) 45.10 0.09 B. Programs not targeted at poor 43,799.63 92.09 8. PROCAMPO 15,024.50 31.59 1. Apoyos directos al productor por excedentes de comercialización (Ingreso Objetivo y otros) 6,961.23 14.64 2. Programas Hidráulicos 3,807.40 8.01 3. Infraestructura hidroagrícola 3,297.10 6.93 4. Alianza para el Campo 2,976.98 6.26 5. Actividades recurrentes de apoyo productivo dentro del presupuesto regular de SAGARPA 2,863.70 6.02 6. Fondo de Compensación a Costos Energéticos Agrícolas 2,684.20 5.64 7. Fondo de Apoyo a la Competitividad de las Ramas Productivas 2,258.30 4.75 8. Programa Ganadero (PROGAN) 2,070.00 4.35 9.- Programa Normal de Sanidades 801.40 1.69 10. AGROASEMEX (Fondos de Aseguramiento y de Microseguros) 520.00 1.09 11. Programa de atención a Contingencias Climatológicas FAPRACC 337.82 0.71 12. Fondo de Apoyo a las Organizaciones Sociales, Agropecuarias y Pesqueras (PROSAP) 197.00 0.41 Source: Oliver and Santillanes (2008) using data from SHCP 33. In contrast to the importance of the non-farm sector in the rural economy, PEC reflects an agro-centered view of rural development. Some 70 percent of spending in production-oriented programs of a public goods type is for 22 agricultural activities and 30 percent for programs open to other rural productive activities (Table 2-10). The main programs in the latter category as shown in Table 2-9 are the skills development program of the Ministry of Education, which supports the enhancement of different rural technical skills (25 percent of non- agricultural programs), a part of FAIS, which supports productive infrastructure (23 percent), rural roads (9 percent), and forestry and other environmental programs (7 percent). The weight of the agriculture programs is probably underestimated since a good part of the resources of many not agriculture-specific programs is probably, in fact, used to support agricultural activities. Table 2-8: Main Agricultural Programs (2006) MxP Percent Million (%) AGRICULTURAL PROGRAMS 47,559.97 100.00 A. Public Goods 8,736.69 18.37 1. Infraestructura hidroagrícola (Comisión Nacional del Agua ­ SEMARNAT) 3,297.10 10.13 2. Actividades recurrentes de apoyo productivo dentro del presupuesto regular de SAGARPA (organismos sanitarios, y de investigación, información y capacitación) 2,853.80 8.77 3. Alianza para el Campo 1,079.39 3.32 4. Programa Normal de Sanidades 801.40 2.46 5. Programa Integral de Agricultura Sostenible y Reconversión Productiva (PIASRE) 462.90 1.42 6. Fondo de Apoyo a las Organizaciones Sociales, Agropecuarias y Pesqueras (PROSAP) 197.00 0.61 7. Microcuencas (FIRCO) 45.10 0.14 B. Private Goods 38,823.28 81.63 1. PROCAMPO 15,024.50 38.70 2. Ingreso Objetivo (maíz, frijol, sorgo, algodón, arroz, trigo, soya...)/Apoyos directos al productor por excedentes de comercialización 6,961.23 17.93 3. Alianza para el Campo 3,981.14 10.25 4. Programas Hidráulicos 3,807.40 9.81 5. Fondo de Compensación a Costos Energéticos Agrícolas 2,684.20 6.91 6. Fondo de Apoyo a la Competitividad de las Ramas Productivas 2,258.30 5.82 7. Programa Ganadero (PROGAN) 2,070.00 5.33 8. FOMAGRO (FIRCO) 746.70 1.92 9. AGROASEMEX (Fondos de Aseguramiento y de Microseguros) 520.00 1.34 10. Programa de atención a Contingencias Climatológicas FAPRACC 337.82 0.87 11. Fondo de Tierras 248.50 0.64 12. Fondo Nacional de Apoyos para Empresas en Solidaridad (FONAES) 173.59 0.45 13.Actividades recurrentes de apoyo productivo dentro del presupuesto regular de SAGARPA 9.90 0.03 Source: Oliver and Santillanes (2008) using data from SHCP Table 2-9: Main Non-Agricultural (Productive) Specific Programs (2006) MxP Percent Million (%) NON_AGRICULTURAL SPECIFIC PROGRAMS 20,309.67 100.00 A. Public Goods 15,307.05 75.37 1. Programa de desarrollo de capacidades (SEP) 4,975.50 24.50 2. Fondo para la Infraestructura Social Municipal dentro del Ramo 33. Aportaciones a estados y municipios 4,695.20 23.12 3. Caminos rurales 1,823.20 8.98 4. Forestal (incluye Proárbol) 913.50 4.50 5. Sistema Financiero Rural dentro de las acciones de SAGARPA 677.80 3.34 23 6. Microrregiones (incluye FONAES) 638.00 3.14 7. Alianza para el Campo (apoyo a acuacultura y pesca) 449.75 2.21 8. Otros de Medio Ambiente 448.10 2.21 9. Fomento y Organización Agraria 143.50 0.71 10. Financiera Rural 127.70 0.63 11. Otros 414.80 2.04 B. Private Goods 5,002.62 24.63 1. Alianza para el Campo (acuacultura y pesca) 918.96 4.52 2. Programa de Empleo Temporal (componente operado por la SCT) 791.90 3.90 3. Forestal (incluye Proárbol) 635.16 3.13 4. Programa de la Mujer en el Sector Agrario (PROMUSAG) 482.40 2.38 5. Programa de Opciones Productivas 400.60 1.97 6.- Dar Continuidad a los Apoyos al Sector Pesquero 314.20 1.55 7. Fondo Nacional de Apoyos para Empresas en Solidaridad (FONAES) 311.40 1.53 8. Fondo de Apoyo para Proyectos Productivos (FAPPA) 291.40 1.43 9. Financiera Rural 202.30 1.00 10. Programa de Empleo Temporal (componente operado por la SEDESOL) 196.40 0.97 11. Otros 457.90 2.25 Source: Oliver and Santillanes (2008) using data from SHCP 34. Agricultural programs are skewed towards private goods and non- agricultural programs towards public goods. More than 80 percent of the agricultural programs are private, while more than 75 percent of those non- specifically oriented to agriculture are public (Table 2-10). Echoing the argument with productive goods in general, the low level of public spending for agricultural activities is noticeable given the evidence of its superior growth impact. As indicated above, non- agriculture specific programs focus on the provision of public goods such as skills training and rural infrastructure. Non-agriculture specific programs other than the ones mentioned above have tended to be somewhat ad-hoc pilot-type experiences reflecting the difficulty in designing non- agricultural rural productive programs in the absence of a multi-sectorial national rural development strategy. Table 2-10: Productive programs cross-referenced with public/private and agriculture/non-agriculture (2006) Not specifically Total Cost Agricultural agricultural MxP Percent Percent Percent (%) MxP Million MxP Million Million (%) (%) Total 67,869.64 100.00 47,559.97 100.00 20,309.67 100.00 Public 24,043.74 35.43 8,736.69 18.37 15,307.05 75.37 Private 43,825.59 64.57 38,823.28 81.63 5,002.62 24.63 Source: Oliver and Santillanes (2008) CHANGES IN THE 2008 PEC 35. An effort was made in 2007 to consolidate the 2008 PEC budget. In 2007, a new, more compact functional classification scheme was introduced, consolidating the wide multiplicity of PEC programs into nine broad groups and 24 16 programs (see Table 2-11). Some simplification in the operational rules of programs was also introduced in the 2008 budget. However, this reform did not involve an actual integration or coordination of the original programs, which in the 2008 PEC numbered 73, many of them containing multiple subprograms. The principal group within the broader classification, Competitiveness, represents almost 40 percent of PEC and includes productive programs as traditionally defined (Alianza), the ASERCA compensatory programs, input subsidies, and even targeted education spending (including Oportunidades). Table 2-11: 2008 PEC Budget PEC 2008 Total budget (million MxP) 204,000 Groups (Programs) Competitiveness 39.4% Educación 13.3% Apoyos Directos al Campo 8.2% Adquisición de activos productivos 8.0% Atención a problemas estructurales 5.8% Soporte 1.8% Inducción y Desarrollo del Financiamiento al Medio Rural 1.6% Atención a Contingencias Climatológicas 0.4% Apoyo a la participación de actores para el desarrollo rural 0.4% Infrastructure (Infraestructura en el medio rural) 20.6% Social (Atención a la pobreza en el medio rural) 16.3% Health (Atención a las condiciones de salud en el medio rural) 9.3% Environmental (Uso sustentable de recursos naturales para la prod. Primaria) 6.5% Administrative (Gasto administrativo) 4.9% Financial (Financiamiento y aseguramiento al medio rural) 1.6% Labor (Mejoramiento de condiciones laborales en el medio rural) 0.8% Land Reform (Atención de aspectos agrarios) 0.6% Source: Scott (2008) 36. The new structure of PEC reveals little spending in some critical areas like finance, labor and land related programs. Together, the first two categories (finance and labor) absorb just 2.4 percent of PEC, equivalent to half of administrative spending. Only MxP 1.7 billion is allocated to rural employment programs, which is less than 1 percent of the 2008 PEC. Land related programs (known as programas agrarios in Mexico), which include, inter alia, access to land, regularization of land rights, legal and other forms of support related to land rights for ejidatarios and small farmers in general, and resolution of land conflicts, absorb only 0.6 percent of PEC, in contrast to the importance of land issues in Mexico. Social security/protection could probably be better identified as a separate category but is included in the health and poverty programs. Until very recently, it was practically non-existent in rural areas. This has changed dramatically since the introduction and rapid growth of the Seguro Popular since 2004, and a minimum non-contributive universal rural pension started in 2007 (Adultos Mayores de 70 años y Más en Zonas Rurales). In 2008, following a 25 large increase in the approved budget for the latter program (MxP 9.9 billion vs. the originally proposed MxP 6.4 billion), these two programs account for approximately 10 percent of PEC. OTHER FISCAL COSTS OUTSIDE PEC 37. Although PEC captures the majority of public expenditure for rural development except for universal social services (such as education and health), there are policy measures implying fiscal costs that are not budget- based and hence fall outside PEC but do contribute to the rural agenda. They include tax exemptions, the Tarifa 9, which subsidies electricity tariffs to pump groundwater, water tariff subsidies, agricultural credit subsidies, and subsidies to crop insurance. The amounts of these implicit agricultural subsidies are significant. Because of difficulties in quantifying the magnitude of each program, only agricultural tax exemptions and Tarifa 9 are examined here. 38. The agricultural sector benefits from several tax exemptions totaling MxP 24.5 billion, which is approximately 5 percent of total tax exemptions and 8 percent of total income tax exemptions. There are 6 different exemption/incentive categories for agriculture. The largest is the reduction in income tax, which totals MxP 11.2 billion (Table 2-12). Other exemptions include deductions for temporary farm worker subsistence (MxP 6.0 billion), for land dedicated to agriculture (MxP 54 million), of income tax for individuals with up to 20 times the agriculture minimum wage (salarios mínimos agropecuarios [SMA], MxP 4.8 billion), and for associations up to 40 times the agriculture minimum wage (MxP 2.4 billion). Table 2-12: Tax exemptions and preferential treatment of the agricultural sector (2006) % of National MxP million GDP Tax Exemption Total 24,547.6 0.27 1. 44.83% reduction in Income Tax (ISR) 11,203.8 0.12 2. Exoneration of 20 minimum agricultural wages (SMA) for each associate or member (cap of 200 times the general 4,801.6 0.05 minimum wage of Mexico's Distrito Federal) 3. Deduction of 19% of revenue of self employed farmers to cover casual work employment, animal feeding and incidental 5,997.5 0.07 expenses (no documentation required) 4.Exoneration of personal income tax (ISR) up to 40 minimum agricultural wages (SMA) 2,382.8 0.03 5. Deduction of investments under the "intermediate regime" 107.9 0.00 6. Deduction for lands devoted to agricultural activities 54.0 0.00 (anticipated) Source: Oliver and Santillanes (2008) using data from Presupuesto de Gastos Fiscales 2006 39. Agriculture uses 76 percent of available water resources in the country but contributes only 1.9 percent of the water use fees and bulk water tariffs 26 collected by the Comisión Nacional de Agua (CONAGUA). Despite the water scarcity situation in many parts of the country, farmers pay only a portion of the costs involved in delivering irrigation water. In addition, water charges, although theoretically based on the volume of water used, are normally calculated per hectare because of the lack of water measuring devices. The current water tariff structure does not fully cover the operation and maintenance costs and depreciation costs of irrigation infrastructure. Also, farmers pay a small fee when their water use exceeds their concession (MxP 0.1 per cubic meter for excess water use). This does not take into account the opportunity cost of water, which would set the price at the marginal cost. Low water pricing and lack of measurement encourage excessive water use and wasting, a situation that is far from being unique to Mexico. In fact, in many countries, irrigation water costs are highly subsidized, with farmers paying part only of the operation and maintenance costs.29 Cost recovery for operation and maintenance of irrigation systems is generally better in Mexico than in other developing countries. 40. The electricity subsidy is costly and promotes an unsustainable use of scarce groundwater resources. The so-called Tarifa 9 is a subsidy on the tariff for electricity use in agriculture, and is mainly used for pumping groundwater from wells. The actual cost of generating and transmitting water in Mexico is estimated to be MxP 1.44/kWh but farmers pay on average MxP 0.33/kWh, i.e. only 23 percent of the actual cost.30 The subsidy implies that farmers pay on average 29 percent of the electricity they consume, whereas industrial users pay 94 percent and urban domestic users 43 percent. The total cost of the electricity subsidy for farmers was estimated to be MxP 8.0 billion in 2006.31 Underpricing of water together with Tarifa 9 have strong environmental impacts, promoting the overuse of Mexico's scarce water resources. The effects of Tarifa 9 are particularly detrimental on groundwater resources, which also have serious consequences for climate change adaptation strategies (see Box 2-3), also the findings of World Bank (2008d) shows that it has very little positive impact on poverty reduction since the majority of Tarifa 9 beneficiaries are well-off farmers and poor farmers typically do not pump groundwater, instead they use water wheels (norias) or surface water.32 Box 2-3: The role of agricultural subsidies in Mexico's water crisis Mexico now faces a "water crisis" that includes the overexploitation of 102 of its 653 aquifers, accounting for more than half of groundwater extraction in the country. The National Water Commission (CONAGUA) estimates groundwater over-extraction at almost 40 percent of total groundwater use. The value of the over-extracted groundwater in agricultural production alone is estimated at more than US$1.2 billion or 0.2 percent of GDP. The depletion of many aquifers leads to non-price and unregulated rationing, distorting growth in Mexico's most dynamic economic regions. 29 World Bank (2007) 30 Muñoz Piña, Avila Forcada, et. al (2006). 31 Primer Informe de Gobierno del Presidente Calderon as cited in Gomez and Santillanes (2008). 32 Of the 105,000 recipients of the subsidy, 68,000 receive an annual subsidy of less than MxP 30,000, while 33 users receive a subsidy of over MxP 500,000. This amounts to a Gini coefficient of 0.91 for the distribution of the subsidy (World Bank, 2007a). 27 The map below shows the locations of the 188 most important aquifers and their levels of exploitation. Most of the extremely overexploited aquifers are in the central-northern regions of Mexico where water scarcity is an issue and where the most active water-using economic sectors operate. The geographical distribution of this resource implies scarcity where the economic activity is taking place. The 188 Most Important Aquifers and their Levels of Exploitation Source: INE (2006) using data from CNA, Water Statistics in Mexico, 2005 edition Some apparent contradictions between the Constitution and national water laws and regulations further complicate this sort of rationing, particularly in relation to disenfranchised populations. Although some irrigation is shifting to water-saving technologies, the shift is limited, and the crop mix remains largely the same because (i) water and electricity prices still give the false signal that water is abundant, and (ii) irrigation infrastructure is insufficient to allow farmers to shift to specialty crops. Moreover, agricultural producers benefit from low electricity tariffs for pumping (tarifa 9). Consequently, farmers have little incentive to change current practices, which result in over-pumping of aquifers, lowering of the groundwater tables, and in many cases the intrusion of salt water. In addition, the financial cost to society of the approximately US$700 million per year electricity subsidy may represent only a fraction of the full economic cost, since environmental degradation is not valued properly. One such example of environmental degradation caused by the inappropriate use of groundwater is the deterioration of semi-aquatic ecosystem wetlands. These ecosystems are considered among the world's most important because they provide a unique habitat to a large number of species of flora and fauna. They also support the migration cycles of aquatic birds, 12 species of which winter in Mexico's wetlands. For this reason Mexico has supported the preservation of these habitats by incorporating 51 wetlands in the Ramsar Convention. However, many of these sites are threatened by the lack of fresh water provided by groundwater. Such is the case of the Xochimilco wetlands, as well as the Lerma and Aguascalientes springs and many of central Mexico's major lakes (Chapala, Cuitzeo, and Pátzcuaro). Source: Muñoz Piña, C., Avila Forcada, S. et. al., (2006) and Asad and Dinar (2006) 28 RELATIVE SIZE OF ARD EXPENDITURES 41. The fiscal costs outside PEC quantified above amount to MxP 32.6 billion, equivalent to 21.3 percent of PEC.33 Including these costs, total fiscal expenses for ARD increase to MxP 185.5 billion (Table 2-13). Both tax exemptions and the electricity subsidy can be classified as production-oriented, private goods type subsidies, thus changing somewhat the balance among expenditure categories. By including these fiscal expenses, the share of public goods fall from 51.7 to 42.6 percent of total ARD expenditures, while production-oriented expenses increase from 44.4 to 54.1 percent (Table 2-5 and Table 2-13). Table 2-13: Total Fiscal Expenses for ARD Including PEC and other Fiscal Costs in 2006 Indirectly Total Cost Social Productive Productive MxP Percent MxP Percent MxP Percent MxP Percent (%) Million (%) Million (%) Million (%) Million Total 185,480.98 100.00 69,912.94 100.00 100,417.24 100.00 15,150.80 100.00 Public 79,120.68 42.65 40,186.94 57.48 24,043.74 23.94 14,890.00 98.28 Private 106,360.30 57.35 29,726.00 42.52 76,373.50 76.06 260.80 1.72 Source: Oliver and Santillanes (2008) using data from SHCP 42. Total rural expenditure (including universal social services) is estimated to be MxP 382 billion, which represents a quarter of total public spending and 4 percent of national GDP. The estimate of ARD fiscal expenses in Table 2-13 does not include rural public spending on universal social services (general education and health services, social security, and other general subsidies such as the residential electricity subsidy). The rural share of these public expenditures was estimated at 27.5 percent in World Bank (2004a). Adding these to the former estimate of ARD expenditures implies that total rural spending is approximately MxP 382 billion, representing a quarter of total public spending and 4 percent of national GDP (Table 2-14). 43. This represents a very significant fiscal effort on behalf of the rural population. Given the current share of the rural sector in the national population (24 percent), it implies that average public expenditure per capita is now similar in the urban and rural sectors, without an "urban bias" in the allocation of public expenditures and subsidies. In fact, when taking into account the contribution to the fiscal revenue, there may even be a "rural bias" in the expenditure. However, there is evidence that the quality of public services is significantly lower in rural areas.34 44. The magnitude of these expenditures is evident if measured in relation to the weight of the ARD sector in the national economy. Agricultural GDP was 3.5 33 Since only two fiscal expenses were quantified (tax exemptions and tarifa 9), the actual total fiscal expenses are expected to be higher. 34 See for instance World Bank (2004c) Chapter 3.A, and OECD (2007d) pages 132-33. 29 percent of GDP in 2006, but rural GDP is considerably larger. Using ENIGH data, the share of the rural sector in national GDP is estimated to be 11 percent .35 From the above, the total APE for 2006 (agricultural programs within the PEC plus the agricultural tax exemption and tarifa 9) as estimated in this study (MxP 80,108 million), represents 43 percent of Agricultural GDP, while total ARD expenditures (MxP 185,481 million, consisting of the entire PEC plus agricultural tax exemptions and tarifa 9) and total rural expenditures (including also universal social services, social security and residential electricity subsidies, MxP 381,636 million) represent, respectively, 18 percent and 38 percent of rural GDP (Table 2-14). Table 2-14: Agricultural and Rural GDP and Public Expenditure National Agriculture Rural (pop. 0-2,500) Share Share (%) Share (%) (%) 6.0 14% (1) Population (2006, million) 104 100.0% (workers) (workers) 24.5 23.6% (2) GDP (2006, MxP million) 9,175,564 100.0% 318,158 3.5% 1,007,807a 11.0% (3) Public RURAL expenditure (PEC+foregone ag. tax+tarifa 9 , MxP million) 185,481 100.0% 80,108 43.2% 185,481 100.0% (4) Public expenditure ((3) + other universal servicesb, MxP million) 996,037 100.0% - - 381,636c 38.3% (5) Expenditure/GDP [(3)/(2)] - 2.0% - 25.2% - 18.4% (6) Expenditure (incl. universal social services)/GDP [(4)/(2)] - 10.9% - - - 37.9% Source: Rural population from CONAPO's population projection, Agriculture labor population from ENOE (Encuesta Nacional de Ocupación y Empleo) Note: (a) Rural GDP estimated using ENIGH data on share of agricultural and non-agricultural income (approximately 25/75) in rural households (b) "Other universal services" include universal social services (general health and educational services), social security, and residential electric subsidy (c) Rural share of universal social services estimated using 2006 spending figure and a share of 24.2% for the rural sector (World Bank, 2004a) 35 To obtain an estimate of rural GDP, the relative share of agricultural/non-agricultural income in the rural sector as observed in ENIGH (approx. 25/75) is used to estimate non-agricultural rural GDP. 30 31 3. EFFICIENCY ASSESSMENT 45. This section assess the relationship between Agricultural Public Expenditure (APE) and agricultural production and productivity from three complementary perspectives: (i) a comparative analysis of APE and agricultural GDP growth rates using international data, (ii) an analysis of the correlation between APE and production and productivity growth at the state level, and (iii) an analysis of the evolution of agricultural support resources with growth rates for production, cultivated land, and yields of different crops. The assessment is limited to agriculture programs, i.e. to APE, and excludes non-agricultural rural development programs, since there is no uniform international/state level dataset that can be used to compare the budgets and performance of rural development programs. 