Mexico - Agricultural sector memorandum

13425 Docwnent of The World Bank FOR OFFICIAL USE ONLY Report No. 13425-ME MEXICO AGRICULTURAL SECTOR :MEMORANDUM September 16, 1994 Natural Resources and Rural Poverty Division Country Department n Latin Arilerica and the Caribbean Regional Office This dotumeot has a restricted distribution and may be used by recipients only in tbe peformance of tbei.: ofiacial duties. Its contents may not otberwise be disclosed without World Bank authorization.  CURRENCY EQUIVALENTS Currency Unit = Mexican New Peso (N$) US$l = N$3.3 US$303;030 = N$l million (July 1994) FISCAL YEAR January 1 - December 31 WEIGmS AND .MEASURES 1 meter (m) = 3.28 feet (ft) 1 kilometer (km) = 0.62 mile (mi) 1 hectare (ha) = 10,000 m2 = 2.47 acres 1 square kilometer (km'l) = 0.38 square miles (mi'l) = 100 ha 1 metric ton (m ton) = 2,205 pounds  ABBREVIATIONS AND ACRONYMS AGROASEMEX Agroaseguradora de Mexico (Mexican Agricultural Insurance Company) ANA GSA Aseguradora Nacional Agricola, S.A. (National Agricultural Insurance Company) ASERCA Apoyos y Servicios a la Comercializaci6n Agricola (Support Services for Agricultural Marketing) ATP Agricultural Technology Project (Loan 3465-ME) AZUCAR National Sugar Company BANPESCA Banco N acional Pesquero y Portuario (National Bank of Fisheries and Ports) BANRURAL Banco Nacional de Credito Rural (National Rural Credit Bank) BANXICO Banco de Mexico (Bank of Mexico) CCVP Comire Cor.sultativo de Variedades de Plantas (Consultative Committee on Plant Varieties) CDS Convenio de Desarrollo Social (Social Development Agreement) CETES Certificados de la Tesorerfa (Treasury Bills) CFE Comisi6n Federal de Electricidad (Federal Electricity Commission) CFM Commodity Futures Market CICOPLAFEST Comisi6n Intersectorial para el Control del Procesarniento y el Uso de Pesticidas, Fertilizantes y Substancias T6xicas (lntersectoral Commission for the Control of Processing and the Use of Pesticides, Fertilizers, and Toxic Substances) CIMMYT Centro Internacional de Mejorarniento de Mafz y Trigo (International Wheat and Maize Improvement Center) CNA Comisi6n Nacional del Agua (National Water Commission) CONADECA Comejo Nacional para el Desarrollo del Cacao (National Cocoa Development Council) CONACAL Comisi6n Nacional de Caminos Alimentadores (National Feeer Road Commission) CONACYT Comejo Nacional de Ciencia y Tecnologfa (National Science and Technology Council) CONAFRUT Compaiifa Nacional de Frutas (National Fruit Company) CONAPE Comejo Nacional de Poblaci6n (National Population Council) CONASUPO Compaii(a Nacional de Subsistencias Populares (National Basic Foods Company) COPLADE Comite de Planificaci6n Estatal (State Planning Committee) CPCh Colegio de Postgraduados de Chapingo (Chapingo Postgraduate College) CUD Convenio Unico de Desarrollo (Development Agreement between the Federal Government and the States) DDR Distrito de Desarrollo Rural (Rural Development District) DFI Development Finance Institution DGCSA Direcci6n General de Conservaci6n de Suelos y Agua (General Directorate of Soil and Water Conservation) DRD Decentralization and Regional Development Project (Loan 331O-ME) FAOICP Food and Agriculture Organization Cooperative Program FEFA Fondo Especial para Financiamientos Agropecuarios (Special Trust Fund for Agricultural Financing) FEGA Fondo Especial de Asistencia Tecnica y Garant(a para Creditos Agropecuarios (Agriculture Technical Assistance and Loan Guarantee Trust Fund) FERTIMEX Fertil izantes