46. Two sets of data are used for the international comparisons, from the Food and Agriculture Organization (FAO) and the Organization for Economic Co-operation and Development (OECD). The FAO data is the GPAGRURAL database produced by the FAO Regional Office for LAC.36 Although excellent for the period covered (1987-2001), unfortunately it has not been updated since 2001. The OECD data on APE is available as an annual exercise for its member countries. The OECD monitoring effort is both more recent and allows more detail in the breakdown of agriculture-related expenditures and supports. It also allows for a comparison with developed countries. COMPARISON WITH LATIN AMERICAN COUNTRIES 47. Mexico stands out in LAC as the one country spending more that the rest in agriculture measured by a variety of indicators. Relative to total government spending, Mexico spent 6 percent on agriculture in the 5-year average of 1997- 2001 (LAC average was 3.4 percent; Table3-1). This was the highest amount in LAC. Relative to the size of the sector, Mexico also had the highest percentage, 23 percent against a LAC average of 11 percent (Figure 3-1). APE share in public expenditure relative to the sector's share of GDP, captured in the Agricultural Orientation Index (AOI), was again the highest: 1.6 compared to a LAC average of 0.5, with Mexico clearly showing a bias in favor of agricultural spending (Table 3-2). 36 FAO (2005b) 32 Table 3-1: APE in LAC (1987-2001): 5 year average, % of total public spending YEARS 1987-1991 1992-1996 1997-2001 Argentina 2 1.4 1 Bolivia 3.1 0.6 2.4 Brazil 5 3.5 3.6 Chile 2.2 2.4 2.6 Colombia n.a. 1 0.4 Costa Rica 22.4 10.1 4.3 Cuba n.a. n.a. n.a. Ecuador 4.3 4.1 5.1 El Salvador 1.6 0.4 2.1 Guatemala 4.4 2.7 2.9 Honduras 2.3 1.6 2.8 Jamaica 2.9 1.2 0.7 Mexico 9 9.1 6 Nicaragua 5.5 12 9.3 Panama 4.3 2.9 4 Paraguay 5.1 4.3 4.3 Peru 2.1 3.2 2.6 Dominican Rep. 18.9 11 8.3 Uruguay 2 1.7 1.9 Venezuela 3.2 1.4 0.8 LAC 5.6 3.9 3.4 Source: FAO (2005b)37 as cited in Valdes (2008a). Note: n.a. = non available 37 In the FAO study the source of public expenditures figures is the IMF Government Financial Statistics. 33 Figure 3-1: APE relative to Agricultural GDP (%, 1992-2001) %A GDP 35.0 30.0 1992-1996 1997-2001 25.0 20.0 14.7 15.0 10.6 10.0 5.0 0.0 te r ua al a y a la a en as o ay a ia or á on a . os ile C Ja ú do .D l om ua i al tin ic bi am ic c ue R ras er iv LA ad ag r gu ai C Ch m éx om va R du ag en ol P ez m an cu B ru ar ta M B rg ar ol ua ep U ic P E lS C P A H N V G E Source: FAO (2005b) as cited in Valdes (2008a) Table 3-2: Agriculture Orientation Index (AOI): Agriculture expenditure share relative to agriculture's share of GDP YEARS 1985-1990 1995-2001 Mexico 7.1 1.6 Venezuela 1.7 1.2 Dominican Rep. 1.1 0.7 Brazil 1.5 0.7 LAC 111.2.2.2 0.55.5 Nicaragua n.a. 0.5 Panama 1.6 0.5 Costa Rica 3.5 0.5 Ecuador 1.4 0.5 Chile 0.2 0.3 Peru 0.2 0.3 Uruguay 0.5 0.3 Argentina 0.7 0.2 Paraguay 0.2 0.2 Honduras 0.2 0.2 El Salvador 0.2 0.2 Bolivia 0.2 0.1 Guatemala 0.1 0.1 34 Jamaica 0.4 0.1 Colombia n.a. n.a. Cuba n.a. n.a. Source: Valdes (2008a). Based FAO (2005b). GDP and Agriculture GDP from the World Development Indicators, and from Government Finance Statistic Database (IMF). 48. Comparatively, the impact of Mexico APE on agricultural growth seems small. Figure 3-2 shows annual growth rates of agriculture jointly with the ratio of public agricultural spending relative to agricultural GDP. In 1997- 2001, Mexico spent on agriculture an amount equivalent to 23 percent of agricultural GDP while attaining a sectoral growth rate of less than 2 percent. In this figure, well performing countries are located to the left and above the fitted line. Mexico is the country furthest away below the line. Figure 3-2: Agricultural growth rate in LAC in relation to APE (as a percentage of agricultural GDP), Average (1997-2001) AGDP Growth Rates (%) 7.0 6.0 Dominican Rep. Panama 5.0 Peru 4.0 Nicaragua Brazil Costa Rica Ecuador 3.0 Chile Guatemala Argentina 2.0 LAC Bolivia Mexico Paraguay Venezuela 1.0 Colombia Honduras El Salvador 0.0 0.0 5.0 10.0 15.0 20.0 25.0 % of AGDP Source: Valdes (2008a) using FAO (2005b). Agricultural GDP growth rates are based on World Development Indicators Database. 49. Similarly, the impact of APE on Total Factor Productivity (TFP) is also low. Figure 3-3 compares growth rates in agricultural GDP and TFP over 1981-2001 with average APE/AGDP expenditure rates for 1985-2001, ordering countries by the latter. The figure suggests, if anything, a negative correlation between the countries' APE and the growth of AGDP and TFP. With the exception of Costa Rica, the six highest spenders (above 15% of agricultural GDP), show the lowest agricultural GDP growth rates over the period. On the other hand, the high growth 35 economies in both GDP and TFP are concentrated in the lower and middle end of the APE distribution. Figure 3-3: Average yearly agricultural GDP and TFP growth rates and distribution of APE/AGDP (%, 1981/5-2001) 5.0% 40% 35% 4.0% 30% 3.0% 25% GDP, TFP growth APE/GDP (%) 2.0% 20% 15% 1.0% 10% 0.0% 5% a a a r y s a la a ic u a ico le ca ay zil ica do ra ua ivi gu am bi tin al er bl ue hi ra ai gu R ex m du om P ol pu a C ag ra en am B ez an ta cu ru te M B on ica ar Re ol rg en os P U ua E J P C H A N n V C G ica in om 1.0% 0% GDP 19812001 TFP 19812001 APE/GDP 19852001 D Source: APE and Agricultural GDP from FAO (2005b); Agricultural TFP growth rates from Avila and Evenson (2004). COMPARISON WITH OECD COUNTRIES 50. Total government support to the agricultural sector has followed cyclical trends. The total transfers to the sector, measured by OECD's Total Support Estimate (TSE) is defined as the sum of transfers to producers from taxpayers and consumers, net of budget revenues (the import receipts associated with MPS programs). The TSE of Mexico has followed a broadly cyclical pattern: it declined in the second half of the 80s (following the 1983 crisis and 1986 trade liberalization through GATT), increased significantly between 1989 and 1994 (reaching its highest historical level in real terms in 1993), collapsed in 1995 following the "tequila crisis", expanded between 1996 and 2002, fell in 2002- 2004, and started to grow again after 2004. Also, the use of the most distortionary producer support instruments declined in the 1990s but resurfaced again since 2000. 51. Mexican support to the sector is slightly lower than the OECD average and similar to USA levels. Figure 3-4 shows average TSE for the period of 2003-05 for OCED members and selected MICs, as percentage of national GDP. Mexico's TSE is 0.90 percent of national GDP, compared to an OECD average of 1.2 36 percent and a USA average of 0.86. However, the level is significantly higher compared to countries such as Brazil, Chile, and New Zealand. The allocation of Producer Support Estimate (PSE) and GSSE in Figure 3-5 shows that compared to OECD average, Mexico has a higher allocation for budget support to farmers (Mexico 59 percent and OECD average 38 percent in 2006), but a lower allocation for MPS (Mexico 30 percent and OECD average 40 percent in 2006) and GSSE (Mexico 11 percent and OECD average 23 percent in 2006). Mexican PSE levels are slightly lower than OECD average levels, similar to USA levels and much higher than Brazil, Chile, and New Zealand (see Figure 3-6). Figure 3-4: Total Support Estimate (TSE) for Select Countries (as % of national GDP; 2003-05 average) Turkey China Russian Federation Japan European Union OECD Mexico United States Canada Brazil Chile New Zealand Australia 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 Source: OECD, PSE/CSE Database, 2007 and OECD Factbook 2008: Economic, Environmental and Social Statistics Note: EU15 for 2003 and EU25 for 2004-2005. Chile Estimates based on Agricultural Policy Note, OECD. 2008 Figure 3-5: Mexico and OECD's PSE and GSSE (%, 1990/97-2006) 37 Mexico (1997-2006) 100.00% Budget 80.00% Support to Farmers 60.00% PSE 40.00% MPS 20.00% GSSE GSSE 0.00% 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 OECD (1990-2007) 100% Budget Support to 80% Farm ers PSE 60% MPS 40% 20% GSSE GSSE 0% p 19 0 19 1 92 93 19 4 19 5 19 6 19 7 98 20 9 20 0 20 1 20 2 20 3 04 20 5 20 6 07 9 9 9 9 9 9 9 0 0 0 0 0 0 19 19 19 19 20 Source: OECD, PSE/CSE Database 2007 Figure 3-6: Producer Support Estimate (PSE) (% of agricultural GDP; 2003-05 average) 38 60 50 40 30 20 10 0 D n es 1 le n a o ey li a il d tio EC EU pa az ad ic an hi at rk ra ex ra C Br Ja an O St al Tu st de M Ze C Au d Fe te ew ni an U N si us R Source: OECD, PSE/CSE Database, 2007 and OECD Factbook 2008: Economic, Environmental and Social Statistics. Chile Estimates are based on Agricultural Policy Note, OECD. 2008 Note: 1/EU15 for 2003 and EU25 for 2004-2005 52. Expenditure in public goods, earlier identified as being underfunded in Mexico, is higher for marketing and promotion activities and lower for infrastructure and research and development compared to some OECD and MIC countries.38 For all the countries included in Table 3-3, infrastructure and research and development are two categories of common significance. In the case of Mexico, however, spending on marketing and promotion is the highest category (although far less than in the United States), and spending on infrastructure and research and development is relatively low, particularly compared to New Zealand and also to Brazil and Chile. Knutson (2007) also points out that Mexico would need to spend an additional MxP 1.1 billion a year in Sanitary and phyto-sanitary inspection systems if it were to match the US system.39 38 Brazil and Chile is included as part of OECD's regular monitoring and evaluation effort of agricultural policies for a few non-member countries. 39 As a benchmark, Knutson (2007) indicates that in the US, US$0.01 is spent on sanitary and phyto- sanitary system per dollar of crops and livestock sales. Applying this benchmark figure to the Mexican sale of crops and livestocks in 2006 (MxP 303 billion or US$25.3 billion), the cost to the Mexican government of a US equivalent SPS system would be MxP 2.8 billion. This compares with the 2006 Mexican sanitary and phyto-sanitary budget of approximately MxP 1.7 billion. 39 Table 3-3: Composition of General Service Support Estimate in Selected Countries (%, Average 2004-2005) United European New ITEMS Mexico Brazil Chile Switzerland States Union Zealand Research and 16.2 30.7 23.0 5.9 16.4 40.3 18.1 Development Agricultural 21.5 13.7 1.1 0.0 7.8 8.6 3.7 Schools Inspection Services 16.5 3.7 9.3 2.6 6.0 32.3 2.4 Infrastructure 13.8 44.5 55.2 14.0 42.4 18.4 18.1 Marketing and 31.2 0.4 11.0 70.5 19.5 0.0 11.2 Promotion Public Stockholding 0.0 6.9 0.0 0.4 7.3 0.0 8.6 Miscellaneous 0.7 0.0 0.8 6.6 0.6 0.2 37.9 Total % 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Source: Valdes (2008a) estimates based on OECD PSE/CSE database, 2005/2007. Note: Brazil is only for 2004. Chilean agricultural schools are almost without exception privately run by not-for-profit organizations, the budgets of which are not counted here. APE AND AGRICULTURAL PERFORMANCE AT THE STATE LEVEL 53. The allocation of APE by states is correlated to the size of their agricultural sectors. Figure 3-7 shows that the distribution of APE and agricultural GDP are closely correlated, in fact, the correlation coefficient is 0.70, although some states (Tamaulipas, Zacatecas, Tlaxcala, Sinaloa, etc.) receive disproportionately higher shares in relation to their contribution to agricultural GDP, while others (Morelos, Baja California Sur, Aguascalientes etc.) receive disproportionately lower shares. Figure 3-8 shows a map highlighting the APE received by states relative to their agricultural GDP. It shows that the states receiving higher proportions of APE are those in the northern and southern parts of the country, whereas the majority of the central states receive less. 40 Figure 3-7: State Shares in National Agricultural GDP and APE (2005-2006) (%; States ordered from left to right by descending level of state APE received in 2006) 14% 12% 10% 8% 6% 4% 2% 0% ah o G u Chia a s M aju s P u ca Nu ab ila e r a la al lie a Y u c he So co ca uz is ico al ro o o Tl tán M taro Ch ulip a r a n M bla Ba gua Co los Qu forn ntes ih a s m r it o o Oa go Fe o ra n Na ón Ja ua Hi nia ja er í st an ur l Ba Gu otos ra a n pa j a s c lim m lo Ve nor Co lg e v as c o Ro ich at Du acá c C re Di int ia S xa Za ac r Ca y a T u lis n h Le de e Lu éx Qu axc da T a Sin a r e te ca pe ua ifo r it a or é P a C a i Sa A APE AGDP Source: Scott (2008) Figure 3-8: Map of Mexico Indicating the Share APE/Agricultural GDP (2006) Note: Dark blue states = APE is more than 16 % of agricultural GDP (12 states) Medium blue states = APE is between 15 % and 11.1 % of agricultural GDP (11 states) Light blue states = APE is less than 11 % of agricultural GDP (9 states) 41 54. States receiving larger amounts of APE generally contribute proportionally more to national agricultural growth. This is shown in Figure 3-9 where the contribution of states to the growth of national agricultural GDP in the period 1994-2006 is plotted.40 States on the left hand side of the figure, which receive more APE than those on the right hand side, tend to have higher contributions to national agriculture GDP growth rate. Regression results show that a 10 percent increase of APE spent in a state is associated with a 0.2 percent higher state contribution to national agricultural growth.41 Figure 3-9: Average Contribution of the States to the Growth Rates of National Agriculture (Million MxP 1993 prices) [States ordered from left to right by descending level of state APE received in 2006] 200.00 150.00 million MXP (1993 prices) 100.00 50.00 0.00 Nuevo León Baja California Sur Jalisco Baja California Tabasco Tlaxcala San Luis Potosí Tamaulipas Zacatecas Sinaloa Michoacán Durango Chihuahua Sonora Guanajuato Puebla Oaxaca México Hidalgo Guerrero Coahuila Campeche Yucatán Querétaro Colima Quintana Roo Distrito Federal Nayarit Chiapas Veracruz Morelos Aguascalientes -50.00 Source: Authors' calculations based on National Account figures 55. Examining APE efficiency using state level data: Efficiency of APE can be examined by looking upon APE as an investment which improves productive technology. Thus, a higher level of normalized APE (APE/ Agricultural GDP) in year t would be associated with a higher growth rate of Agricultural GDP in years t+1, t+2, etc. Looking at it this way, in order to maximize agricultural growth, APE should be distributed among the states so as to equalize the marginal effect of APE on growth rates. If we find that rates of growth are low in states where normalized APE (APE/AgGDP) is high, this suggests that there are potential gains from reallocating APE from these states to states where the converse is true. 40 This contribution is calculated by multiplying the average annual agricultural GDP of the state in 1994- 2006 by the state's average annual agricultural growth rate in that period. 41 The correlation is less than perfect with a correlation coefficient of 0.55. The correlation is significant at the 1% level (p value = 0.001). 42 56. The allocation of public expenditure to the states is not related to the growth of their agriculture. The relation between the level of normalized APE received and the rate of growth of agriculture is shown in Figure 3-10, which exhibits no correlation between the two variables.42 Under the assumption that the share of allocations to the states has remained reasonably stable over the period, the conclusion that follows is that the federal government has tended to allocate more public expenditure to states with large agricultural sectors but not to those growing faster. The association between APE and the contribution to national agricultural growth is explained, therefore, by the effect of size in the allocation of APE as shown previously in Figure 3-9, not by the fine-tuning of allocations according to the growth potential of the states, as observed by historical growth rates. Figure 3-10: Average Annual Growth Rate of State Agricultural GDP (%, 1994-2006) [States ordered from left to right by descending level of normalized state APE received in 2006] State agricultural growth rate (average of 1994-2006, %) 4.50 4.00 3.50 3.00 2.50 2.00 1.50 1.00 0.50 0.00 Nuevo León Baja California Sur Jalisco Baja California Tabasco Tlaxcala San Luis Potosí Tamaulipas Zacatecas Querétaro Sinaloa Chihuahua Guanajuato Michoacán Durango Puebla Oaxaca México Hidalgo Guerrero Coahuila Campeche Yucatán Colima Sonora Quintana Roo Distrito Federal Nayarit Chiapas Veracruz Morelos -0.50 Aguascalientes Source: Authors' calculations based on National Account figures 57. There is scope for raising the efficiency by re-orienting APE to states which exhibit higher impact of spending. Given that the growth rate in agriculture tends to be higher in states with smaller agricultural sectors, the marginal benefit of APE is higher in those states compared to states with large agricultural sectors, whose growth rates are lower. As a result, small agricultural states have a much higher investment efficiency (or "bang for buck")43 of APE than large agricultural states, which receive a disproportionately higher share of APE. For example, 42 Correlation coefficient is -0.22, which shows little correlation between the two variables (normalized state APE and state agricultural growth rate). Thus, correlation is statistically insignificant (p value = 0.21). 43 This is calculated by multiplying the average agricultural GDP of the state in 1994-2006 by the state's agricultural growth rate in that period, divided by APE in 2006 of that state. 43 Morelos, which received one of the least APE as a share of its agricultural GDP, shows MxP 21 of agricultural GDP for every peso of APE that was allocated to the state, whereas Tamaulipas which received the second most APE as a share of its agricultural GDP shows only MxP 0.4 of agricultural GDP for every peso of APE allocated to the state (see Figure 3-11). This would suggest the existence of potential efficiency gain in re-allocating APE in favor of those states which exhibit a higher growth per peso of spending. Figure 3-11: Investment efficiency of APE ("Bang for Buck") [states ordered from left to right by descending level of APE/agricultural GDP] 25.00 MxP of APE per Ag. GDP spent 20.00 15.00 10.00 5.00 0.00 Nuevo León Baja California Sur Jalisco Baja California Tabasco Tlaxcala San Luis Potosí Tamaulipas Zacatecas Sinaloa Chihuahua Sonora Guanajuato Michoacán Durango Puebla Oaxaca México Hidalgo Guerrero Coahuila Campeche Yucatán Querétaro Colima Quintana Roo Distrito Federal Nayarit Chiapas Veracruz Morelos Aguascalientes -5.00 Source: Authors' calculations based on National Account figures 58. Labor and land productivity also appear to be uncorrelated with APE (see Figure 3-12). The four states with the highest APE/AGDP rates (Tamaulipas, Tlaxcala, Zacatecas and Campeche) present the lowest land productivities among all states except Yucatan. On the other hand, productivity growth is roughly U- shaped: it is positive for some of the states with largest shares of APE, negative for most states in the middle and again positive for the states with the smallest APE shares. 44 VP/worker(hctr) Value of production/worker Ta m Ta au m 80% 60% 40% 20% 0% 20% 40% 60% 80% 100% l au 0 20 40 60 80 100 120 140 Tl ipas a Tl lipa Za xca Za axc s c l c a Ca ate a Ca ate la m cas p m ca p s Qu S ech Qu S ech Source: Scott (2008) in ina e in ina e ta l o ta l o na a na a Ch Roo Ch Roo ia ia So pas S pa Ba Chi ono s Ba Chih nor ja hu ra ja u a a C ah C h al u ifo a al ua ifo r T rn Gu Tab nia Gu ab ia an asc an asc aj o aj o u u H i a to da Hi ato da Co lgo Co l g o l li Oa ima Oa ma x Na a c a xa Gu y a Na c a er rit Value of production/hectare r 45 Gu yar Change in VP/hctr er it M ero r Sa n P ico éx M er o Lu ue Sa éx is b l n Pu ico Po a Lu e Yu tos is bl Po a c í D atá Yu tos c í Qu uran n er g o Du atá Qu n ra n Ve étar ra o er go é c Ve tar Ja ruz ra o Co lisc Nu a o cr ev h u Ja u z i Change in VP/worker Co l i s c Ag Mic o Le la Ba u h ó Nu a h o Percentage change in labor and land productivity: 20002004 ja as oa n Value of production/worker ev ui C ca cá o la A M L Di alifo lien n s t rn te Ba gu icho eón rit ia s ja c cas a o S C a á a l n Fe ur d Di lifo ient M era and Percentage Change in Labor and Land Productivity (%, 2000-2004): st rn es or l rit ia el o S os Fe ur [states ordered from left to right by descending level of APE/agricultural GDP] d M era 0 5 or l 10 15 20 25 30 35 40 45 el os Value of production/hectare Figure 3-12: Average Value of Production per Worker and Hectare (thousand MxP 2004, 2000-2004, 59. There is a negative correlation between spending in private agricultural goods and state level agricultural GDP growth. Regression results in Table 3-4 show that, controlling for the levels of mechanization, fertilization and expenditures in public goods, a 10 percent increase in APE on private goods as a percentage of the value of agricultural production is associated with a 2.6 percent reduction in Agricultural GDP growth.44 This result may seem counterintuitive, but can be explained by several factors. First, and most importantly, APE and infrastructural investments have been concentrated historically in the largest and most developed agricultural states, where additional growth potential may thus be lower. Secondly, the size and concentration of federal APE may have a crowding out effect on private and state government investments in the largest beneficiary states. Finally, the results are consistent with a low productive impact of APE programs at the farm level. On the other hand, spending on public agricultural goods shows a positive but not statistically significant impact on agricultural GDP growth.45 Table 3-4: Effect of APE on Log Average Annual Agricultural GDP growth: 2000-2006 APE Private Goods (% value agr prod) -0.261 [0.014]* APE Public Goods 1.113 [0.316] Mechanization (% of cultivated land) 0.052 [0.006]** Fertilization (% of cultivated land) -0.042 [0.031]* R-squared 0.712 Robust p values in brackets * significant at 5%; ** significant at 1% APE AND THE PERFORMANCE OF DIFFERENT CROPS OVER TIME 60. MPS and output-based ASERCA payments have targeted mostly traditional crops, and not fruits and vegetables. Figure 3-13 to Figure 3-15 compare the evolution of agricultural support resources with growth rates for production, cultivated land, and land productivity of grains, vegetables and fruits over the period 1980-2004. Both MPS and output-linked ASERCA payments have targeted mostly traditional crops, particularly maize and other grains, as well as raw sugar and some animal products like milk and poultry meat (Figure 3-13). 44 This result is in accordance with results by Lopez and Galianato (2007) showing that, ceteris paribus, reducing the share of subsidies to private goods (or, equivalently, increasing the share of public goods) in the government's budget has a large and significant positive impact on rural per capita income. 45 Lopez and Galinato (2007) also show using a 1985-2001 dataset on rural expenditure for LAC that reducing the share of subsidies to private goods (or, equivalently, increasing the share of public goods) in the government's budget has, ceteris paribus, a large and significant positive impact on rural per capita income. 46 Fruits and vegetables, on the other hand, have not received significant support, but have benefited from the liberalization of agricultural markets. Figure 3-13: Support Based on Specific Commodity Production: MPS and ASERCA (grains) (MxP Million, 1980 - 2004) 40,000 30,000 20,000 10,000 0 10,000 20,000 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 Grains MPS Grains ASERCA 50,000 40,000 30,000 20,000 10,000 0 10,000 20,000 30,000 40,000 50,000 60,000 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Diary Meat Sugar Other Source: OECD (2007) as cited in Scott (2008). 47 61. Corresponding to the surge in the support for grains, the production of grains expanded after 2000 while the growth of vegetables and fruits declined. As one would expect, after the liberalization reforms a significant increase is observed in the production of fruits and vegetables, but not of grains. This growth was associated with an expansion in cultivated land in the case of vegetables, and an increase in the productivity of land in the case of fruits. By contrast, after 2000, the growth of vegetable production slows down, and in the case of fruits declines, while grains grow at an average 7.5 percent annually, entirely through growing land productivity (Figure 3-14 and Figure 3-15). Note that the 1988-1994 and 2000-2004 periods present similar trends in the relative behavior of grain vs. fruits and vegetable production and cultivated land, in favor of the former. This coincides with the surge of MPS and output-based support for grains, as well as the expansion of variable input-based support, which is also mostly linked to the latter. 62. The data reviewed thus suggests a conflict between the market liberalization process initiated in the early 1990s and culminating in 2008, and agricultural support policies. Far from being resolved, this conflict has been revived in the present decade, with the gradual shift back towards more distortionary support policies. Subsidies have been biased towards traditional crops (grains) rather than supporting comparative advantage crops like fruits and vegetables. Especially in light of favorable conditions brought about by the current high commodity prices for precisely these traditional crops, subsidizing these crops does not make sense from an efficiency point of view. The findings of an analysis of Mexico's agricultural policies including trade policies in World Bank (2008c) quantify this anti-export bias. The study shows that producers of several import competing crops such as certain grains and dairy benefit from huge income transfers (with a positive nominal rate of assistance of 19.5 percent for the period 2000-04), while producers of exportables such as fruits and vegetables are taxed (with a negative nominal rate of assistance of 12.5 percent).46 46 The Nominal Rate of Assistance (NRA) is defined as the percentage by which government policies raise gross returns to producers above what they would be without government's intervention (or lower them if the NRA is less than zero). This concept is similar to the OECD's Total Support Estimate (TSE) but there are several methodological differences between the two (World Bank, (2008c)). The overall NRA for importables is positive over the period 2000-04 indicating government support to these crops, but with a large degree of variation in the level of assistance to specific products. Some major importables showed almost zero or even negative NRAs (barley, maize, sorghum, soybeans, beans and eggs), while products such as wheat, rice, milk, sugarcane and chicken meat showed high NRAs (40-80 percent). On the other hand, NRAs for exportables are negative over the same period except for beef, indicating that exportables have generally been taxed. The tax was especially high on coffee and tomatoes, exceeding 40 percent in some years. 48 Figure 3-14: Index of Production, Cultivated Land, and Land Productivity in Grains, Vegetables and Fruits (Base year is 1980= 1, 1980-2004) Production 2.5 2.0 1.5 1.0 0.5 0.0 1981 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Grains Vegetables Fruits Cultivated Land 2.5 2.0 1.5 1.0 0.5 0.0 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Grains Vegetables Fruits 49 Productivity (Ton/ha) 2.5 2.0 1.5 1.0 0.5 0.0 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Grains Vegetables Fruits Source: SIAP, SAGARPA as cited in Scott (2008) Figure 3-15: Average Annual Growth Rates in Production, Cultivated Land, and Land Productivity in Grains, Vegetables and Fruits (%, 1980-2004) P r o d u c t io n 1 2 % 1 0 .5 6 % 1 0 % 8 % 7 .5 0 % 7 .5 5 % 6 % 4 .0 8 % 4 % 2 .2 6 % 2 .1 9 % 2 % 0 .1 6 % 0 % 0 .5 0 % 2 % 1 .2 2 % 1 9 8 8 1 9 9 4 1 9 9 4 2 0 0 0 2 0 0 0 2 0 0 4 G ra in s V e g e t a b le s F ru it s 50 C u l t i v a t e d l a n d 8 % 6 .1 2 % 6 % 4 % 3 .1 1 % 2 % 0 .8 4 % 0 % 0 .1 5 % 0 .9 5 % 2 % 2 .3 5 % 2 .3 6 % 2 .4 4 % 4 % 4 .5 5 % 6 % 1 9 8 8 1 9 9 4 1 9 9 4 2 0 0 0 2 0 0 0 2 0 0 4 G ra in s V e g e t a b le s F ru its L a n d P r o d u c t i v i t y 9 % 8 .5 0 % 8 % 7 .4 5 % 7 % 6 % 5 % 4 .5 5 % 4 .0 5 % 4 % 3 .2 4 % 3 % 2 .5 1 % 2 .4 1 % 2 % 1 .3 8 % 1 .2 2 % 1 % 0 % 1 9 8 8 1 9 9 4 1 9 9 4 2 0 0 0 2 0 0 0 2 0 0 4 G ra in s V e g e t a b le s F ru its Source: SIAP, SAGARPA as cited in Scott (2008). 51 4. EQUITY ASSESSMENT 63. The equity of ARD expenditures is evaluated at the level of geographic units as well as that of individual producers and households. The geographic analysis is at the state level and also at the municipal level for programs for which there is information. The distribution of APE is analyzed ordering states by their extreme rural poverty rates (pobreza alimentaria), using the official CONEVAL measures for 2005.47 STATE AND MUNICIPAL LEVEL EQUITY ASSESSMENT 64. The distribution by states of APE is highly correlated with Agricultural GDP but not with extreme poverty. Figure 4-1 shows the cumulative distribution of ARD expenditures and the cumulative state shares of extreme poverty, Agricultural GDP and Oportunidades payments. There is a high degree of correlation between the distribution of APE and Agricultural GDP, but not between APE and extreme poverty. This is in contrast to the distribution of Oportunidades, which shows a high correlation with the distribution of extreme rural poverty. For example, the poorest six states are home to about 55 percent of the extreme poor, and receive about 48 percent of total benefits from Oportunidades but only 7 percent of total APE. 47 This measure is highly correlated with the multivariate Consejo Nacional de Poblacion (CONAPO) marginality index, and is used here in preference over the latter because it represents the current official poverty measure at the state level. Given the correlation, the results are not sensitive to the choice of measure. 52 Figure 4-1: Extreme Poverty, Agricultural GDP and ARD Expenditures (%, 2005-6) [States ordered from left to right by rural poverty rate] 60 100% 90% 50 80% 70% 40 Cummulative shares 60% Poverty rate 30 50% 40% 20 30% 20% 10 10% 0 0% s e l ca la a as o ico ra la oo ón oa ua to ra ca lim ch sc eb ui no iap ua xa al h de Le R éx te ba pe ah ua Pu Sin Co Oa na So aj o Fe M Ch ca Ta m Co ih an ev ta Za o Ca Ch Gu in Nu ri t Rural poverty rate Cummulative poverty shares Qu st Cummulative Agric. GDP shares Cummulative APE shares Di Cummulative Oportunidades shares Source: Scott (2008), CONEVAL for rural poverty rates, INEGI for Agricultural GDP, and Oliver and Santillanes for APE and Oportunidades expenditure 65. The distribution of APE per rural capita for the principal programs is concentrated in the upper half of the poverty-ordered state distribution, with the highest benefits allocated to the large agricultural states of Tamaulipas, Sinaloa, Chihuahua, and Sonora. These four states are among the principal per capita beneficiaries of PROCAMPO reflecting their agricultural land assets. But their disproportionate participation in APE is explained by the extreme concentration of Apoyos, diesel and the electricity subsidies (tarifa 9). At the other extreme, the poorest states receive support mostly from PROCAMPO and Alianza, obtaining barely a tenth of the per capita support benefiting the former states (see Figure 4-2). 53 Figure 4-2: Principal APE Programs (MxP, 2006) [states ordered from left to right by extreme poverty rate] 6,000 5,000 Anual spending per rural capita (MxP) 4,000 3,000 2,000 1,000 0 m as Ve l a So o D u che ev tes os is a t Ag Co ala Ba rn a ia ro Ta ruz ich co m n ih ro Gu pas M n Na co Sin as Pu sí er a Tl ra ali r a an go Qu Hid o na o Yu oo M ón Ja a Za yari C Su fo m s c ui l Lu ac Qu alo hu c Ca acá in alg tá t to eb rn no Ta atec re lip ua Ch éta el lis M ba s i Gu ran Le R éx c Nu lien ali Coli c ja ia ca n Oax ia pe u a ah Po ua fo ax or ra er au aj o o Ch c a ta C Sa ja Ba PROCAMPO (tradicional) PROGAN Apoyos a la comercialización Diesel Alianza para el Campo Source: Oliver and Santillanes (2008) as cited in Scott (2008) 66. State-wise, Alianza is better distributed than PROCAMPO, which in turn is better distributed than Apoyos. Considering individual programs, Alianza is the most progressively distributed at the state level, with 28 percent of transfers going to the poorest five states, followed by PROCAMPO, with 22 percent. The distribution of Apoyos is highly regressive, with the three poorest states receiving just 7 percent (see Figure 4-3). 54 Distribution Distribution Ch Ch 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 0% 2% 4% 6% 8% 10% 12% Gu iapa Gu iapa er s er s Sa re Sa re n Oa ro n Oa ro Lu x Lu x is aca is aca Po Po t t Distr. 2002 Distr. 2002 Pu osí Pu osí Ve ebl Ve ebl ra a ra a Ta cruz Ta cruz M ba M ba ic h sc ic h sc o o Ca oac Ca oac m án m án pe pe Distr. 2006 Distr. 2006 D ch D ch G u ura e G u ura e an ng an ng aj o aj o u u Qu H ato Qu H ato in ida in ida ta lg ta lg na o na o Yu Roo Yu Roo ca ca M tán M tán éx éx Na ico Na ico Cummul. Distr. 2002 Cummul. Distr. 2002 Za ya Za ya r r Ta cate it Ta cate it m ca m ca 55 au s au s lip lip S as S as Apoyos Qu ina Qu ina Procampo e lo e lo Ch rét a Ch rét a ih aro ih aro ua ua h h Ja ua Ja ua lis lis So co So co Cummul. Distr. 2006 Cummul. Distr. 2006 n n Tl ora Tl ora ax ax Ag C ca Ag C ca ua oa la ua oa la sc hu sc hu N u a lie ila Nu alie ila ev nte ev nte o s o s L L [States ordered from left to right by extreme poverty rate] Ba M eón Ba M eón ja or ja or C el C el a l C o os al Co os Ba ifor lim Ba ifor lim ja nia a ja nia a C S C S al u al u ifo r ifo r rn rn ia ia Figure 4-3: Distribution of the Three Main Agricultural Programs (%; 2002, 2006). 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 10% 20% 30% 40% 50% 60% 70% 80% 90% Cummulative rural poverty shares 2006 Cummulative rural poverty shares 2006 100% 100% Cummulative distribution Cummulative distribution Alianza 8% 100% 90% 7% 80% 6% 70% Cummulative distribution 5% 60% Distribution 4% 50% 40% 3% 30% 2% 20% 1% 10% 0% 0% Ja u a ra a Ch rét a Nu alie ila is aca Tl ora m án Gu ura e aj o Qu H ato Yu Roo M eón ifo r o na o N ico ih aro er s Pu osí au s Si a s ua oa la o s al Co os ja nia a ia Ta cruz M tán n Oa ro So co Ta cate it al u Gu iapa ev nte m ca e lo Ve ebl Ba ifor lim D ch ic h s c an ng ta lg r Ag C xca rn C S re lip el Za aya lis sc hu h Ca oac t éx Qu n a in ida u M ba n ca Lu x pe Po ua L or a Ch C Sa ja Ba Distr. 2002 Distr. 2006 Cummul. Distr. 2002 Cummul. Distr. 2006 Cummulative rural poverty shares 2006 Source: Scott (2008), calculated based on Oliver and Santillanes (2008) and World Bank (2004a) 67. State-wise expenditures on both public and private goods are regressively distributed. Interestingly, expenditures in public goods are more regressively distributed than those in private goods, with per capita benefits rising in the upper half of the state distribution (Figure 4-4). Expenditures in private goods are concentrated in the large agricultural states, Tamaulipas, Sinaloa, Chihuahua, and Sonora, reflecting the concentration of Apoyos, diesel and the electricity subsidies (tarifa 9). 56 Figure 4-4: Per Capita Expenditure in Public and Private Goods in APE (MxP, 2006) [States ordered from left to right by extreme poverty rate] 6,000 350 300 5,000 250 4,000 Private (MxP,bars) Public (MxP,line) 200 3,000 150 2,000 100 1,000 50 0 0 Nuevo León Zacatecas Aguascalientes Morelos Chiapas Veracruz Tamaulipas Oaxaca Campeche San Luis Potosí Puebla Sinaloa Sonora Tlaxcala Colima Chihuahua Coahuila Baja California Sur Baja California Guerrero Durango Nayarit Querétaro Tabasco Guanajuato Quintana Roo México Hidalgo Jalisco Michoacán Yucatán Private goods Public goods Source: Scott (2008). Based on data by Oliver and Santillanes (2008) 68. Results at the municipal level for PROCAMPO and Ingreso Objetivo confirm the regressive geographical distribution of APE. It is possible to analyze the distribution at the municipal level using administrative data for two main programs, PROCAMPO and Ingreso Objetivo, ordering municipalities by their extreme rural poverty rates. Both programs are regressively distributed, but Ingreso Objetivo extremely so, with high per capita payments for a small fraction of municipalities, and no payments for most of the rest. In comparison, PROCAMPO benefits are densely distributed throughout. Figure 4-5 shows that the poorest 70 percent of municipalities receive 40 percent of PROCAMPO transfers, but less than 6 percent of Ingreso Objetivo, and in the latter case these resources are concentrated in a few municipalities so that the great majority of the poorest half of municipalities (and all those in the poorest third) receive no transfers from Ingreso Objetivo. 57 Figure 4-5: Average per Rural Capita of PROCAMPO and Ingreso Objetivo Transfers by Municipalities (MxP, 2005-6) [ordered from left to right by extreme rural poverty rate] Procampo 100% 16000 90% 14000 80% Per rural capita expenditure (MxP) 12000 70% Cummulative distribution 10000 60% 50% 8000 40% 6000 30% 4000 20% 2000 10% 0% 0 Municipalities Procampo Pesos/cap Procampo Cumm.distr. Procampo Cummulative rural poverty distr. Ingreso Objetivo 100% 4500 90% 4000 80% 3500 Per rural capita expenditure (MxP) 70% Cummulative distribution 3000 60% 2500 50% 2000 40% 1500 30% 1000 20% 10% 500 0% 0 Municipalities Pesos/cap Cumm.distr. Cummulative rural poverty distr. Source: ASERCA administrative data bases; CONEVAL municipal poverty measures. 58 PRODUCER AND HOUSEHOLD LEVEL EQUITY ASSESSMENT 69. The availability of household and producer data bases reporting both agricultural support programs and a relevant measure of household/producer wellbeing or wealth to allow distributive analysis is limited. Three kinds of data sources are used, which are complementary but not strictly comparable: (i) general household surveys including ARD transfers (ENIGH 2004, 2006; ENNVIH 2002), (ii) evaluation surveys for specific programs (Alianza, Oportunidades), and (iii) administrative data of the programs (PROCAMPO, Ingreso Objetivo). 48 The distribution of benefits is analyzed using two types of ordering corresponding to the different data sources. In the case of administrative data, producers are ordered by land holdings, which is the only measure of wealth/welfare reported in these data. In the case of the national household surveys, benefits received are analyzed by population deciles ordered by (current) income per capita. Household and producer data allow coverage of the principal ARD programs, including the principal agricultural support programs, and rural social programs like Oportunidades and the Programa de Empleo Temporal, as well as a number of other smaller rural development programs. Using the distribution of irrigated land as a proxy for the distribution of hydrological, hydro-agricultural and tarifa 9 electricity subsidies, the agricultural support programs covered in this incidence analysis represent 75 percent of APE. 70. There is a high degree of regressivity for the major agricultural programs. To compare the principal APE programs, Figure 4-6 presents concentration curves derived from the administrative data for PROCAMPO, Ingreso Objetivo, and Alianza PDR. 49 For comparative purposes, the distribution of rainfed and irrigated land is also shown. There is an extreme concentration of benefits for all programs. The poorest producer decile (in terms of land) receives a tenth of a percentage point of Ingreso Objetivo and similarly insignificant fractions of the other programs. At the other end of the distribution, the richest producer decile receives the following shares of transfers: (i) 45 percent of PROCAMPO,(ii) 55 percent of the Alianza PDR, 50 , (iii) 60 percent of energy and hydrological subsidies (proxied by irrigated land), and (iv) 80 percent of Ingreso Objetivo. 48 General household surveys have the important advantage of being nationally representative and including high-quality data on income and other measures of household well-being, but their sample size is not designed to capture specific transfer programs accurately, especially when these have limited coverage or concentrate a large share of their benefits in a relatively small proportion of producers. The other two sources are of course designed to capture the program beneficiaries and transfers accurately, but are not nationally representative and generally contain limited or no income household data. The analysis obtained from the three sources must therefore be interpreted carefully, and in a complementary way. 49 The curve for Alianza PDR is derived from Word Bank 2006a. 50 The PDR is aimed at low-income groups through explicit targeting criteria. However, the Alianza evaluation data reveals a failure to comply with these criteria. The program's rules require that at least 70% of its resources be allocated to Very High or High marginality localities (as defined by CONAPO's marginality index), but in 2004 only 32% of the expenditures associated with PDR were spent in these localities ­ less than 2% in Very High marginality localities. 59 Figure 4-6: Concentration Curves of Major Agricultural Programs and Land (%) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Procampo Ingreso Objetivo Alianza-Desarrollo Rural Irrigated Land Rainfed Land Source: Scott (2008) using ASERCA administrative data, FAO (2005) and World Bank (2006). 71. PROCAMPO stands in contrast to Oportunidades. In Figure 4-7 the distribution of PROCAMPO and Oportunidades resources is compared using ENIGH data.51 The contrast between Oportunidades and PROCAMPO is evident from their concentration curves in the income space. Although both programs are more progressive than the income distribution, Oportunidades is highly progressive (pro-poor) while PROCAMPO is regressive. 51 The distribution of PROCAMPO benefits differs substantially according to whether ENIGH data or the administrative program data of ASERCA are used, the latter resulting in a substantially more regressive distribution than the former. This is because, as most household surveys, ENIGH does not capture well household incomes at the top of the distribution. In view of this, we report in some sections two sets of results for PROCAMPO, one using ENIGH data and the other using ASERCA data. 60 Figure 4-7: Comparison of PROCAMPO and Oportunidades Transfers by Rural Population Deciles (%, 2006) [ordered from left to right by pre-transfer income per capita] Rural distribution 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0% Oportunidades Procampo Income Source: Scott (2008), calculations are based on data by ENIGH 2006 72. ARD expenditures are an inefficient redistributive instrument while APE is not only inefficient but actually contributes to increasing income inequality. To assess the global impact of ARD expenditures on rural income inequality, the APE programs are contrasted with nine social and rural development programs reported in ENIGH 2006 and in a special "Social Module" commissioned by SEDESOL with ENIGH 2004. Concentration Coefficients (CC) are used to measure equity. 52 As shown in Figure 4-8, the CC of ARD expenditures is regressive in absolute terms (CC = 0.24), but still progressive relative to the distribution of pre-transfer income (CC = 0.35). APE is more regressive (CC = 0.58) than the distribution of pre-transfer income (CC = 0.35), thus actually contributing to increasing income inequality. By contrast, rural development expenditure (RDE) is progressive (pro-poor) in absolute terms (CC = -0.2.6), with the notable exception of the two productive SEDESOL programs: Opciones Productivas (CC = 0.12) and Crédito a la Palabra (CC = 0.45). 52 The Concentration Coefficient (CC) is a generalized form of the Gini coefficient that shows the concentration of a particular income source x (in this case payment from ARD programs) when recipients are ranked by an index y (in this case pre-transfer income). The CC ranges from -1 when all transfers are received by the poorest households, through 0 when all households receive the same amount of transfer income, to +1 when all transfers are received by the richest households. 61 Figure 4-8: Concentration Coefficients of ARD Expenditures, Income and Land (2006) O portunidades 0 .3 1 5 P iso Firm e 0 .3 1 2 IM SSO portunidades 0 .3 1 5 PET 0 .3 0 9 Total RD 0 .2 6 5 D esay unos D IF 0 .2 3 2 D espensa D IF 0 .1 4 2 Seguro P opular 0 .1 3 7 V iv ienda Rural 0 .0 7 3 Procam po (ENIG H ) 0 .1 1 7 O pciones P roductivas 0 .1 2 1 Total ARD 0 .2 3 8 P retransfer incom e 0 .3 4 9 Credito a la P alabra 0 .4 5 4 P rocam po (ASERCA) 0 .5 0 2 Rainfed land 0 .5 0 3 Total land 0 .5 2 4 Total APE 0 .5 8 4 Irrigated land 0 .6 5 7 Ingreso O bjetiv o 0 .8 0 9 4 5 % 2 5 % 5 % 15% 35% 55% 75% Source: Scott (2008) using ENIGH 2006, ENIGH 2004 (Social Module), ASERCA Beneficiary data bases. Note:Data for 2004 is used for the following programs: Piso Firme, Programa de Empleo Temporal (PET), Deasyunos and Despensas DIF, Vivienda Rural, Opciones Productivas, and Crédito a la Palabra. 73. There is a sharp contrast between the distribution of APE and RDE. Under certain assumptions, 53 the distribution and income incidence of total ARD expenditures in 2006 was estimated. A very different story emerges for APE and RDE. More than half of APE is concentrated in the richest household/producer decile, whereas RDE is progressively distributed, with the extreme rural poor (the poorest 20 percent of rural households) obtaining 33 percent of transfers (Figure 4-9). APE is regressive even relative to the distribution of pre-transfer income, so its incidence actually increases income inequality. 53 The assumptions used are as follows: (i) The RDE transfers whose distribution could not be estimated directly (40 percent) are distributed on average as those that were estimated (60 percent). This probably overestimates the progressivity of RDE, given the weight of Oportunidades in the estimates. (ii) The APE programs whose distribution could not be estimated (37 percent) are assumed to be distributed as total (rain-fed & irrigated) cultivated land, as reported in the ASERCA data bases (except for the energy and hydro-agricultural expenditures which are assumed to be proxied by the distribution of irrigated land). This is probably a lower bound for the repressiveness of APE. (iii) Given the important degree of underreporting of household income in ENIGH when compared to the National Accounts, to obtain a realistic estimate of the incidence of ARD expenditures, HH income was adjusted by the relevant factor (1.87) to make it consistent with the National Accounts. Since it is reasonable to assume that underreporting in Mexico is more significant at the top than at the lower end of the income distribution, both adjusted and unadjusted estimates are estimated. 