de Mexico (Mexican Fertilizer Company) FICART Fideicomiso para Credito en Areas de Riego y Temporal (frust Fund for Rainfed and Irrigated Areas) FIDELIQ Fideicomiso Liquidador de Instituciones y OrganizaCiones Auxiliares (frust Fund for the Liquidation of Auxiliary Institutions and Organizations) FINAGRO Fideicomiso Nacional para el Mercado de Granos y Oleaginosas (National Trust Fund for Grains and Oilseeds Markets) FlRA Fideicomisos Instituidos en Relaci6n con la Agricultura (frust Fund for Agriculture) FIRCAVEN Fideicomiso para la Restructurad6n de la Canera Vencida (frust Fund to Restructure the Overdue Portfolio of BANRURAL) FIRCO Fideicomiso de Riesgo Companido (frust Fund for Shared Risk) fMIA fundaci6n Mexicana de Ia Investigaci6n Agrfcola (Mexican Foundation for Agricultural Research) FONDO Fondo de Garantfa y Fomento para la Agricultura, Ganaderfa y Avicultura (Credit and Guarantee Trust Fund for Crop, Livestock, and Poultry) FONAES Fondo Nacional de Empresas de Solidaridad (National Fund for Solidarity Businesses) FOPESCA Fondo de Garantfa y Fomento para las Actividades Pesqueras (Fisheries Credit and Guarantee Trust Fund) FRR Financial Rate of Return FSAL Financial Sector Adjustment Loan (Loan 3085-ME) GAIT General Agreement on Trade and Tariffs GDP Gross Domestic Product GLASOD Global Assessment of Land Degradation GOM Government of Mexico HYV High Yielding Variety IARC International Agricultural Research Center IDB Inter-American Development Bank ID Irrigation District IDSP Irrigation and Drainage Sector Project (Loan 3419-ME) INE Instituto Nacional de Ecologfa (National Ecology Institute) INI Instituto Nacional Indigenista (National Indigenous Institute) INMECAFE Instituto Mexicano del Cafe (Mexican Coffee Institute) IMPA Instituto para el Mejoramiento de la Producci6n de Azucar (Institute for Sugar Cane Production Improvement) INEGI Instituto Nacional de Estadistica. Geograffa e Informiitica (National Institute for Statistics, Geography and Information) INIFAP Instituto Nacional de Investigaciones Forestaies y Agropecuarias (National Institute for Forestry, Agriculture, and Livestock Research) IU Irrigation Unit LMRC Long-run Marginal Cost NAFTA North American Free Trade Agreement NAFIN Nacional Financiera, SNC (National Finance Company) NPC Nominal Protection Coefficient NPV Net Present Value NTB Non-tariff Barrier OECD Organizaton for Economic Cooperation and Development OFMIN On-Farm and Minor Irrigation Networks Improvement Project (Loan 3704-ME) O&M Operations and Maintenance OP Other Producer PA Procuradurfa Agraria (Office of the Agrarian Attorney General) PBI Productor de Bajos Ingresos (Low-Income Producers) PECE Pacto de Estabilizaci6n y Crecimiento Econ6mico (Stabilization and Economic Growth Pact) PEMEX Petr61eos de Mexico (Mexican Petroleum Company) POA Programas Operativos Anuales (Annual Operating Programs) PROCAMPO Programa de Apoyos Directos al Campo (Farm Support Payments Program) PROCATI Programa de Capacitaci6n. Asistencia Tocnica e Investigaci6n (Agricultural Training. Technical Assistance and Research Project) PROCEDE Program a de Certificaci6n de Derechos Ejidales y Titulaci6n de Solares Urbanos (Ejidal Land Certification and Urban Parcel Titling Program) PRODERITH Programa de Desarrollo Rural Integral para el Tr6pico Humedo (program for Integrated Rural Development of the Humid Tropics) PRONASE Productora Nacional de Semillas (National Seed Production Company) PRONASOL Programa Nacional de Solidaridad (National Solidarity Program) PSE Producer Subsidy Equivalent PVP Plant Variety Protection RAN Registro Agrario Nacional (National Agrarian Registry) RI Restrictiveness Index RNVP National Registry of Plant Varieties (Registro Nacional de Variedades de Plantas) SAR Staff Appraisal Report SARH Secretarfa de Agricultura y Recursos Hidraulicos Secretariat of Agriculture and Water Resources SCT Secretarfa de Comunicaciones y Transporte (Secretariat of Communication and Transport) SECOFI Secretar(a de Comercio y Fomento Industrial (Secretariat of Trade and Industrial Development) SECOGEF Secretarfa de la Contralorfa General de la Federaci6n (Secretariat of the Comptroller General of the Federation) SEDESOL Secretarfa de Desarrollo Social (Secretariat of Social Development) .; SEDUE Secretarfa de Desarrollo Urbano y Ecologfa (Secretariat of Urban Development and the Environment) SDCSA Subdirecci6n General para la Conservaci6n de Suelo y Agua (Sub-Directorate for Soil and Water Conservation) SEPESCA Secretarfa de Pesca (Secretariat of Fisheries) SFOR Subsecretarfa Forestal (Forestry Subsecretariat of SARH) SHCP Secretarfa de Hacienda y Credito PUblico (Secretariat of Finance and Public Credit) SIEBAN Sistema de Estfmulos a la Banca y a los Productores Incentive System for Banks and Producers SNICS Servicio Nacional de Inspecci6n de Semillas (National Seed Inspection Service) SNIM Sistema Nacional de Informaci6n de Mercados (National System of Market Information) SPP Secretarfa de Programaci6n y Presupuesto (Secretariat of Budget and Planning) SRA Secretar(a de Reforma Agraria (Secretariat of Agrarian Reform) TABAMEX Tabacos de M~xico (Mexican Tobacco Company) TFP Total Factor Productivity TRI Trade Restrictiveness Index UCT Unidad de Cooperaci6n Tecno16gica (Technology Cooperation Unit) UNAM Universidad Nacional Aut6noma de Mexico (National Autonomous University of Mexico) UPOV International Convention for the Protection of New Varieties of Plants USLE Universal Soil Loss Equation VAT Value Added Tax WUO Water User Organization  MEXICO AGRICULTURAL SECTOR MEMORANDUM TABLE OF CONTENTS EXECUTIVE SUMMARY 0 0 0 0 0 0 0 0 0 • 0 0 • • 0 0 • 0 • • • 0 • • • • 0 • 0 • • • 0 • • • • • • • 0 • • • 0 • 0 0 • 0 i 10 Background.. 0 0 0 • 0 0 • • • 0 • • • • • • 0 • • • • 0 • 0 • • • • • • • • • • 0 • • • 0 0 • • • • • • • • •• i II. Chief Sectoral Issues • 0 • • • • • • • • • • • • ii • • • • • • • • • • • • • • 0 • • • • 0 • • • • 0 • • 0 • 0 •• A. Growth. Economic Incentives and Public Policy (Chapter 2) ii 0 • • • •• 0 • • • • • 0 • • 0 • 0 •• B. CoJDIDerCialization (Chapter 3) . . 0 • • •iv • • • • • • • • • • • • • • • • • • • • • • • 0 • • • • 0 • •• C. Irrigation and Water Markets (Chapter 4) . . . . . . . . . v 0 ••••••••• 0 0 •• 0 0 0 0 0 0 •• D. Agricultural Research. Intellectual Property and Extension (Chapter 5) . vii 0 0 0 0 • 0 • 0 •••• E. Rural Finance (Chapter 6) •.•. 0 ••••• viii 0 0 •••••••••• 0 • 0 •••••••••••••••• F. Fiscal Policies for the Rural Sector (Chapter 7) •.•• x 0 •••••••• 0 • • • • • • • • • • • • •• G. The Land Market (Chapter 8) . . • . • . . . . . . . . . . • . • . . . . . . . . . . . . . xiv 0 ••••••• H. The Tropics (Chapter 9) • . • • . . . • . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . .. xiv I. Soil Conservation (Chapter 10) . . . . . . . . xvii 0 0 • 0 ••••••••••••••••• 0 • • • • • •• 1. IN1'RODUCTION . . . . • . . • 0 • • • • 0 • 0 • • • • • • 0 • 0 • 0 • • • • • • • • • • • • • • • • • • • • • • • • •• 1 I. The Overall Economic Environment and Policy Reforms ... . . . . . . . . . . . . • . . . . . . . 