62 74. Results show a flat ARD distribution for the poorest 70 percent and a sharp increase for the richest 10 percent, mainly through APE. The top 10 percent receive more than half of total APE, but less than 2 percent of total RDE (Figure 4-9). Total APE transfers represent 20.7 percent of the adjusted average income of the richest decile (almost 40 percent unadjusted), but just 7.6 percent of the poorest (14 percent unadjusted) (Figure 4-10). On the other hand, total RDE adds 53 percent (almost 100 percent unadjusted) to the poorest deciles pre-transfer income, but barely adds to the income of the top decile. To summarize, the distribution of public ARD expenditures is flat for the poorest 70 percent, at less than MxP 500 per capita per month, but increases sharply at the tenth decile, where rural households obtain on average more than MxP 3,000 monthly per capita (Figure 4-11). Figure 4-9: Estimated Distribution of APE and RDE (%, 2006) [rural household deciles ordered from left to right by pre-transfer per capita income] 60% 52.3% 50% 40% 30% 20% 17.2% 15.8% 13.2% 13.4% 12.0% 12.1% 9.4% 9.2% 10% 6.9%7.1% 6.6% 5.2% 4.8% 3.4% 4.0% 1.6% 1.6% 2.5% 1.8% 0% 1 2 3 4 5 6 7 8 9 10 household deciles APE RD expenditure Source: Scott (2008) , using ENIGH 2006 63 Figure 4-10: Estimated Incidence of APE and RDE in Rural Household Income (Transfer as % of Pre-transfer income, 2006) [rural household deciles ordered from left to right by pre-transfer per capita income] 60% 52.9% 50% 40% 30.9% 30% 20.5% 20.7% 20% 16.3% 15.2% 12.6% 9.8% 10.2% 9.0% 10% 7.6% 6.0% 7.1% 7.4% 8.1% 6.3% 5.1% 4.9% 3.1% 0.5% 0% 1 2 3 4 5 6 7 8 9 10 household deciles APE RD expenditure Source: Scott (2008) , using ENIGH 2006 Figure 4-11: Estimated Average Monthly Transfer per Capita to Rural Households from APE and RDE (MxP, 2006) [rural household deciles ordered from left to right by pre-transfer per capita income] 3,500 3,000 78 2,500 2,000 MxP 1,500 3,005 1,000 175 500 201 328 255 199 719 339 336 287 291 417 159 210 282 49 55 84 126 0 1 2 3 4 5 6 7 8 9 10 household deciles APE RD expenditure Source: Scott (2008) , using ENIGH 2006 75. APE cancels out approximately half of the redistributive impact of RDE. The accounting effect of APE on the rural Gini coefficient is to increase it by 6.7 64 percent (11.5 with unadjusted income), while RDE decreases it by 14 percent (24.8 percent unadjusted), with a net reduction of 6.5 percent associated with total ARD expenditures (see Table 4-1). In other words, the regressive nature of the APE is so great that it cancels out approximately half of the redistributive impact of RDE on relative inequality, measured through the Gini coefficient. Table 4-1: Redistributive Effects of APE and ARD Expenditures Distribution Transfer Incidence HH Deciles Transfers Pre- Post-transfer income transfer + APE APE RDE income + APE + RDE APE RDE Total & RDE Income: Unadjusted 108,572 76,925 467,957 (Million MxP) 1 1.6% 17.2% 2.9% 2.7% 4.9% 4.4% 14.2% 99.0% 113.2% 2 1.6% 15.8% 4.4% 3.9% 6.0% 5.3% 9.5% 57.9% 67.4% 3 2.5% 13.2% 5.5% 5.0% 6.6% 5.9% 11.2% 38.3% 49.5% 4 3.4% 12.0% 6.5% 6.0% 7.3% 6.7% 13.2% 30.4% 43.7% 5 4.0% 12.1% 7.1% 6.6% 7.8% 7.2% 13.8% 28.4% 42.2% 6 5.2% 9.4% 8.5% 8.0% 8.7% 8.2% 15.1% 18.4% 33.5% 7 6.9% 7.1% 10.0% 9.5% 9.6% 9.2% 16.8% 11.8% 28.6% 8 9.2% 6.6% 11.6% 11.2% 10.9% 10.6% 19.1% 9.2% 28.3% 9 13.4% 4.8% 13.7% 13.8% 12.5% 12.7% 23.6% 5.8% 29.3% 10 52.3% 1.8% 29.7% 33.5% 25.8% 29.7% 38.7% 1.0% 39.7% Total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 23.2% 16.4% 39.6% Gini 0.5839 -0.2652 0.3486 0.3887 0.2620 0.3118 Change in Gini 11.5% -24.8% -10.6% Income: Adjusted 875,291 (Million MP) 1 2.9% 2.7% 4.0% 3.8% 7.6% 52.9% 60.5% 2 4.4% 4.1% 5.3% 5.0% 5.1% 30.9% 36.0% 3 5.5% 5.2% 6.1% 5.8% 6.0% 20.5% 26.5% 4 6.5% 6.2% 7.0% 6.6% 7.1% 16.3% 23.3% 5 7.1% 6.8% 7.5% 7.2% 7.4% 15.2% 22.6% 6 8.5% 8.2% 8.6% 8.3% 8.1% 9.8% 17.9% 7 10.0% 9.7% 9.8% 9.5% 9.0% 6.3% 15.3% 8 11.6% 11.4% 11.2% 11.0% 10.2% 4.9% 15.1% 9 13.7% 13.7% 13.0% 13.1% 12.6% 3.1% 15.7% 10 29.7% 31.9% 27.5% 29.7% 20.7% 0.5% 21.2% Total 12.4% 8.8% 21.2% Gini 0.3486 0.3721 0.2990 0.3259 Change in Gini 6.7% -14.2% -6.5% Source: Scott (2008) Note: RDE = Rural Development Expenditure 65 5. M&E AND INSTITUTIONAL ASPECTS OF ARD PROGRAMS MONITORING AND EVALUATION 76. The Monitoring and Evaluation (M&E) system in Mexico has been undergoing major changes in the legal and institutional framework since the late 1990s. In the 1970s and 1980s, evaluations were implemented, if at all, by internal departments within the programs themselves, with limited transparency or accountability. In 1997, the Progresa/Oportunidades program began its highly acclaimed M&E system using multi-annual panel data sets and rigorous experimental or quasi-experimental methods. 54 In 2000, annual evaluations became mandatory for all targeted programs in the 2000 Federal Budget Law. Similarly, tax/expenditure incidence analysis was made mandatory for all public spending in 2001 by the 2001 Federal Budget and Tax laws. Evaluations are required to be submitted to Congress to feed into the annual federal budgetary process. In 2005, the Consejo Nacional de Evaluación (CONEVAL) was created by the Ley General de Desarrollo Social, with the dual mandate of generating official poverty measures and coordinating the evaluations of all social development programs, including all targeted rural social development programs. 55 In 2006, results based budgeting and a performance evaluation system were made mandatory for all federal spending programs by the Ley Federal de Presupuesto y Responsabilidad Hacendaria. The Logical Framework was the methodology adopted to generate the relevant performance indicators. In 2007, General Guidelines for the Evaluation of Federal Programs (Lineamientos generales para la evaluación de los Programas Federales de la Adminisración Pública General) were issued by CONEVAL, and in May 2008, the results based budgeting and the performance evaluation system were expanded to local (state and municipal) government programs through a constitutional reform. 54 The credibility of the program's positive and internationally recognized results can arguably be credited as a major factor in explaining the political survival of the program to two changes in administration, a change in political party, and even a political regime change (from effectively one-party rule to alternating parties). 55 CONEVAL is a decentralized federal public agency attached to SEDESOL. Its functions include establishing M&E methodologies and performance indicators, disseminating evaluation results, identifying priority intervention areas, setting up criteria for poverty measurement, and giving suggestions to Government with respect to M&E of social programs and poverty measurement. CONEVAL is chaired by SEDESOL and consists of six experts of well-established reputation and a technical secretary. Evaluation work is jointly coordinated by CONEVAL, SHCP and SFP, which integrated the "Evaluation System". 66 77. To organize the evaluation work, Annual Evaluation Programs are established each year indicating which programs will be evaluated and what type of evaluation will be carried out.56 Programs have already been compiled for 2007 and 2008. The Annual Evaluation Programs reports 130 and 202 evaluations to be submitted to Congress in 2007 and 2008 respectively. However, the majority were basic "consistency and results" evaluations, which often suffer from insufficient depth, and only 3 (6) were impact evaluations in 2007 (2008). 78. The rapid development of the institutional M&E framework since the late 1990s, and especially since 2006, is remarkable but the actual implementation of this ambitious framework is facing practical challenges, especially in ARD programs. Annual evaluations have been limited and heterogeneous in quality, depth, objectivity and comparability. This is especially true for programs in rural areas where the beneficiaries are dispersed. These limitations stem from issues on the supply side, i.e. the thin pool of qualified expert consultants who are able to conduct quality evaluations, as well as the demand side, i.e. the limited resources available to CONEVAL, SHCP, Secretaria de Funcion Publica (SFP), and Congress to regulate and process the rapidly growing number of evaluations. Also, although the Logical Framework methodology adopted is in principle a simple and easily replicable method to achieve consistency between instruments and objectives across programs, the consistent and informative application of this method may be difficult to achieve in practice. The Logical Framework method should be used as a minimum common denominator for the M&E system, to be complemented with more substantive evaluation methods. 79. The evaluation of agricultural programs is particularly tricky for several reasons. First, they are usually demand-led, and hence the actual beneficiaries are not known beforehand. Second, they pose complex measuring challenges (of farm related inputs, outputs, prices and assets), which require carrying out farm management type surveys. Third, results vary among farms according to agro- ecological conditions and factors such as weather and water availability, which makes it difficult to select control groups and compare results. Finally, beneficiaries are dispersed and costly to reach. Because of these challenges, specific methodologies, beyond general evaluation principles and methods, are not readily available to assess the impact of these programs. This is an area where methodological research and trial and error experimentation is highly needed. 80. Limited availability of experienced talent in M&E hinders the implementation and oversight of good M&E. There is a shortage of well- trained M&E experts in universities and consulting firms in Mexico to carry out the evaluations, as well as on the side of the contracting institutions (ministries and agencies) to act as counterparts in the evaluations and monitor quality. 56 Various types of program evaluations are considered in the Lineamientos: consistency and results, indicators, processes, impact, and ad hoc aspects. In addition, there are also strategic evaluations related to the policy, institutional or strategic aspects of programs or sets of programs, and new programs must undergo a design evaluation in the first year. 67 81. ARD programs are monitored and evaluated only at the program level and not at the level of PEC to track progress in rural development. Beyond offering a minimum of consistency between evaluations, the new M&E framework tracks results at the program level, and fails to provide a broader perspective considering the links, complementarities and synergies among programs. There is no autonomous and technical evaluation institution for ARD programs looking at the entire ARD portfolio comparable to the role of CONEVAL for social development, though even the latter has yet to institute an evaluation methodology capable of converting the growing tide of individual evaluations into a comprehensive evaluation of social spending. 82. Currently, there is no unified information system and data base for different ARD programs, or even different agricultural support programs to coordinate the distribution of resources at the beneficiary level. On the contrary, each program (even within SAGARPA) has their own database of beneficiaries and it is often not possible to link them. This makes it difficult to analyze aggregate impacts, identify overlaps, or track the total resources received by individual producers from different programs to allow for the establishment of a ceiling on total APE transfers received, or identifying horizontal and vertical inequities in these transfers. More ambitiously, a unified system would allow for the targeting of complementary support packages as a function of the specific socioeconomic characteristics and productive potential of producers. At present, only a handful of programs, notably Oportunidades, have adequate beneficiary data bases. An ambitious project to construct a unified data base of the principal social programs was initiated by SEDESOL in the previous administration.57 An IDB technical assistance project has been launched recently with the ASERCA unit of SAGARPA to build a system of this kind for the agricultural support programs as well as to design an improved Logical Framework for PROCAMPO.58 83. Annual evaluations are costly to implement and unnecessary in most cases. Currently, external evaluations of targeted programs are carried out annually. This is expensive and in most cases unnecessary because of few program changes from one year to the next. As a consequence, quality suffers, "evaluation fatigue" sets in, and there is a tendency of program managers not to take evaluation results sufficiently into account in their decision-making. It would be preferable to have less evaluations of better quality and more relevance. Also, rather than covering all program aspects, evaluations could cover in depth specific aspects in separate years. The new CONEVAL evaluation framework seems to allow making these adjustments. 57 This has been hampered by the failure to build a functional identification number system for the Mexican population. The Clave Única de Registro Poblacional was created for this purpose, but currently contains multiple registers. 58 The "Improving systems and operational procedures of PROCAMPO" program was approved in March 2008. It is a technical assistance project with US$100,000 financed by the IDB and US$25,000 financed as counterpart from SAGARPA. 68 84. The weakest and most critical link in the improved M&E framework is the insufficient feedback of the evaluation results into managerial and budgetary decision-making. Even the most sophisticated and functional M&E system becomes irrelevant for policy making if there are no effective mechanisms to translate its results into actual managerial and budgetary decisions, from policy reforms to changes in the behavior of final service providers in response to accountability and incentive mechanisms. To date, there is little evidence that the annual evaluations of ARD programs or the overall distributive analysis of public spending/taxes have been used in budgetary decisions by the executive or legislative branches of government, thus creating a space for political interferences in the determination of the public spending program. A possible way to provide feedback is for evaluation exercises to be followed by the setting up of an action plan for program managers to address the issues identified in the evaluations. The plan would be agreed between evaluators, program managers and representatives of the evaluation system. Compliance with the actions agreed upon in the plan would need to be monitored. INSTITUTIONAL ISSUES 85. In this section some selected institutional issues are examined. They succinctly cover two broad topics: (i) the political economy circumstances in which ARD policy-making is conducted, and (ii) the challenges found in the implementation of ARD programs.59 86. A crucial political-economy issue is the pressure of rent-seekers and interest groups to influence programs and their design. This pressure has been behind the creation or extension of some programs in Mexico as well as in other countries, including high income countries, and resisting it may be difficult. One reason is that ARD policies are regarded by some sectors as an arena of resource dispensation rather than as economic instruments to improve distribution and growth. 60 Congruent with this is a view of ARD programs as a cost of maintaining social tranquility in the countryside, to be adjusted according to social negotiation or voting pressure. However, the political and public administration system has increasingly employed a much more technical or rational view in the last decades, and hence views of the above type are rapidly receding. The present moment seems particularly appropriate for accelerating this transition. The formulation of a national ARD strategy and the creation of an ARD expenditure planning system advocated in Chapter 6 could be the triggers of a stronger position vis-à-vis this type of pressures. The strengthening of CONEVAL's M&E function would also favor this. 59 More detailed analysis and factual information is contained in World Bank 2005b on the first topic and World Bank 2006a on the second topic. 60 This trend is not unique to Mexico. It is often observed in other countries as well, developing and developed alike, where agricultural programs are often exempt from the usual scrutiny and evaluation accorded to publicly funded programs. 69 87. A related issue is the often-strained relationship between the executive and legislative branches of government in the formulation of ARD policies and budgetary allocations. More interaction between the Executive and the (strong) rural lobby in Congress, particularly, but not only, during the budget formulation process would improve the chances for a more rational ARD policy framework. 88. Another political economy issue is the relation between the federal and state governments in ARD policies. With a few exceptions (mainly Alianza and FAIS-FISM (Fondo para la Infraestructura Social Municipal), ARD programs and expenses are fully centralized. Program design, funding and operational rules are usually handled at the federal level.61 Under these circumstances, since state governments do not have control of the rural agenda in their states, they do not internalize the political and economic costs and benefits of ARD actions, successes and failures. There is an ambiguous state of affairs in which neither the federal government nor state governments feel truly accountable for ARD outcomes. State governments are not necessarily unhappy with this, because, if need be, it allows them to blame the federal level for program shortcomings and avoid hard, possibly unpopular, decisions in the allocation of scarce funds. This pushes the situation towards low-level policy equilibrium, particularly since state governments would have clear advantages in designing and carrying out many ARD actions. Certain courses of action to counter this trend are proposed in Chapter 6. 89. Implementation matters, and is particularly complex in production-oriented programs. How are programs implemented is almost as important as how they are designed. Implementation challenges may jeopardize, even reverse, the achievement of objectives. 62 These challenges are particularly acute in production-oriented programs compared to infrastructure or cash transfer programs. The reasons are that for good implementation production-oriented programs require a medium- to long-term perspective, a particular type of synergy between ground level program operators and the responsible agency, high quality participation of beneficiaries, and a large recurrent budget. The bottom line is that these programs deal essentially with private goods, and public-private convergence is hence of the essence.63 The above conditions are difficult to meet, and appropriate implementation is therefore a challenge. Several aspects of the operational challenges are highlighted below. 90. Macro type political and administrative circumstances may place constraints on implementation. One such circumstance is the 3-year, no reelection electoral system of municipal authorities, which hinders their involvement in medium- to long-term endeavors like ARD. Another circumstance is the annually based 61 See World Bank (2006a) for an analysis of the decentralization of rural programs in Mexico. 62 For an analysis of implementation challenges of rural development programs in Mexico see World Bank, 2005b, Rural Poverty Study, Chapter 6. For the case of the rural development program of Alianza, see World Bank, 2006a, Rural Decentralization Study, Chapter 3. 63 An illuminating analysis of this matter, based on the case of public programs in the State of Ceará, in Brazil, is that of Judith Tendler (1997). 70 budgetary system existing in Mexico, which does not favor the continuity of programs and views required by ARD, and does not help giving security to beneficiaries regarding the stability of program benefits. 91. Execution of the budget is limited to a short time window and relies on deferred payment modalities. Overall the execution rate of the PEC is quite high. The execution rate for the latest year available, 2007, shows an execution rate of 97 percent for the entire PEC program (Table 5-1). However, the existence of administrative and budgetary norms make it, in practice, difficult to disburse program funds and typically limit the disbursement period to several months in the year, thus placing big pressure on those executing the program64. An examination of the quarterly execution rates of the PEC budget, at glance, present a smoothed disbursement pattern over the quarters (Table 5-2). Table 5-1: 2007 PEC Budget by Ministry PEC 2007 (millions of pesos) Budget Branch (ramo) Approved Modified Executed % Executed TOTAL 176,794.4 170,122.6 165,354.4 97% SEGOB 300.0 300.0 300.0 100% SRE 71.4 82.8 82.8 100% SHCP 8,291.9 8,100.2 7,534.5 93% SAGARPA 58,536.9 57,677.4 56,519.4 98% SCT 2,857.8 2,857.7 1,707.5 60% SE 838.5 837.3 837.2 100% SEP 23,686.8 24,592.0 23,220.0 94% SSA 15,940.5 10,752.0 10,752.0 100% STPS 66.9 66.6 66.6 100% SER 4,779.8 4,701.4 4,701.4 100% SEMARNAT 14,289.8 13,026.1 12,551.1 96% AP-SS 5,512.9 5,512.9 5,512.9 100% SEDESOL 16,525.1 16,524.5 16,517.3 100% SECTUR 76.5 76.5 76.5 100% PSE 250.0 250.0 250.0 100% TSA 694.0 689.6 649.6 94% AP-FED 24,075.6 24,075.6 24,075.6 100% Source: SHCP Table 5-2: 2007 PEC Quarterly Execution Rate (Million MxP) Quarter Allocated Executed (cumulative, incl. CLCs) % Execution (cumulative) 1Q 42,530.70 29,196.80 17.2% 2Q 42,530.70 85,341.60 50.2% 3Q 42,530.70 126,309.20 74.2% 4Q 42,530.70 165,354.40 97.2% Total 170,122.80 165,354.40 97.2% Source: Centro de Estudios para el Desarrollo Rural Sustentable y la Soberanía Alimentaria 64 Many APE programs are very sensitive to the timeliness with which the resources reach farmers, especially in areas where the rainy season is short. Operational and budgetary norms often stand in the way of timelines, and result in the late arrival of subsidies. 71 92. Recurrent funds for program operation are scarce. Vigilance from SHCP that program money is well spent, as much as possible going to final beneficiaries, is to be praised, but it must be understood that shortage of operational funds may jeopardize programs. It is important for budgeting authorities to realize that production-oriented programs are particularly costly in recurrent costs if high quality implementation is sought. Also, the level of operating cost is expected to differ depending on project design, e.g. PROCAMPO requires very little operational funds once the beneficiary roster has been compiled, while others, like Alianza's PDR requires much more funding for effective implementation. 93. Organizational systems and cultures often lead to poor implementation. Due to the segmentation of government agencies, rural programs often have different breakdowns of geographical regions, different definitions, norms and procedures for similar things, different timing and disbursement methods, and create their own separate counterpart organizations. This makes program coordination extremely difficult. Another key issue is how to overcome the culture of short- term achievement often present in ARD program activities characterized by focusing on quick results with insufficient strategic focus. The pressure to disburse program funds is part of this. Pursuing quick results hinders inter- agency and agency-client cooperation and beneficiary selection, and leads to low expenditure quality. 