0 3 II. Natural Resource Endowment . • . . . . . . . . . . . . . . . . • . . . 0 • • • • • • • • • • • • • • • • •• 6 m. History Structure of Mexican Agriculture . • . . . . • . . . . . • . . . . . . . . . . . . . . . • . . . . 8 2. GROWTH, ECONOMIC INCENTIVES, AND PUBUC POUCY . • . • • . . . . • . . . . . . . • • • . . ., 12 I. Decomposition of Growth of Output Value • . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 12 II. Nominal Protection, Transfers and Price Stability • . . . . . . . . • . . . . • . • . . . . . . . . . . . 15 m. Direct and Indirect Effects of Output Price Distortions . . . . . . • . . • . . • . . . . • . . . . • . . 21 IV. Other Transfers . . . . . . . • • . • . . . . • . . • . . . . . . . . . . . . • . . • . • . . . . . . . . . . • . 2S V. Effects of Policy Reforms on Profitability of Individual Commodities . . . • . . . . • . . . . . • . 27 VI. The Effects of NAFrA and PROCAMPO . . . . . • . . . . . . . . . . . . . . . . . . . . . . . • . . . 29 A. NAFrA . . . . . . . . . . . . . . . . . . . • . . . . • . . . . • . . . . . . . . . . . . . . . . . . . . . . 30 B. PROCAMPO . . . . • . . • . . . . . . . . . . • . • . • . . . 0 • • • • • • • • • • • • • • • • • • • • • • 33 VII. Conclusions and Recommendations . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . • . . . • 36 3. COMMERCIAUZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 I. The Proposed System of Commercialization . . . . • . . . . • . . • . . . . • • . • . . . . . . . . . . . 38 II. Risk ManageIIleIlt Options • . . . • • . . . . . • . • . . . . . . . . . . • . • . . . . • . . . . . . . . . . . 39 m. Mexico and Commodity Futurea Markets . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • 41 IV. Integration of Existing Markets . • . • . . • . • . . . . . • . . . . . . . • . . . . . . . • . . . . . . . . . 44 A. Market Integration in the Long Run . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 B. Quantifying the Long Run . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 V. Conclusions and Recommendations . . . . • . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . 55 The report wu prepared by John McImirc (Tut. Manaler, LAlNR), with lobo collabontioo of John Batre. ("ODOmotric" market price anaIylit), Silvia CuI.ro (nil'" fiJwlce), LoWIe Cord, (rural fileal policy), aod PIDOI Varan,eiI (market analyst); aod cotlllUltaDll Jorp Martinez. Julia RoitmaD aod BIeonon Oatcia (rutal fileal policy), John Barton (lelal aod replator)' f'ramework), aod K. William BIter (walAlr marketa). JOieph Rio, ~ilia Lim aod Sarah Crow provided ptoductioo report. Tbo DivilioD Chief ia Michael Bauor aod lobo COUJIlfy DopIUUDOIIl ~1Ot ia Bdilberto Septa. ChaplAlr Nil» ia baed OD !be "R.evirN of A,piculWraJ Develop_Ill Experience. aod Potential in the Tropic." (February 1994), prepared by a FAO/CP team led by Fraociaco ReifachDCidcr.  TABLE OF CONTENTS (Continued) Page No. 4. IRRIGATION AND WATER MARKETS . • . . • • . • . . . . . . . . . . . . . . . . . . . . . • . . • . . . . . . 56 I. Irrigated Agriculture . . • . . . . . . . . . . • . . • . . . . . . . . . . . . . . . . . • . . . • . . . . . . . 56 n. Problems of Irrigated Agriculture . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . • 57 A. Slowing Expansion of Irrigation . • . . • . . • . • • • . . • . . . . . . . . . . . . . . . . . . . . . . 57 B. Cost Recovery • . . . . . . • . . • • . . . . . • . . . . . . . . • . . • . . • . . . . . • • . . • . . . . . 59 m. Recent Policy Changes in Irrigation . . . . . • . . . . . . • • . • • . . • • . • • . . . . . . . . . . • . . 60 IV. Water Marlcets . . . • . . . . . . . . • . . . . . • . . . . . . . . . • . . • . . . . . • . . . . . . . . . . . . 61 A. Costs of Establishing Water Markets . . • . . . . • . . • . • • . . • • . . • . . . . • • . . . . . . . 62 B. Experience in Developing Countries . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . 63 C. New Institutional and Legal Ammgements in Mexico . . . . • . . • • . . • . • . • . . . . . . . . 65 D. Is the New Law Adequate to Creste an Efficient Water Market? . . . . . . . . . . . . . . . . . 66 V. Recommendations . . . . . . . . . . . . . . . . . . • . . • . . . . . . • . . • . . • • . . • . . . . . . . . . 67 5. RESEARCH, INTEll.ECTUAL PROPERTY AND EXTENSION . • . . . . . . . . . • . • . . . . . . . . .• 69 I. Research . . . • . . . . . . . . . • . . . . . . . • . . • • . • . . • • • . . . . . . . • . . . . . . . . • . . . 69 n. Intellectual Property Issues . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . • . . . . . . . . . • 76 A. Seed Policy . . • • • • . . . . . • . . . • . . . . . . • . . • . . . • . . • . . • . . . . • . . • . . . . . . 76 B. The Regular Patent System . . . . . . . . . . • . . • . . . • . . . • . . . . . . . • . . • • . . • . . • 78 m. Extension •.. . • . • • . . . . . . . . . . . . . . • . . • . . • . . . . . . . . . . . . . . . . . . . . . • . . 82 IV. Recommendations .. . . . • . . . • . . • • . . . . . . . . . . . . . . . . • • . . . . . • . . • • . • . . • • 85 A. Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , 85 B. Intellectual Property . . . . . • . . . . . . . . . . . . . . . . . . • . . • . . . . . . . . . . . . . . . • 86 C. Reforms in Extension • . • . • . . • . . • . . • • . . . . . . . • . . . . . . . . . . . . • . . • • . . • . 87 6. RURAL FINANCE. . . . . • . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . 88 I. Background.................................................... 88 n. Agricultural Financing . . . . . • . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . • . . • . . . . 89 m. The Cost of Rural Finance. . . . • • . . • . . . . . . . • . . • . . . . . . • • . . • . . . . . . • . . . . . 91 A. The Cost of Government Rural Finance Institutions . . . . . . . . . . . . . . . . . . . • . . . . . 93 B. Estimated Loss in Value of the GOM Transfers. . . . . . • . . . . • . . . . . . . . . . . . • . . • 99 C. Benefits Accrued to Borrowers of Government Funds • . . . . . . . . . . . . . . . . . . . . .. 100 IV. Rural Finance during the Salinas Administration . . . • . . • . . . . . . • . . • . . . . • • . . . . . 101 A. Streamlining the Responsibilities of Development Banks and Trust Funds . . . • . . . . . .. 101 B. Elimination of Financial Subsidies . . . . • . . • . . • • . . • . • . . • • . . . . . . . . . . . • .. 104 C. Other Reforms . • . . . . • . . . . . . . . . . • . . • . . . • . . . . . . . . . . . . . . . . . . . . .• 105 D. Nonperforming Aaricultural Loans . . • . . • • . . . . . • . . . . . . . . . . . • . . . • . . • . . 106 V. Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 108 7. FISCAL POUCIES FOR THE RURAL SEcrOR . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . .. 111 I. Evolution of Rural Development Expenditures . . . . . . . . . . . . . . . . . . . . . . . . • . . . . 