94. There is insufficient client orientation and empowerment of beneficiaries. Three main issues may be highlighted here: (1) poor program dissemination, which is essential in demand-driven programs; (2) weak direct accountability of program operators to clients/beneficiaries and weak "client satisfaction" incentive criteria; and (3) in some cases, an opportunistic and rent-seeking behavior of program operators and service providers. 95. There are few incentives for program operators to perform well. Program operators dealing with field activities and the interface with clients are generally poorly paid and their knowledge and experience little valued. This refers not only to the staff of the implementing agencies, but also to contracted private technical service providers. The poor economic compensation and weak incentives to boost morale for bottom level program operators is generally inconsistent with the importance of their function. Some of the issues are: no link of payments to performance and client satisfaction; little valuation of the operators' function, with scarce consultation with them on program matters; insufficient training and dissemination of best practices; no systematic performance evaluations; and an insufficient networking system and a client-orientation culture. 72 6. OPTIONS FOR MOVING FORWARD 96. Mexico's large ARD program represents a very significant fiscal effort by the GOM on behalf of the rural population. Average public expenditure per capita is now similar in the urban and rural sectors, unlike that in many Latin American countries, where there is still an "urban bias" in the allocation of public expenditure. The present ARD program represents the outcome of the far- reaching reform efforts which began in the late 1980s to modernize the sector and introduce a more efficient, equitable and less distortionary policy environment. 97. Mexico's ARD policy reforms included many novel features signaling a clear departure from past policies. They resulted in some international best practices such as the success of the Oportunidades program in rural poverty reduction, the decoupled design of PROCAMPO, and the strong global leadership of the GOM in mainstreaming climate change in ARD issues. This chapter offers options and alternatives for the GOM to move forward on further improving its ARD public expenditure program. 98. The basic diagnostics in previous sections identify several challenges in the current structure of ARD expenditure: Low cost-effectiveness--a high level of spending in rural areas, with low impact on growth, climate change adaptation, and poverty reduction. Imprecise ARD strategic objectives and rationale. Distortion in some important programs--causing the prices paid or received by private parties to differ from their values to society, discouraging producers and consumers from making economically efficient decisions and in some cases resulting in negative environmental externalities. Poor coordination among the many expenditure programs. Regressive overall spending, especially in agricultural support programs. Limited built-in monitoring and evaluation. Implementation constraints. KEY FEATURES OF THE ARD PROGRAM IN MEXICO 99. Resources for ARD in Mexico are large. The amount of resources spent by Mexico in ARD is large, bigger than in other countries in Latin American and bigger also than in many OECD countries, in relation to the sizes of the agricultural sector, the rural population, and the public expenditure bill (see chapter 2). Yet the results have been less than expected (see chapter 3). So the 73 challenge for Mexico is not the size of ARD expenditures but their allocation and quality--the thrust of Mexico's expenditure reform. 100. Some ARD programs have contradictory goals. There seems to be little coherence among the objectives of some support programs to farmers. SAGARPA and CONAGUA programs improve efficiency in irrigation water use, but tarifa 9 militates in the opposite direction. Similarly, Alianza, Apoyos a la Competitividad de las Ramas Productivas, and PROCAMPO Capitaliza aim mostly at production intensification, competitiveness, and employment generation. This aim is contradicted by Ingreso Objetivo, which mostly supports crops that generate little employment and have no comparative advantage. Similarly, input support programs, like the electricity and diesel subsidies promote production patterns based on comparative advantage. Mexico should address these inconsistencies in policy messages and incentives. 101. Public goods are undersupplied. The importance of public goods to promote agricultural development is well established (see Box 2-2). Public goods ought to be the main thrust of a competitively oriented APE program. Yet, considering all fiscal costs examined in this study (budgetary and nonbudgetary), public goods are only 24 percent of the cost of supports directly related to agricultural development--and 43 percent of those related to total ARD. Rebalancing ARD expenditures in favor of public goods would be one of the crucial aspects of agricultural policy reform. 102. Some ARD programs have the perverse effect of encouraging activities that generate negative externalities. Ideally, the programs should generate positive externalities. The most obvious challenge is the incentive to deplete aquifers created by the tarifa 9 subsidy to groundwater pumping. Others include incentives to overuse agrichemicals, resulting in contamination of waterflows, and to cultivate marginal lands, resulting in erosion and increased flood risk downstream. Programs like tarifa 9 that encourage overuse of water also result in maladaptation to future conditions, which are likely to be drier since Mexico is expected to experience a decline in water runoffs of 10­20 percent nationally and up to 40 percent along the Gulf Coast wetlands.65 Some policies, particularly subsidies to grain production, are also known to contribute to deforestation.66 103. Redirecting agricultural expenditures to encourage activities that generate positive externalities would produce substantial environmental gains. Several windows of the ProÁrbol Program administered by CONAFOR already seek to do this; opportunities should be sought to recast other programs along similar lines. Such measures could also generate additional financing--for example, from local water users and from national and global buyers of carbon emission reductions. 65 The decline is in the difference between precipitation and evapotranspiration (World Bank 2008b). 66 Klepeis and Vance (2003), in a study of Southeastern Mexico, suggest that PROCAMPO's requirement to keep land in production is linked to deforestation and deterioration of soil fertility. 74 POLICY OPTIONS FOR REFORM 104. This section discusses some options and suggestions to reform ARD expenditures in the light of the analysis of previous chapters. ARD expenditure as a whole is examined, and options for a new overall structure are presented. The present moment is particularly appropriate for these changes because of the opportunities created by the medium- to long-term scenario of high agricultural prices. The buoyant state of markets should make the transition smooth and the changes more palatable to those who would see their subsidies diminished. 105. To make well informed decisions, the GOM will need to weigh the tradeoffs of policy options based on considerations such as cost, sequential importance, technical difficulty, risks, and impact. While some options promise greater benefits in the longer run, others could be targeted for quick action. The chapter summarizes policy options in five areas: 6) Improving the ARD planning system. 7) Rationalizing the overall farmer support system. 8) Improving the support system for small producers through gradual decentralization. 9) Improving the M&E and institutional aspects of ARD programs. 10) Increasing the positive environmental externalities of ARD programs and supporting the national climate change strategy. 1) Improving the ARD Planning System 106. A more effective planning system for ARD spending could be put in place. The guidelines for allocating budgetary resources and the authority and planning capacity of the Comisión Interministerial para el Desrrollo Rural Sustentable (CIDRS) would have to be strengthened for PEC to be effective . A recent OECD review of rural development issues in Mexico (OECD 2007c) highlights the challenge faced by Mexico where SAGARPA, a sectoral ministry without sufficient authority over peer sectoral ministries is mandated to plan the ARD expenditures as chair of CIDRS. Box 6-1 summarizes the OECD analysis on coordination of rural policies of its member countries (OECD 2006a). The OECD proposal to move CIDRS from SAGARPA to the President's office would better reflect the multisectoral character of rural development, and strengthen the Commission's planning capacity. Another option is to place CIDRS under SHCP because SHCP is primus inter pares as a ministry, has a multisectoral mandate and view, and is responsible for looking into the effectiveness and quality of public expenditure and preparing the federal budget. In addition to strengthening CIDRS, a strong nucleus of agricultural policy analysts, staffed with experts of 75 recognized academic excellence could be formed to operate autonomously under the SAGARPA. Box 6-1: OECD Analysis on Inter-Institutional Horizontal Co-ordination OECD sees interinstitutional coordination as an aspect of meta-governance--or "the governance of government and governance." In this framework horizontal coordination attempts to overcome sectoral approaches in favor of an integrated policy approach to rural development. Coordination is needed to "encourage the various institutional and managerial systems which formulate and implement rural policy to work together." To ensure consistency--that is, "that individual policies are not contradictory and that they converge in a coherent strategy"--several horizontal co-ordination options are considered following OECD member-country experience: "a special unit reporting directly to a head of government or parliament [France]; an integrated ministry to address several issues of importance to rural regions [UK, Germany, Japan]; `policy proofing' [UK, Canada]; and inter-ministerial co-ordination via working groups and formal contracts [Mexico, Italy]" In Mexico OECD highlights the innovations from the LDRS and CIDRS, but acknowledges that "SAGARPA has been much on its own leading the promotion and implementation of the LDRS [and that] the fact that the CIDRS is chaired by one sector limits the multi-sectoral objective of the law." In this last respect, "experience from OECD countries indicates that a horizontal commission which is chaired by one sector (in this case, agriculture) may be limited in pursuing multi-sectoral objectives and hinders the full involvement of other ministries in a national rural strategy". The alternatives proposed to strengthen the CIDRS are: "Assigning a meta-ministerial leadership to the CIDRS (...), which could be filled by the Chief of the Executive". "Rotated leadership of CIDRS (...) among different ministries. In this way the works of the commission are not seen as monopolized by one institution" "The creation of an ad hoc independent institution in charge of rural policy with multi-sectoral perspective, with budget and normative arrangements to enforce collaboration from the different ministries". "Strengthening the legal attributions of CIDRS (...), with stronger budgetary allowances rather than the formality of presenting a `rural budget' to congress". Source: OECD, 2006a, pp. 110­114, and OECD 2007c, pp. 118­122 107. The planning system for ARD spending should be based on a comprehensive national ARD strategy. Without a comprehensive ARD strategy and planning system for ARD expenditures it is difficult to achieve coherence among ARD programs and their objectives. A comprehensive ARD strategy with consistent, explicit, and selective objectives as well as quantifiable results and performance indicators would be an important step toward better ARD policy. The strategy could serve as a framework for individual programs, to be assessed by their contribution to its objectives. M&E of individual programs could also be carried out. Along with the strategy, an ARD budgetary planning system would be required to ensure consistency between the strategy and budgetary allocations. As seen, PEC has led to substantial progress in this direction, but so far it provides only the framework--not the content--of such a system. 76 2) Rationalizing the Overall System for Farmer Support 108. The federal public expenditure system in productive programs should be reoriented in three ways. Gradually discontinuing MPS to inputs67 and outputs, as well as compensatory cash transfers (see Technical Note 1 in the Appendix of this chapter which describes New Zealand's successful agriculture policy reform experience in the Appendix to this chapter). Substantially increasing allocation to irrigation improvement and management; drainage and transport infrastructure; other rural infrastructure (such as that provided by FAIS); research, extension, and training; sanitary and phyto-sanitary services; market intelligence (including collecting weather information through weather stations and agricultural statistics); emergency programs; and other public goods.68 Increasing support to small producers. 109. Non-targeted production-oriented programs benefiting mostly commercial farmers should be separated from productive and natural resource management (NRM) programs targeting small producers, vulnerable groups, and the poor which should be decentralized to the states. Resources for non- targeted production-oriented programs could be maintained at the federal level, but funds now spent by different ministries for targeted programs could be pooled. The pooled funds could be distributed to state governments as block grants according to an objective formula to allow state governments to create their own programs that promote small producers. In addition, given the current imbalance in favor of non-targeted programs, some resources could be moved from non-targeted programs to the pooled fund. These decentralized funds could also support programs that assist small producers in adapting to climate change and increasing positive externalities in NRM. 110. Federal non-targeted programs that support on-farm and value chain investments should be rationalized. These programs include PROGAN, Apoyos a la Competitividad de las Ramas Productivas/Adquisición de Activos Productivos, and the non-targeted programs of Alianza. One option is to merge these types of 67 Given the likely difficulty of immediately discontinuing tarifa 9, a major input subsidy program that subsidizes electricity for farmers to pump groundwater, Asad and Dinar (2006) present some options for reforming the program: (i) decoupling, so that each farmer receives the average subsidy; (ii) allocating the subsidy based on historical consumption; (iii) assigning subsidies only to water concession holders, thereby stimulating a more efficient and legal use of water and electricity; and (iv) a combination of one of the preceding options with a payment per hectare approach to further target the subsidy. Each of these options has political economy consequences that would need to be considered in the context of the policy dialogue, but none of the options would be as politically charged as eliminating the subsidy altogether. 68 Certain functions within these activities should be carried out by the private sector due to the private goods nature of the activity, e.g. implementing appropriate management practices, financing of the certification process, developing traceable systems etc. 77 program into a new single non-targeted support program for on-farm and agriculture value chains investments. Another is to maintain different programs but redesign them and perhaps eliminating some to avoid duplications, sharpen their objectives, and make them more effective and results oriented. 111. Non-targeted programs, which benefit mostly commercial farmers, are better administered at the federal level. This is because these producers tend to be more specialized, better linked to the value chains, and less in need of a close rapport with agency staff. A commodity or value chain approach rather than a territorial one could be used for these programs, so it makes sense to decentralize the mandate and resources for supporting small producers while managing non- targeted programs at the central level. 112. The logic of deficiency payments needs to be reassessed. Deficiency payments protect farmers' incomes from market events by providing a subsidy equal to the difference between the market price and a reference price calculated to ensure a "fair" income to farmers. Isolating farmers from price movements (particularly price falls) interferes with market adjustment. Depending on the crop supported, deficiency payments can encourage overuse of water resources, as in Mexico, where water-intensive basic grain production is supported. Deficiency payments may be justified on a strictly temporary basis under emergency conditions but not as a permanent support system. One such deficiency payment program is Mexico's Ingreso Objetivo which covers 10 basic crops (mainly grains) in certain states and is used as a permanent system of income support to beneficiary farmers. 69 The program is expensive and inequitable and distorts production decisions. More efficient market-based instruments based on price risk management tools, insurance, and other public goods could also protect farm incomes while promoting competitiveness, employment, and growth in a less distortionary manner than deficiency payments. 113. The logic of decoupled compensatory transfers also needs to be reassessed. Lump-sum compensatory transfers have been much celebrated because, being decoupled from production and prices, they are a non-distortionary payment to farmers. They have been widely used in the European Union, United States, and other countries to cushion the transition to market liberalization and the lifting of price supports. Originally designed for a transition period of 15 years (1994­ 2008), Mexico's PROCAMPO is this type of a large compensatory transfer program. Having reached the end of the transition period and having cushioned the NAFTA impact, the economic logic of this program should be analyzed, given how expensive and inequitable it is. The GOM can reassess whether to keep paying farmers more or less indefinitely because they planted corn (or other eligible crop) 18­20 years ago, especially with the recent high agricultural prices. 69 This program is akin to the U.S. Marketing Assistance Loan program, a major countercyclical program that allows farmers of designated crops to receive a loan from the government using production as collateral. The program effectively guarantees a minimum price for the crops since producers can forfeit the crop to the government when the loan matures and keep the loan principal as payment, if market prices are lower than the guaranteed price. 78 114. The role of compensatory transfer programs on easing rural liquidity constraints should be clarified. Rural credit in Mexico is very limited, seriously constraining investment capacity in rural areas.70 Because of this, public subsidies can help ease liquidity restrictions for rural producers. But it would be better to tackle the underlying challenge of increasing rural finance (Box 6-2)--for example, by adjusting the regulatory framework to better reflect the circumstances of rural financial markets,71 by picking up part of the risk of rural credit operations, by subsidizing some fixed costs of providing financial services to small dispersed producers, by supporting microfinance institutions with training and technology, or by expanding the capacity and improving the focus of government rural financial institutions such as Fideicomisos Instituidos en Relación con la Agricultura (FIRA), Fondo de Capitalización e Inversión del Sector Rural and Financiera Rural. Furthermore, promoting price risk management tools and insurance can increase farmers' access to credit because financial institutions regard these tools as guarantees in approving credit. Box 6-2: Challenges Facing the Rural Finance Sector in Mexico Rural finance in Mexico is characterized by inadequate outreach and limited access to finance. Nearly 70 percent of municipalities do not have a single bank branch, and close to 80 percent of the rural population does not have access to financial services Financial services in rural areas are provided mostly by numerous small, often weak institutions. There are three main challenges facing the sector at the retail level: Consolidating local institutions and enhancing depositor confidence by ensuring that institutions are regulated and supervised and by establishing a deposit protection fund offering a reasonable level of security to savers. Deepening penetration and outreach of the sector by strengthening the branch network of existing institutions and by deploying innovative non-infrastructure-based approaches such as correspondent/agent banking and mobile banking. Adapting lending processes to the characteristics of low-income rural clients to mitigate self-selection out of the formal banking system by potential clients and introducing financial products that respond to needs of the local population (small savings, medium- to long-term maturity loans, national and international remittances, among others). There are several new types of agencies ("para-financieras," SOFOLES, SOFOMES) that are becoming important sources of rural finance using different means. It would be useful to evaluate their effectiveness. The national and policy levels also face a range of issues, notably more schemes supported by multiple institutions, which permits arbitrage among recipient banks/institutions and unhealthy competition among public institutions (in lending and in channeling subsidies), and a relatively large subsidy in interest rate and transaction costs. The main challenges at the policy level are initiating action to consolidate and integrate the various schemes and programs under one or two agencies, carrying out a stringent evaluation of the different forms of subsidy and their outcomes to improve efficiency of public funds, and reviewing 70 Rural finance is an extremely important policy issue in Mexico, but it is not addressed in this study partly because of its technical nature and partly because, although there are some rural finance programs in PEC, this being mainly a private area, most associated issues are not visible in a public expenditure analysis. 71 Some examples of proposed adjustment to the regulatory framework include amending legislation of the minimum deposit requirement for niche financial companies such as housing mortgage companies and consumer lending companies to enable them to accept deposits, which is a more demanded service in rural areas than credit; reforming the structure of Financiera Rural to separate its role as a second-tier bank and first-tier bank to alleviate unfair competition in the first-tier banking market; and defining norms to standardize the operations of auxiliary financial service firms in areas such as credit registration forms, credit reporting to government, and accounting reporting forms (Buchenau and del Angel 2007). 79 the roles of the nodal agencies in the changing sector institutional capacity (for example, FIRA's ability to continue it subsidy program is expected to diminish significantly over the next five to six years). 115. Rural non-farm development has large potential for complementing agriculture policies in fostering rural development. As shown in chapter 2, rural nonfarm activities are extremely important for income generation and employment in rural areas, but they receive only a small part of ARD expenditures. Reaching this sector with promotional activities is a big challenge, but so is the potential payoff. Difficulties arise from the multi-sector character of rural nonfarm activities and the absence of organizations whose members are engaged in rural nonfarm economic activities, 72 which gives the sector little visibility and common voice. The government's sectoral organization in economic development also makes it difficult to formulate policies that support the sector. This is particularly true because some important measures ­ including safety nets in the short term and increasing human capital in the longer term ­ are the responsibilities of different governmental agencies. 116. The suggestions described above for rationalizing the overall system for farmer support imply, at the individual program level, the gradual discontinuation of PROCAMPO and Ingreso Objetivo as well as the transformation of Alianza. The rationale for these options and discussion of alternatives follows. 117. PROCAMPO. The future of PROCAMPO has elicited many ideas. So far, notwithstanding the end of the NAFTA transition period, the program has been maintained without major modifications. But there are airs of change, and proposals are being voiced. Some attempt to make the program less regressive by capping payments (making payment per hectare a decreasing function of size or limiting eligibility to farmers under a certain size, say 5, 10, or 20 hectares). Another proposal is to open the program to all farmers and link it to present farm size rather than to the historical acreage of eligible crops. Thus, all farmers, by virtue of being such, would receive a subsidy linked somehow, more or less regressively, to farm size, with or without a limit on that size or a cap on payments. Bolder proposals are to discontinue PROCAMPO and use its resources for something else to benefit farmers or the rural population (such as expanding Oportunidades, carrying out new environmental programs, or improving rural education). The most detailed proposal divides PROCAMPO into two parts: an unconditional cash transfer and a conditional cash transfer (see Box 6-3; Winters and Davis 2007). Box 6-3: Proposal by Winters and Davis (2007) on PROCAMPO Reform The proposal by Winters and Davis (2007) divides PROCAMPO into an unconditional cash transfer and a conditional cash transfer. The unconditional transfer would be paid to all farmers, regardless of crop, of up to 5 hectares,1 as a fixed, lump-sum. Eligible land is defined as "the average total land [under] production 72 This includes small traders, artisans, small entrepreneurs in different industries (dressmaking, shoemaking, construction materials, and the like), providers of personal and technical services (hairdressers, mechanics, electricians, and the like), and others. 80 over the previous three agricultural seasons."2 Conditional transfers would be paid to all farmers, irrespective of crop, of up to 10 hectares, proportional to the amount of land, and subject to certain conditions on "capacity building and/or the adoption of certain agricultural practices of resource management."3 Participation in the conditional transfer would be optional. Farmers of less than 5 hectares could receive both transfers. Conditionality would be "region specific and determined by regional SAGARPA offices." The total payment to a household from the two transfers could be capped. For households also participating in Oportunidades the payment from the PROCAMPO transfers plus the payment from Oportunidades could be capped. Observations. This proposal improves on PROCAMPO but has some shortcomings. With regard to the unconditional transfer, what is the sense of making an unconditional lump-sum transfer to small farmers when Oportunidades already exists? Oportunidades is well targeted and covers the entire rural area, so a farming household that is not a beneficiary is not likely to qualify as poor.4 Why should it then receive an unconditional cash transfer from government? What makes this household more deserving than other nonpoor households? If the household already participates in Oportunidades, why should it receive a separate payment from PROCAMPO that other Oportunidades beneficiaries do not receive? There is little economic sense in this. The conditional transfer raises two questions. First, exactly what conditionality would be applied and how would it be monitored? The implementational difficulties of production-oriented programs suggests many monitoring challenges. Second, is a size-related transfer conditional on certain recipient behavior better than alternative ways to promote farm competitiveness for the size range of farms considered. Among the alternative forms is support (through co-financing) of on-farm investments like those carried out by Alianza. Much duplication and overlapping would be likely between this transfer and the Alianza PDR program, which is targeted to the same type of farmers. 1 The authors indicate that the cutoffs of 5 and 10 hectares chosen for small and medium-size farmers are only suggested definitions and that careful consideration should be given to other cutoffs. The authors also mention that the beneficiaries of the first component could be selected by their poverty status, using a method similar to that of Oportunidades. 2 No mention is made of livestock farms, and no difference is stated between rainfed and irrigated land. 3 The authors also say: "The transfer could be conditional on adopting certain agricultural practices or attending public lectures on productive activities and natural resource management and from these lectures presenting a management plan for improving on-farm production practices." 4 Oportunidades, like any other targeted program, suffers from exclusion errors. But Oportunidades is acknowledged to be well targeted and well implemented, and hence this error should not be large. In any case, it would make little sense to keep PROCAMPO as a default payment system for those missed by Oportunidades. Source: Winters and Davis (2007). 118. Considerations for modifying or discontinuing PROCAMPO in its present form. Although the rationale for discontinuing PROCAMPO may be clear economically, especially given the existence of other programs that cater to small farmers, modifications to the program may be more feasible for the short term. In designing a new PROCAMPO, important considerations include: Ensuring that a modified PROCAMPO does not cause major negative environmental externalities, or, better yet, that it promotes environmentally friendly activities. 81 Enhancing, or at least maintaining, the level of decoupling between production and the subsidies to avoid distorting production decisions.73 Improving equity--for example, by capping payments or reducing payment per hectare as the number of hectares increases. Building in incentives to induce farmers to switch to high value crops, and Announcing a final date to discontinue the program and possibly a timetable for reducing payments until full discontinuation. 119. Possible alternative uses of PROCAMPO funds. Funds from the existing PROCAMPO could be allocated to a completely new program with different eligibility or operating rules or could be combined with an existing program that would keep PROCAMPO's identity. Worthwhile alternatives include an augmented Oportunidades program ("CAMPOrtunidades"), new environmental programs, and improved rural education programs. Another option is to expand the noncontributory rural pension scheme, which pays individuals age 70 and older who live in locations of up to 20,000 inhabitants MxP 500 a month. Given this coverage and benefit amount and 2005 Conteo Nacional de Población population data, the program would cost about MxP 10.2 billion. Expanding to cover the population age 65 and older would increase the cost to MxP 15.2 billion, and to cover the population age 60 and older (as in Brazil) would cost MxP 21.7 billion. Since the 2007 programmed cost of PROCAMPO was MxP 14.9 billion, the funds from the discontinued PROCAMPO would more than allow the pension program to cover rural residents age 60 and older. 120. Ingreso Objetivo. Ingreso Objetivo has much smaller coverage than PROCAMPO and is generally regarded as a distortionary program. Most experts have proposed discontinuing rather than modifying it (Box 6-4; Sumner and Balagtas 2007). Another option being discussed by GOM is to use price bands rather than a single reference price, but details of this proposal are still being developed. Box 6-4: Proposal by Sumner and Balagtas (2007) on Ingreso Objetivo Reform Sumner and Balagtas (2007) make two recommendations for Ingreso Objetivo reform. The preferred one is to discontinue the program for five reasons: The subsidy is inequitable, concentrating on better-off farmers and states. Expected high cereal prices in the coming years mean that payments will be very small or nil, making this a unique opportunity to eliminate the program. It would be better to use countervailing duties to compensate producers for the reduction in U.S. export prices due to U.S. domestic support policies. Resources used in the incremental production of program crops due to the subsidy would be allocated to other uses, increasing efficiency. Eliminating the program would reduce production from participants, thus raising domestic prices and favoring nonparticipants. 73 PROCAMPO is currently partially decoupled because subsidy payments are based on historic production patterns. But a condition of receiving payment is that the land in question is still in active productive use (including forestry or certain ecological use). Thus, for example, land that is fallow would not be considered as under eligible use. This has raised concerns over its effect on deforestation and soil degradation. 82 They also recommend improvements to the program if political economy reasons prevent discontinuation. The first is to reduce target prices so that "they reflect the `safety net' feature of the program," setting target prices "so that payments are made only when market prices are truly low by historical standards recognizing the long-run trend toward lower commodity prices in inflation adjusted terms." Another is to decouple payments, converting them into direct transfers to "all producers with a history of marketing the program crop, with no requirement that they maintain production in that or another crop." Observations. Reducing target prices seems a reasonable second-best option, but a program of decoupled direct transfers to farmers with a historical record of producing the program crops already exists: PROCAMPO. If amended this way, Ingreso Objetivo would duplicate PROCAMPO while being more poorly targeted. Another option is to pay the price subsidy not on the full amount produced but on a smaller, intra-marginal amount, so that the Ingreso Objetivo payment does not affect producers' decisions. This would improve the allocative distortions from the subsidy but not its distributional bias. Related to this, Rosenzweig (2008) proposes to discount PROCAMPO transfers from the Ingreso Objetivo subsidy. Source: Sumner and Balagtas (2007) 121. Alianza. Alianza para el Campo, Mexico's flagship agriculture investment support program, is the largest demand-led program for co-financing private on- farm investments. But Alianza is more than that: it is a financing umbrella for a wide variety of SAGARPA activities, from fisheries to sanitary and phyto- sanitary services, farmer organization, support to rural development councils, training, research, information systems, and a host of other expenses. Many actions financed under Alianza are regular SAGARPA activities--that is, part the regular work of any ministry of agriculture. It is unclear why these activities are financed under a specific, annually renewed program, subject to yearly-approved operation rules. 122. Reorienting Alianza. Three measures are proposed to realign the program: Separating all the subprograms, components, and activities that naturally belongs within the regular program of work of SAGARPA and budgeting them under SAGARPA's regular expenses. Merging funds from Alianza PDR with those of other production-oriented targeted programs to be decentralized. Either (i) merging untargeted investment support Alianza subprograms with other untargeted investment support programs, such as PROGAN and Apoyo a las Cadenas Productivas, to create one single large program to co-finance on-farm and value chain investments, or (ii) maintaining different production-oriented untargeted programs and redefining them to structure a comprehensive system that provides general investment support to on-farm and agriculture value chain investments. The policy instruments available for commercial agricultural growth, including Alianza, must be rationalized and made more effective. Rosenzweig (2008) points out the weak focus on productive activities in the current APE and emphasizes the need for new instruments for a more effective agricultural policy. 83 Restructuring the investment support system for commercial agriculture would be a major part of this effort.74 3) Improving the Support System for Small Producers through gradual decentralization 123. Small producers face big and distinct challenges. Small producers' restrictions and potentials differ from those of medium-size and large commercial farmers. Small producers needs require adequate consideration in ARD expenses to make them more competitive, especially given the highly regressive nature of the agricultural support programs (see chapter 4). Several programs from various ministries are already oriented toward small farmers (Alianza PDR, for example). But challenges remain: the numerous programs overlap and lack a clear-cut rationale, investments are so dispersed that effectiveness is reduced, and technical assistance, training, and business focus are insufficient. Technical Note 2 in the Appendix of this chapter summarizes Chile's technical assistance program, which is clearly targeted toward small farmers. 124. Gradual decentralization to state governments of the mandate and resources to support small producers would help address challenges.75 There are several reasons for this: Decentralization would facilitate a territorial approach for rural development, which would help clustering investments (see below). The number, dispersion, and heterogeneity of small producers favor responsibilities at the sub-national level, where constraints and opportunities can best be identified. The sub-national level is also more appropriate for attaining synergy between productive programs and other investments and services (such as NRM, infrastructure or education) favoring small producers. Resources could be more rationally used at the sub-national level if state governments apply them according to state-level strategic plans for the development of the small producers (see below). 125. In principle, such decentralized funds should be targeted mainly at states where it can be most efficiently used by small producers, i.e., where it can get the biggest "bang for buck" in terms of growth and poverty reduction. In practice, 74 Rowenzweig (2008) offers the following as guidelines for such a restructuring: (i) employment generation should be a critical aim of the support system because it is a major way for the growth of the commercial sector to trickle down (crops and farming systems that are labor intensive should thus be supported); (ii) instead of rigid operational rules of present programs, flexibility is needed to operate severeal support instruments, from competitive funds to liquid guarantees, co-financing of relevant investments, kickoff grants, support to strategic alliances between producers and other market agents, tax concessions, and others (the use of these instruments should be selective rather than universal); and (iii) the support system should embrace not only farmers but other agents in the value chains who create synergies. 75 There is a detailed analysis of a proposal to decentralize rural programs to state governments in World Bank (2006a) and World Bank (2007c). 84 political realities will mean that there are some tradeoffs necessary, but at the least, the programs should be designed to allow monitoring of results over time, so that future allocations can take past performance into account. 126. Because decentralization takes time to implement, a starting point could be a discussion between the federal government and state governments to clarify the following aspects: Federal programs, subprograms, and components that could be decentralized--that is, programs that are production-oriented and targeted to small producers. The formula for distributing funds to the states. Other resources that should be transferred, such as staff, vehicles, operation funds, equipment, buildings, and so on. Eventual institutional support to some state governments from the federal level to enable them to effectively carry out their new responsibilities. The broad guidelines and the M&E system to be established by the federal government for the use of funds. Responsibilities for the federal government and state governments. Federal government responsibilities would include setting minimum standards and benchmarks for use of funds by state governments, conditioning fund transfer on preparation of strategies and programs, setting up M&E systems for these strategies and programs, enhancing state government implementation capacity, and auditing use of block grants, independent of the audits by state legislative bodies. State responsibilities would include creating and approving a state strategy for developing the small producer sector and a logical framework with clear objectives and measurable outcomes that could be evaluated as well as any necessary counterpart funding from their own resources. 127. After the transition period, decentralization would imply transferring to state governments, in addition to the present FAIS to finance small rural infrastructure, block grants for the decentralized development of small rural producers (in productive and NRM activities). 76 This, in turn, implies (i) 76 The idea is to decentralize to state governments the budgetary funds of programs oriented toward improving the productive and competitive capacity of small producers. These funds are currently budgeted under federal production-oriented targeted programs of various ministries, such as SEDESOL (. Opciones Productivas), SECON (FONAES), SAGARPA (Alianza PRD), and SRA (FAPPA). Funds would be distributed to the states as block grants, where the resources from different programs would be merged, determined by an objective formula based on demand considerations, which should not be regressive. Future programs that assist farmers in adapting to climate change and managing natural resources to increase positive externalities would also be good candidates for funding under such block grants, since such they would need to be carefully tailored to local agro-climatic and production conditions. 85 phasing out corresponding federal ARD programs and allowing state governments to replace them with their own programs, so that they can exert strong ownership of the rural development agenda for the development of small producers in their states; (ii) pushing ahead with administrative federalization (federalización administrativa) to transfer to state governments the assets, staff, and other operational resources of federal ministries required for state governments to properly operate the newly decentralized funds; and (iii) strengthening the implementation capacity of state governments to be able to takeover these responsibilities. 77 Decentralization of funds to support small producers would facilitate the use of a territorial approach to rural development (Box 6-5). Box 6-5: Applying a Territorial Approach to Rural Development in a Decentralization Option Within the states the regional level is key to implement rural development activities with a territorial approach. The collection of municipalities (regions) in which Mexican state governments divide their states for planning purposes seem appropriate for a territorial approach to rural development--that is, for clustering investment and implementing strategic territorial programs. Strong regional economic coordination institutions are very important to that effect. Some states, for instance Michoacán, have already created regional development councils. If strengthened with technical capabilities and provided funds to co-finance productive programs for small producers, these councils could become the focal point for rural economic development in their territories. In most states, however, the coordination entities for rural development at the regional level are the Comités Distritales de Desarrollo Rural Sustentable (CDDRS). These councils could be appropriate meso- level entities for economic coordination of ARD activities in their regions but their institutional capacity and participatory nature should be strengthened. Regional councils (whether CDDRS or others designed by state governments) could combine participatory principles with sound technical and market criteria to design long-term development strategies for small producers in their territories. These strategies would be based on the potential of territorial assets and the identification of development axes. Investment clustering could be achieved through programs decided in a participatory manner, capable of attracting the interest of a sufficient number of producers in order to reach a critical mass of output, and through the use of a value chain approach. Regional councils could also design large projects capable of triggering rural regional development. To fulfill these functions, regional councils would need to be staffed by core groups of well selected and motivated technical staff, to be paid using the funds decentralized to the states. In addition, they should receive a share of the decentralized resources to co-finance the financial investments in their regional programs. Thus, regional councils would be the main operational vehicle of the rural development strategies and policies of the states. Source: World Bank (2006a). 4) Improving the M&E and Institutional Aspects of ARD Programs 128. Alternatives exist to enhance the M&E system of ARD expenditures. Chapter 5 presented the recent important advances and areas for improvement in the M&E of social programs: evaluations not only of single programs but also of the entire PEC, development and integration of databases of the beneficiaries of ARD 77 See Wolrd Bank (2006a) for further discussion on options for decentralization in Mexico. 86 programs, deeper program evaluations at longer intervals instead of the current annual cycle, and systematic use of evaluation results for managerial and budgetary decisions based on action agendas agreed upon by the evaluators, the implementing agency, and a third party. Adherence to these agendas should be monitored closely. As a logical follow on to the Ley de Desarrollo Rural Sustentable, Mexico could also start a "rural proofing" system to anticipate and monitor national policies' impact on the rural sector. Canada and the United Kingdom both found success with rural proofing systems (called "rural lens" in Canada; Box 6-6). Box 6-6: Rural Proofing in the UK and Canada The U.K. Government's Rural White Paper of 2000 obliges all domestic public entities to put a rural- proofing mechanism in place through to systematically check policy design and implementation for impact on rural areas, and to mitigate impacts where appropriate. Much like the "rural lens" approach in Canada, the process has helped interdepartmental coordination. Interest in rural proofing has spread from central governments, becoming a tool for local authorities to analyze and improve. The British Commission for Rural Communities, which acts as supervisory body to verify that rural proofing has been undertaken, produces an annual report. The Treasury guidelines on SR04 highlight the need for rural proofing, but the responsibility was on departments to translate it into their delivery plans for 2005­08. In 1996, Canada produced the "Thinking Rural" report, detailing new and explicit guidelines for rural policy. The Federal Government created a Rural Secretariat within the Ministry of Agriculture and Agri- Food to administer Canada's rural policymaking and to coordinate federal agency activity affecting rural areas through an interdepartmental working group on rural issues. Subsequently, a Secretary of State for Rural Affairs was established at the Cabinet level. All new policies were subject to a rural policy impact assessment. In 1998 a rural lens checklist was introduced to determine whether a policy or program addresses priorities for rural areas. The checklist covers the priority rural areas and the consideration, delivery, communications, and M&E or rural impacts. The Rural Secretariat administers the Canadian Rural Lens with staff from other departments, scrutinizing new policy initiatives. Rural lens staff can advise the Minister to support (or not to support) new policy proposals. Although the Minister has only one voice at the Cabinet table, opportunities to involve regional development agencies and their ministers are sought. This gives departments an incentive to take the rural lens comments into account. If the rural lens staff think that the rural perspective has not been properly addressed, they can influence the policy proposal accordingly. The objective is not to advocate for putting rural considerations first, but to ensure that decisions are fully informed of the implications for rural communities. Source: OECD (2006), The New Rural Paradigm UK The Countryside Agency (2004), Rural Proofing in 2003/04 129. Alternatives exist to improve program implementation. Implementation challenges are complex and varied. Many are institutional limitations that take time to resolve. Simplifying the number and type of programs would certainly help respond to the implementation challenge, as would the approval of a comprehensive ARD strategy. To help identify and address these issues, CONEVAL has the process evaluations instrument, which could be widely used to maintain the focus on implementation challenges. Issues that could be addressed include: 87 Revising the budgetary and administrative regulations under which ARD programs are implemented to ensure that they provide sufficient flexibility and do not unnecessarily impair or slow down implementation. Paying due attention to the recurrent/operational cost needs of programs, which are particularly large for production-oriented programs. Improving dissemination of programs and the accountability of operators vis- à-vis beneficiaries and using "client satisfaction" criteria to assess performance. Improving the economic incentives and morale of program operators and ensuring that incentives are well aligned with program objectives. 5) Increasing the Positive Environmental Externalities of ARD Programs and Supporting the National Climate Change Strategy 130. Strategies for supporting agriculture and for dealing with climate change need to be mutually reinforcing. For the past two decades over 80 percent of economic losses from weather-related disasters occurred in the agriculture sector. 78 Agricultural policies and climate change policies thus need to be mutually reinforcing by incorporating two principles. First, farmers need to respond to localized changes in climate depending on their individual conditions and constraints. Policies should expand their options rather than encourage them to make choices that do not correspond well to individual circumstances. Second, farmers should not be encouraged to "maladapt" by overusing scarce resources or increasing production in high-risk areas. Both principles have implications for agricultural policies that are congruent with the National Climate Change strategy. 131. Expanding payments for environmental services (PES) programs beyond forestry 79 could create positive synergies between agricultural production and the environment. Specific investment programs to reduce emissions that are identified by the National Climate Change Strategy and that could be scaled up include programs to increase the use of biomass (for example, through high- efficiency wood burning stoves in rural communities and renewable energy sources for rural areas as supported by Fideicomiso de Riesgo Compartido, or FIRCO), livestock programs to rehabilitate degraded rangelands, and 78 Saldaña-Zorrilla (2007). 79 These include Pago por Servicios Hidrologicos Ambientales (PSAH), Programa Nacional de Reforestacion (PRONARE), Programa de Conservación y Restauración de Ecosistemas Forestales (PROCOREF) and Programa de Plantaciones Forestales Comerciales (PRODEPLAN) and would support reforestation, more efficient forest management practices, and recovery of degraded forest lands. Forestry programs, which have seen exponential growth in public expenditure allocation since 2000, could be further scaled up, taking into careful consideration the current and future projection of ecological suitability of the land in terms of soil characteristics and slope, as well as the financial viability of the productive operation, to ensure that the right species of forests are being promoted in areas where they make sense ecologically and financially. 88 hydrometrological hazard risk management and water resources management.80 Scaling up these programs could be accomplished either through national programs or through decentralized programs. Such scaling up should be accompanied by appropriate monitoring measures to ensure that the programs are responding to their objectives. In particular, there is considerable scope to replace distortionary programs such as Ingreso Objetivo with targeted environmentally- friendly payments for environmental services programs, which would reduce distortions and provide income support to small farmers while also providing valuable national and global externalities. 132. The discussion below on Mexico's umbrella forestry program, ProÁrbol, highlights some challenges faced in PES programs and offers suggestions for reorientation. 133. ProÁrbol. The ProÁrbol program is an umbrella framework under CONAFOR that supports Mexico's forestry sector. Its far-reaching goals are aimed at conducting forestry planning, improving production and productivity, promoting conservation and forest restoration, and increasing sectoral competitiveness. In terms of conservation efforts, it houses programs for reforestation, soil restoration, forest fire prevention, sustainable forest management, and payment for environmental services (Pago por Servicios Ambientales del Bosque, PSAB). 81 Previous and current administrations have allocated a significant amount of budget to the sector since 2000, and the ProÁrbol budget has been growing exponentially since,82 despite initial implementation and administrative challenges.83 The program's performance still leaves much to be desired in terms of meeting its goals (productivity, production level, conservation, and competitiveness). 80 This may include improved water resources management programs that emphasize preserving environmental services such as the Alianza para el Pueblo program, which supports efficient use of water for irrigation, modernizing and rehabilitating irrigation districts, strengthening climate data collection, processing, using and disseminating water, and recognizing ancestral practices and lessons learned by communities affected by climate variability. 81 PSAB now receives over MxP 100 million a year in funding. Some 1.4 million hectares were under conservation contracts in early 2008; 2008 contracts should bring this total to over 2 million hectares. PSAH (and, subsequently, the water window of PSAB) pays landowners to conserve existing forests. Payments are made ex post, once the conservation has been verified. Eligibility is based on both spatial criteria (there is a map of eligible areas, based primarily on indicators of importance for water services) and various prioritization criteria (each application receives from 0 to 5 points if they meet the stated criteria, and the applications with most points are accepted); the actual criteria have evolved over time in response to both technical and political considerations. Conservation contracts are for five years, and are renewable providing applicants have sufficient points. Payments are uniform countrywide. They are stated in multiples of the minimum wage, and amount to about US$40 per hectare per year for cloud forests and US$30 per hectare per year for other forests. 82 Budget for the forestry sector in Mexico grew by 2,094 percent between 2001 and 2008 (Merino, Rodriguez et. al., 2008). 83 For example, to streamline the work, the program initially conducted all transactions by bank transfers, rather than by checks, which caused some delays and discontent among beneficiaries. Also, some administrative guidelines--such as the verification procedures for plantations subsidies-- were delayed. 89 134. The status quo of the PSAB (ProÁrbol). Mexico undertook careful preparation for the PSAB program. But these good intentions were overwhelmed by intense implementation pressures. For example, criteria for payments for hydrological environmental services were changed to spread payments more broadly, irrespective of relative importance for water services, when it became clear that focusing payments on areas where aquifers are most overexploited would concentrate payments in only a few states. The program has also been used as a vehicle to address unrelated activities, such as a commitment that Mexico made at the Bishkek Mountain Summit to increase spending on conservation in mountain areas; these were added to the eligibility criteria. The result was very poor targeting, at least initially (Muñoz et. al. 2008). From 2003 to 2005 as much as 90 percent of forest area under contract was in areas with aquifers in equilibrium or underexploited aquifers, and as much as 72 percent was in areas of low or very low risk of deforestation. More recent assessments are not available, but efficiency is thought to have increased (for example, location in an area of high deforestation risk is now a priority criterion). A politically-driven requirement for uniform payments also means that payments are often ill-suited to local conditions--paying much more than opportunity costs in some areas (resulting in much higher demand for participation than funding allows) and much less than opportunity costs in other areas (resulting in limited participation in areas that could provide very high levels of environmental services). 135. Reorientation of PSAB (ProÁrbol). PSAB shows that programs aimed at improving environmental conditions are feasible and can attract considerable interest from land users. But it also shows that such programs are as vulnerable as other public programs to pressures that divert them from their stated objectives. Considerable efforts are needed to ensure that these programs reach their objectives efficiently.84 Even so, even an imperfect program aimed at improving environmental conditions may well be better than one that generates negative externalities. Technical Note 3 in the Appendix of this chapter presents some international perspectives on programs for payments for environmental services. 136. Table 6-1 summarizes possible actions for each of the five policy themes discussed above, identifying tradeoffs among the various criteria of policy options such as cost, sequential importance, technical difficulty, risks, and impact. The matrix highlights that while some options promise greater benefits in the longer run, others could be targeted for quick action. Actions that are likely to have the highest impact, and yet could be tackled in the short term include those to improve the overall national ARD planning system, M&E system and some changes in institutional or operational aspects. Other actions could be taken on as 84 The World Bank, with GEF co-financing, is supporting a project to increase the efficiency of Mexico's PSAB program. The Mexico Environmental Services Project has three components: developing new financing sources based on payments by local service users (primarily water users, but also biodiversity users such as the tourism industry, and carbon buyers), increasing the efficiency of the current payment program (primarily by moving away from the current countrywide one-size-fits-all approach to a differentiated approach tailored to the conditions of particular areas), and one to support participation by poorer ejidos. 90 a mid to long term goal, such as decentralization of small farmer programs to state governments. However, even for this longer term action, concrete immediate steps are also identified, such as identifying the responsibilities of each level of government and national benchmarks, that would facilitate the transition to the longer term goal but could be implemented in the short term. Table 6-1: Summary Matrix of Policy Actions Technical Implemen- Likely Difficulty/ Fiscal Cost Policy Options tation Impact Risks or Savings 1. Improving the ARD Planning System (1) Prepare and approve a national Short High Medium Low Cost ARD strategy (2) Reposition CIDRS within the federal government to strengthen its No Cost or coordination capacity, and build an Short Medium High Savings effective ARD budgetary planning system (3) Create a nucleus of high-level agriculture policy analysts under Short Medium Low Low Cost SAGARPA umbrella. 2. Rationalizing the Overall System for Farmer Support (1) Increase expenditure allocation to public goods: irrigation improvement and management, drainage and transport infrastructure; other rural Short, Medium infrastructure (such as that provided High Medium High Cost and Long by FAIS); research, extension and training; sanitary and phyto-sanitary services; market intelligence; emergency programs; and others (2) Modify PROCAMPO No Cost or Short Medium Medium Savings (3) Discontinue PROCAMPO Medium High High High Savings (4) Modify Ingreso Objetivo No Cost or Short Medium Medium Savings (5) Discontinue Ingreso Objetivo Medium High High High Savings (6) Modify input support programs No Cost or Short Medium Medium (tarifa 9, agro-diesel, others) Savings (7) Discontinue input support programs with or without temporary Short High High High Savings compensatory payments (tarifa 9, agro-diesel, others) (8) Rationalize federal non-targeted programs oriented to support on-farm and value chain investments (mainly PROGAN, Apoyos a la No Cost or Short, Medium High High Competitividad de las Ramas Savings Productivas/Adquisición de Activos Productivos, and the non-targeted programs of Alianza). (9) Reduce the allocation to non- Short, Medium High Low Medium 91 Technical Implemen- Likely Difficulty/ Fiscal Cost Policy Options tation Impact Risks or Savings targeted production-oriented programs and Long Savings to increase support for small producers (see 3 (1))* (10) Increase support to the rural Short, Medium Medium Cost nonfarm sector High High and Long (11) Strengthen the rural finance Short, sector to increase service access Medium, and High High Medium Cost Long 3. Improving the Support System for Small Producers through gradual decentralization (1) Increase expenditure allocations to Short, Medium Medium High Low support small producers* and Long Cost (2) Pool resources from targeted production/NRM-oriented programs of different federal entities, decentralize those resources to state governments No Cost or Medium, Long High High according to an agreed distribution Savings formula and decentralization protocol, and discontinue the corresponding federal programs (3) Accelerate the decentralization of Short, ARD federal offices (federalización Medium, and Medium High Low Cost administrativa) Long (4) Agree with state governments on a protocol containing guidelines for decentralization of funds to support small producers: (i) national No Cost or Short Medium Medium benchmarks, (ii) fund distribution Savings formula, (iii) M&E system, (iv) implementation support system, and (v) state government responsibilities 4. Improving the M&E and Institutional Aspects of ARD Programs (1) Adjust budgetary and administrative regulations of ARD No Cost or programs to ensure flexibility and Short Medium High Savings timely execution (2) Increase recurrent cost allocations Short, Medium for production oriented programs and Long Medium Medium Medium Cost (3) Improve the dissemination of ARD Short, Medium programs and Long Medium Low Low Cost (4) Improve the accountability of Short, Medium program operators at all levels and and Long High Medium Low Cost introduce "client satisfaction" criteria to assess performance (5) Improve the economic and moral Short, Medium incentives of medium and bottom level and Long High Medium Medium Cost program operators (6) Carry out evaluations of the entire Short, Medium Medium Low/Medium Low Cost PEC and Long (7) Develop and integrate databases of Short, Medium Medium Medium Low Cost 92 Technical Implemen- Likely Difficulty/ Fiscal Cost Policy Options tation Impact Risks or Savings ARD program beneficiaries and Long (8) Carry out in-depth program Short, Medium Medium Low Low Cost evaluations at less frequent intervals and Long (9) Organize a system to agree on and monitor action agendas deriving from Short Medium Low Low Cost evaluation results 5. Increasing the Positive Environmental Externalities of ARD programs and Supporting the National Climate Change Strategy (1) Improve SAGARPA norms to No Cost or regulate slash and burn agriculture in Short Medium Low Savings light of increasing forest fires (2) Increase research allocation to research and technology transfer of reducing emissions from agriculture Short, Medium High Medium Medium Cost such as no-till technology, reducing and Long GHG gas from livestock, reforestation, efficient use of inputs (3) Revise regulations to incentivize Medium farmers to sell electricity generated by Medium Medium High Savings biomass to the grid** (4) Improve targeting of geographical area and payment levels of payment Short, Medium High Medium High Cost for environmental services programs, and Long and scaling up its operations (5) Replace distortionary support Short, Medium No Cost or programs with payments for High Medium and Long Savings environmental services Implementation term: Short (less than 2 years), generally associated with changes in laws or regulations; Medium (3­5 years), requiring some institutional changes; Long (more than 5 years). Effects term: Short (less than 2 years); Moderate (3­5 years); Long (more than 5 years). Impact: High, medium, low (relative to other policy options proposed in these notes, not relative to other more general reform options). Technical difficulty/risks of undertaking the options suggested: High, medium, low. Fiscal cost to government: High (large public investment programs), medium (some public program expenditure required), low (little public expenditure). Fiscal saving: high, medium, low. * The suggested proposal is to shift resources from non-targeted production oriented programs to targeted ones (eventually decentralized to state governments) to support small farmers. Thus, the net effect should be budget neutral--that is the overall expenditure should not increase or decrease. ** The suggested proposal is expected to increase investments by farmers to generate biomass energy. Although, this proposal per se is budget neutral, in the long-run this should result in fiscal cost saving since as farmers begin to generate biomass energy, they would buy less subsidized electricity from the grid. 93 APPENDIX TO CHAPTER 6: TECHNICAL NOTES Technical Note 1: New Zealand's Successful Agriculture Policy Reform Experience New Zealand historically pursued a policy of agricultural protection (1960-1980), but after structural macroeconomic adjustment and trade openness the agricultural sector experienced a gradual liberalization from subsidies and price controls. In the period from the 1960s until 1980 New Zealand had strong price controls and subsidies to farmers in order to keep production levels constant while encouraging exports (Johnson, R.W.N., 2000). By 1983, for instance, New Zealand's producer support estimate (PSE) was 35% of producers' income and the effective rate of assistance surged to 123% of total rural household income (OECD, 2008). After 1983, New Zealand Government began to reduce export and farm subsidies in the agricultural sector. During the 1990s further deregulation reforms enabled the agriculture sector to benefit from greater synergies, better use of resources, and a competitive industry structure that allowed better adjustment to markets. These reforms led to an improvement of TFP growth to average 2.5% a year compared to a 1.5% growth in the pre-reform period. Source: MAF, 2006 Policy Impacts of Reforms Sector Growth From the period between the mid-1960s until the mid-1980s the share of agriculture production to GDP fell steadily from 14% to 5.7%, given heavy regulation, price distortions, low incentives to production and unfavorable exchange rates. A decade after deregulation in the agriculture labor productivity has doubled. Despite less land devoted to livestock and arable farming--from 14 million hectares in the 1980s to 12 million hectares in 2000s--, productivity increased 85% in a 20 year period. 94 Source: MAF, 2006 The competitive market structure generated after the reforms led to smooth adjustments to external shocks and, as a result, resources were more efficiently allocated to high productivity sectors. For example, among pastoral industries the reformed caused resources to divert into dairy farming, a highly productive sub-sector. Consequently, total stock units in the dairy sector increased by 65% while beef and sheep sectors declined by 2 and 43%, respectively. A consequence of reforms was to allocate more resources into innovation, research and other important services that foster agricultural production efficiency. The high value of agriculture production has allowed New Zealand to have one of the highest propensities to invest ratios, and one of the highest percentages of GSSE in relation to TSE for the development in Agriculture among OECD countries (OECD, 2008). New Zealand's GSSE on Agriculture 1986-2008 of TSE GSSE as % ofTSE 90 Mexico New Zealand United States European Union OECD- Total 80 70 60 50 40 30 20 10 0 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 Source: OECD, 2008 References: Johnson, R.W. 2001. "New Zealand's Agricultural Reforms and their International Implications". IEA, U.K. Anderson, K., Martin, W., and Valenzuela, E. 2006. "The Relative Importance of Global Agricultural Subsidies and Market Access". World Trade Review. Vol. 5 No. 3. 95 MAF. Agriculture in New Zealand: Past, Present and Future. ABARE and New Zealand Ministry of Agriculture and Forestry 2008 OECD. 2008. Case Study: Domestic Reform, Trade, Innovation and Growth in New Zealand's Agricultural Sector. TPWP No. 74. Paris. Technical Note 2: INDAP: Technology Transfer Targeting Small Farmers in Chile Chile's Institute for Agricultural Development (INDAP) began in 1962 as an agency of the Ministry of Agriculture, oriented exclusively at enhancing the productivity of small, family farms through lending and technical assistance. INDAP's definition of small farm families, which has remained unchanged over the decades, is simple: families with 12 or fewer hectares (irrigated equivalent), with a net worth less of than US$100,000, and with income deriving principally from agriculture. Other than INDAP, there is no government- supported extension service; commercial operations are left to contract such services with private agents and remain unsubsidized. INDAP is also the only significant source of formal long-term credit for the low-income, small-farm sector. Chile was one of the first countries to introduce public funding and private delivery of agricultural technical assistance. INDAP has evolved over time while maintaining a degree of continuity as it adjusts to changing conditions. The basic principle of underwriting private extension services for only small farms has been maintained for three decades, although the structures of INDAP's operation have adapted to experience and new realities. The overall approach has proven itself operationally as a manner of delivering extension, although its success in graduating farmers is less clear. It has also maintained political support and funding, which have increased in real terms. Questions remain about program coverage and success in aiding viable farmers to reach commercial competitiveness. Impact of INDAP's small farmer assistance programs The growth in clients of INDAP's technical assistance programs and subsidies and expenditures has been well documented. A third of INDAP's transfers are now credit subsidies, and two-thirds are subsidies for on-farm investment, technical assistance, and managerial training (either directly to individual farmers or to farm groups and municipality-run programs in poor areas). According to the OECD, INDAP's credit and technical assistance programs reached in some form (perhaps minimally) approximately 116,000 people in 2006, or 42 percent of small farms (as counted in the 1997 agricultural census). The share receiving significant technical assistance is approximately half. The role of INDAP's technical assistance in moving farmers into a sustainable, higher- income trajectory--their insertion into commercial structures--is less well known, and there are simply no good measures of how many former clients have left the program and succeeded on their own. Unlike the USDA extension service, which is meant to be a permanent support for information dissemination to agriculture as an industry and not specifically oriented to small farmers, INDAP is meant to focus on a specific subgroup of poorer farmers who could, with assistance, reach viable scales. 96 The two most recent evaluations conducted in the mid-1990s show that INDAP contributes to more intense agricultural input use, including family labor, which increases gross and net farm income. Lopez (2000), however, finds that INDAP programs do not raise family incomes due to an off-setting reduction in off-farm income and that more intense input use does not raise total factor productivity. The key determinant in total household income appears to be internal household characteristics rather than external factors, including geographic location. Berdegue and others (1998) find that INDAP programs increase household income, but the effect could have been due to selection bias. They also note that INDAP has less of an effect in some geographic areas. But these studies should be complemented by an examination of what happens to these families over time, an analysis that INDAP has neglected. INDAP's basic mission--to graduate viable farmers into commercial agriculture--has not been explicitly used in evaluating the program. It is unclear whether INDAP is treating new cohorts of farmers over time or aiding the same group of beneficiaries repeatedly. The number of small farmers is declining, but INDAP has been growing. During the early 1990s INDAP's budget increased 11 percent a year in real terms, and 4.4 percent in the late 1990s. By the early 2000s government funding still covered about 85 percent of INDAP costs. Some Chilean experts speculate that the agency attends to more or less the same group of beneficiaries over time, implying that the fiscal outlays per client could amount up to US$3,000 or more per year--or about US$50,000 over the last 15 years. These would be large sums relative to the incomes of the Chilean rural poor. Panel data following INDAP clients would be useful for finding the proportion of farmers that reach commercial viability and the factors contributing to their success.. When applying the lessons from Chile's INDAP to other countries, this deficiency should be remedied. A summary of best practices for technical assistance from INDAP's experience Define the target audience within the small farm sector, identifying the subset of farmers that could eventually generate sustainable income without government subsidies. Do not pretend to cover all small farmers at the same time. Set up a monitoring system to assess a program's ability to graduate clients to sustainability after five to seven years. Periodic quantitative evaluations are needed (recognizing selection effects between clients and the control group) that go beyond detecting impacts on production and input use to assessing the impact on household incomes from all sources. For farm families that cannot reach a viable scale in production, focus aid on transition to other income activities and on social safety nets and noncontributory pensions. To reduce the costs and inefficiencies associated with government programs, do not subsidize extension service for commercial farms. There is a rationale to support basic and adaptive research for agriculture as a whole. Government support for--and supervision of--private delivery of extension services can provide technical assistance more efficiently than embedding extension delivery in a government bureaucracy. 97 Do not rely on a voucher system that asks individual small farmers to make contracts with extension providers because it may not generate the scale of demand that could sustain a private market for extension services. Use centralized coordination of extension services for standardized commodities, not for products that require closer vertical coordination along the marketing chain. Recognize that an extension delivery system cannot "get it right" from the beginning, but must adjust to mistakes and changing conditions. Information and monitoring, and flexibility, are key to adjustment. References Diaz, Juan. 2007. "Family Farm Agriculture: Factors Limiting its Competitiveness and Policy Suggestions," note prepared for the OECD Review of Agricultural Policies: Chile. Berdegue, Julio, and RIMISP (Red Internacional de Metodologías de Investigación en Sistemas de Producción), STOAS, Grupo de Investigaciones Agrarias (GIA) and EMG Consultores. 1998. "Evaluación de Instrumentos de Fomento Productivo: El Programa de Transferencia Tecnológica de INDAP." Santiago, Chile. Lopez, Ramon. 2000. Determinants of Rural Poverty in Chile: Evaluating the Role of Public Extension/Credit Programs and Other Factors," Chapter 9 of R. Lopez and A. Valdes, eds., Rural Poverty in Latin America, St. Martin's Press. Sotomayor, Octavio. 1994. "Políticas de modernización y reconversión de la pequeña agricultura tradicional Chilena," ODEPA and IICA, Santiago. Source: Valdes and Foster (2008). Technical Note 3: Payments for Environmental Services in Natural Resource Sectors Managing natural assets well is difficult because the value of many--if not most--of their environmental services cannot be internalized by private actors; they are externalities and public goods. Individuals in society generally value natural assets, but the incentives guiding private action rarely reflect their value to society as a whole. The key to better managing natural assets is to design institutions that can measure the social benefits and costs of environmental services and translate that information into appropriate incentives for individual decisionmakers, such as farmers. The classic policy proposals for better management focused on reducing negative externalities: environmental taxes on emissions, mandated best practices, property rights allocation, and support for institutions for common property management. In agriculture and forestry environmental taxes are rarely used due to the high cost of monitoring emissions. Emissions permits outside agriculture tend to be pollutant specific. Better known in the natural resource sector are best management practices to reduce environmental damage. Such practices are now more common--for example, when managing native forests--and are often pushed by tourism-linked considerations and by consumer preferences in wealthier countries valuing not only the product but the environmental friendliness of production. In fact, there is growing emphasis on private, third-party certification and voluntary implementation of standards, as with organic produce in fresh fruits and vegetables for exports. This is due in part to the normal 98 slowness of governments as many developing countries to adjust the design and enforcement of best practices. Recent policies have emphasized both market-oriented and government-supported subsidies of environmental goods and services. These policies, such as incentives for better land management, focus on Payments for Environment Services (PES) provided by those who benefit from such services, including local, regional, and global beneficiaries. PES programs can be found increasingly in both developed and developing countries, ranging from contracts between farmers and local industries whose profits are sensitive to water availability, to contracts between one country's industry and groups in distant countries seeking carbon sequestration. According to the FAO, there are four main PES markets: climate change mitigation, watershed services, biodiversity conservation, and landscape aesthetics. In developing countries markets for reducing carbon emission and conserving biodiversity are potential sources of new revenue for agriculture. The ultimate beneficiaries of environmental services are usually dispersed, so intermediaries such as governments and international public and private concerned entities are required. To date, PES programs have been pushed by governments and are geographically focused. They are usually direct payments in exchange for better land management, such as soil and water conservation measures, and tree plantations for carbon sequestration. Beneficiaries can pay via a surcharge to water bills or with fees for park visitors. There are two key steps for designing efficient PES programs. First is targeting farmers or other managers of natural assets whose adoption of better management practices can yield the highest environmental benefits for the lowest costs. Second is the structure of compensation, which not only induces better management but also is "self-enforcing" by reducing monitoring costs and assuring compliance. Perhaps the most studied case of PES in Latin America is the Costa Rican National Forestry Financing Fund, which began in 1997 and has led to similar efforts in other countries. The fund officially recognizes that forest owners provide a bundle of environmental services, including watershed protection, biodiversity conservation, scenic beauty, and carbon fixation and sequestration. The PES system arose when the government was searching for sustainable funding for both forest conservation and aid to the forestry sector. A 2002 evaluation of the program found that the Costa Rica PES had contracted more than 280,000 hectares of private forests, with more than 800,000 hectares pending, spending US$57 million between 1997 and 2002. Forest conservation contracts paid over US$210 per hectare, and reforestation contracts US$538 per hectare. Funding came primarily from fuel taxes, complemented by international donors. One special consideration is the potential role of PES in alleviating poverty. Associated with poverty are low levels of education, ill-defined property rights over natural assets, and a lack of capital that could support households during long-term investments in natural assets (such as planting trees). Poverty often leads to environmental degradation. Being able to target poor farmers for PES would bring both environmental benefits and poverty reduction. Yet while targeting the poor on paper might be logical, implementing a PES plan would be constrained by the very structural characteristics that link poverty to 99 environmental degradation in the first place: ill-defined property rights that make contracting and long-term projects uncertain for both buyer and seller of the service; small farm holdings that increase the transaction costs per unit of compensation and reduce the viability of better management (such as crop rotation) and thus increase the required compensation to entice farmers to better management practices; and lack of credit as a short-term backup makes farmers more risk averse to adopting a new system of management. Nevertheless, the World Bank, while recognizing that PES is not in itself an antipoverty strategy, has suggested some pro-poor considerations in PES design: devise specific mechanisms to counter high transaction costs and understand the social context in which a PES would be applied to avoid adverse impacts on the poor and design appropriate remedial measures. Making PES programs pro-poor, however, might increase their costs, so funding must be found beyond the beneficiaries of environment services. Donors who are more directly concerned about poverty could finance these additional costs. Both the World Bank and the FAO have programs to help policymakers and interested parties design PES strategies. The World Bank has supported PES projects in many countries, particularly in Latin America. Past World Bank­supported PES projects have been implemented in Colombia, Costa Rica, and Nicaragua; projects are currently under way in Costa Rica, Mexico, and Panama; and projects are in preparation in Brazil, Colombia, and Ecuador. FAO implements a program called Payment for Environmental Services for Agricultural Landscapes (PESAL), under its Agricultural Development Economic Division (ESA). The ESA provides web links and tools for designing and implementing PES with information on ecosystem services markets, recommendations on setting up PES schemes, and links other international organizations in this field. Source: Valdes (2008b). 100 101 7. REFERENCES Asad, M., and A. Dinar. 2006, "The role of water policy in Mexico: Sustainability, equity, and economic growth considerations." Latin America and the Caribbean Region Sustainable Development Working Paper No.27. Washington: World Bank. Avila, A. F., and R. E. Evenson. 2004, Total Factor Productivity Growth in Agriculture: The Role of Technological Capital. Chap. 31 in R. E. Evenson, P. Pingali, and T. P. Schultz, eds., Handbook of Agricultural Economics. Amsterdam: Elsevier. Buchenau, J and del Angel, G, 2007, "Consultoría sobre Servicios Financieros Rurales" (background paper for the IDB study on Mexico Agriculture Public Expenditure, 2007). Cortés, F, Felipe López Calva, Rodolfo de la Torre, 2005, "Evaluación de la pobreza en México", Seminario Universitario de la Cuestión Social, Instituto de Investigaciones Sociales, UNAM. August 24, 2005. Engel, S., S. Pagiola, and S. Wunder. 2008, "Designing payments for environmental services in theory and practice: An overview of the issues." Ecological Economics, Special Issue on Payments for Environmental Services in Developing and Developed Countries, 65:4, pp.663-674. Fan, S., Hazell, P., Thorat, S., 1999, "Linkages between government spending, growth and poverty in rural India". FAO, 2005a, Evaluación Alianza para el Campo 2004. FAO, 2005b, "Base de Datos de estadisticas e indicadorese del gasto publico agricola y rural (GPAGRURAL), Oficina regional para American Latina y el Caribe, Santiago. FAO-SAGARPA, 2004, Evaluación Nacional de Alianza para el Campo, Informe General. Gordillo, G., De Janvri, A. and E. Sadoulet, 1999, La Segunda Reforma Agraria de México: respuestas de familias y comunidades, 1990-1994, FCE, Mexico. 102 Guerrero García Rojas, H., 2007, "Politicas y gasto publico federal en infraestructura de riego" (background paper for the IDB study on Mexico Agriculture Public Expenditure, 2007). Klepeis, P. and Vance, C., 2003, " Neoliberal policy and deforestation in Southeastern Mexico: An assessment of the PROCAMPO program", Economic Geography, Volume 79, Issue 3. Knutson (2007), "Assessment of Mexico's Sanitary, Phytosanitary, and Food Safety Policies and Programs and Their Implementation: Diagnosis and Proposals for Reforms" (background paper for the IDB study on Mexico Agriculture Public Expenditure, 2007). Leyva-Parra, G., 2005, "El Ajuste del Ingreso de la ENIGH con la Contabilidad Nacional y la Medición de la Pobreza en México", in Székely (ed.). Lopez, R and Galinato G. I., 2007, "Should Governments stop subsidies to private oods? Evidence from rural Latin America", Journal of Public Economics 91, 1071-1094. Merino, L., Rodríguez, J., et. al., 2008, "Estudio Estratégico Sobre el Sector Forestal Mexicano". (Unpublished draft prepared for the World Bank). Muñoz Piña, C., Avila Forcada, S. et. al.,2006, "Agricultural Demand for Groundwater in Mexico: Impact of water Rights enforcement and electricity user fees on groundwater level and quality" (background paper for the World Bank study "Economic Assessment of Policy Interventions in the Water Sector". Muñoz-Piña, C., A. Guevara, J.M. Torres, and J. Braña. 2008. "Paying for the hydrological services of Mexico's forests: Analysis, negotiations and results." Ecological Economics, Special Issue on Payments for Environmental Services in Developing and Developed Countries, 65:4, pp.725-736. Musgrave R., 1959, "The Theory of Public Finance". McGraw-Hill, New York. OECD, 2006a, "Agricultural and Fisheries Policies in Mexico". OECD, 2006b, "OECD Rural Policy Reviews: the New Rural Paradigm: Policies and Governance" OECD, 2007a, "Producer and Consumer Support Estimates OECD Database 1986-2006". OECD, 2007b, "Producer and Consumer Support Estimates OECD Database 1986-2006: USER'S GUIDE". OECD, 2007c, "Mexico: Estimates of support to agriculture, Producer and Consumer Support Estimates OECD Database 1986-2006". 103 OECD, 2007d, "OECD Rural Policy Reviews: Mexico" Oliver, L.G. and Santillanes, S., 2008, "Cuantificación y clasificación del gasto público rural en México", Background paper for this study. Pagiola, S., G. Platais, and A. Arcenas. 2005, "Can payments for environmental services help reduce poverty? An exploration of the issues and the evidence to date from Latin America." World Development, 33, pp.237-253. Pagiola, S., and G. Platais. 2007, Payments for Environmental Services: From Theory to Practice. Washington: World Bank. Pagiola, S. 2008, "Payments for environmental services in Costa Rica." Ecological Economics, Special Issue on Payments for Environmental Services in Developing and Developed Countries, 65:4, pp.712-724. Rosenzweig. A., 2008, "Apuntes para el Grupo de Trabajo sobre Procampo: la Política de Fomento Productivo en México en el Sector Agropecuario", Sagarpa, mimeo. Ruiz-Castillo, J., 2005, An evaluation of "El ingreso rural y la producción agropecuaria en méxico 1989-2002", published in 2004 by the SIAP (Servicio de Información y Estadística Agroalimentaria y Pesquera) of SAGARPA (Secretaría de Agricultura, Ganadería, Desarrollo Rural, Pesca y Alimentación). Saldaña-Zorrilla, S.O., 2007, Socioeconomic vulnerability to natural disasters in Mexico: rural poor, trade and public response"", CEPAL, Mexico. Scott, J., 2005, "Transferencias Públicas (y Otros Ingresos) en Especie en la Medición de la Pobreza", in Székely (ed.), 2005. Scott, J., 2008, "Assessment of ARD Public Expenditures", Background paper for this study. Sumner, D.A. and Balagtas J.V., 2007, "Economic Analysis of the Ingreso Objetivo Program in Mexico" (background paper for the IDB study on Mexico Agriculture Public Expenditure, 2007). Székely, M. and Rascón, E., 2004, "México 2000-2002: Reducción de la Pobreza con Estabilidad y Expansión de Programas Sociales", May 2004. Székely, M. (ed.), 2005, "Números que Mueven al Mundo: La Medición de la Pobreza en México", ANUIES-CIDE-SEDESOL-Miguel Ángel Porrúa. Taylor, Yunez Naude, Gonzalez, 2007, "Estudios sobre Politicas Publicas para el Sector Rural en Mexico" (background paper for the IDB study on Mexico Agriculture Public Expenditure, 2007). 104 Tendler, J, 1997, "Good Government in the Tropics", Baltimore: The John Hopkins University Press. Valdes, A., 2008a, "Agricultural Public Spending: Description and Assessment Relevant to Latin America", Background paper for this study. Valdes, A. 2008b, "Payments for Environmental Services in Natural Resource Sectors", Background paper for this study. Valdes A. and Foster, W., 2008, "INDAP: Public Funding and Private Delivery Technology Transfer targeting Small Farmers in Chile", Background paper for this study. Winter, P. and Davis, B., 2007, "Designing a new PROCAMPO program: Lessons from Oportunidades" (background paper for the IDB study on Mexico Agriculture Public Expenditure, 2007). World Bank, 2002, "State-level Public Expenditure Review: The Case of Guanajuato". World Bank, 2003, "State-level Public Expenditure Review: The Case of Veracruz- Llave". World Bank, 2004a, "Mexico Public Expenditure Review". World Bank, 2004b, "Access to Finance for Farmers and Agriculture Enterprises" (Unpublished). World Bank 2004c, "Poverty in Mexico: an Assessment of Conditions, Trends and the Government Strategy". World Bank, 2005a, "Beyond the City the Rural Contribution to Development". World Bank, 2005b, "Income Generation and Social Protection for the Poor". World Bank, 2005c, "Infrastructure Public Expenditure Review (PER)". World Bank, 2006a, "Decentralization, Poverty, and Development in Mexico". World Bank, 2006b, "Mexico's Competitiveness: Reaching its Potential". World Bank, 2007a, "Water Public Expenditure Review" (Unpublished). World Bank, 2007b, "Policy Note on Sensitive Crops (Corn, Sugar, Beans)" (Unpublished). 105 World Bank, 2007c,"Mexico 2006-2001: Creating the Foundations for Equitable Growth". Policy Notes. World Bank. 2007d, "State and trends of the carbon market 2007." Washington: World Bank. World Bank, 2008a, "World Development Report 2008 Agriculture for Development. World Bank, 2008b, "Climate Change Aspects in Agriculture: Mexico Country Note", Unpublished. World Bank, 2008c, "Distortions to Agricultural Incentives in Latin America", Edited by K. Anderson and A. Valdes. World Bank, 2008d, "Poverty and Social Impact Analysis of Groundwater Over- exploitation in Mexico" (Draft). 106