111 A. Federal Expenditures . . . . • . . • . . • • . . • . . . . . . . . . . . • . . . . . • . . . . . . • . .• 111 1. Level of Expenditures • . . • . . • . . . . . . . . . . . • . . . . . . . . . • . . • . . . • . . • . 112 2. Composition of Expenditures • . . . . . . . . . . . . . . . . . . • . . • . . . • . . • . . . . .. 115 B. State and Municipal Expenditures . . • . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . 118 n. Evaluation of Rural Expenditure Policy . • . . . . . • . . • • . • • . . • . . . • . . • . . . . . . . .. 121 A. Efficiency in Rural Expenditures. . • . • . • • . . . . . . • . . . . . . • . . . • . . • . . . . . .. 122 1. Expenditure Strategy . . . • • . . . . . . . . . . . . . . . . • . . . . . . . . . • . . . . . . . .. 122 2. Decentralization of Rural DeVelopment Finance . . . . . . . . . • . . • . . . . . . . . . . . 125 3. Multi-Year Capital Budgeting. . . . . . . . . . . . . • . . . . . . • . . . . . . . . . . . . . .• 128 4. The Budget Process . . . . . . . • . . . • . . • . . • . . . . . . . . . . . . . . . . . . . . . . .. 128 TABLE OF CONTENTS (Continued) Page No. B. Public Experulitures and Poverty Alleviation. . . . . • . . . . . . . . . . . . • . • . . . . . . .. 129 1. Revenue Sharing . • . . . . . . . • . . . . • . • . . . . . . . . . . . . . . . • . . . . . . . . . • 130 2. Solidarity . . . . • . . • . . . . • . . . . • . • . • . . . . • . • . • . . • . • • • . . • • . • . . .. 131 3. Distribution of Marketing (CONASUPO) and Credit (DFIs) Subsidies . . . . . . . . . .. 132 4. SARB . . . . . . . . . . • . . . . . . . . . . . . • . • . . • . • . . . . . . . • . . . . . . . . . •. 133 m. Fiscal System and Incentives for Rural Development Investments . . . . . . • . . . . . . . . . .. 133 A. Introduction . . • • . . • . • . . . . • . . • . . • . . . . . . • . . . . . . . • . . . . . . . . . . . . . 133 B. The Incotne Tax • . . . . . . • . . • . • . . . . . . . • . . . . • . . • . . . . . . . . . . . . . . • .. 135 C. The Gross Asset Tax • . . . . . . . • . • • • . . . . . . . • . . • . . . • . . . . • • . • . . . . • .. 136 D. The Value Added Tax . • • . . . . . . • • • . . . . . • . • . . . . . . . . . . . . • . . . . . . . • . 137 E. Property Taxation • . • . . . . . . . . • . • • • • . • . • • . • . . • . . • . . . . . . . . . . . . • .• 138 F. Social Security and Payroll Taxes • . . • . . . . . • . . . • . . • . . • . • . • . . . . . . . . . . . 138 IV. Recommendations and Conclusions . . • . • . . . . • • . . . . • . • . . • . . . . • . . . . . . • . . . 139 A. Trends in Rural Development Finance •••.•......•.•..........•....... 139 B. Improving the Efficiency of Rural Development Expenditures . . . . . . . . . . . . . . . . •. 140 C. Public Expenditures and Poverty Alleviation. • . . . . . • . . . . • . . . . . . . . . . . . . . .. 143 D. Impact of Taxation on Incentives for Private Investments in Rural Development. . . . . .. 144 8. TIlE LAND MARKET . . . . . . . • . . . . . . . . . . . . • • . • . • . . . . . . . • . • . . • . . . . . . . • . . 146 I. The Old System . . . . • . . • . . . . • . . . . • . . . . . . . • . . . . . . . . . . . . . . . . . . . . .. 147 A. The Legal Basis . . . . . • . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 147 B. Operational Restrictions on Land Use • . . . . . • . . • . . . • . . . . . • . • . . . . • . • . . .. 147 C. Productivity Effects . . • . • . . . . . . . • . . . . • . . . . . . . . . . . . . . . . . . . • . . . . .. 148 II. The Recent Reforms . . . . . . . . . . . • . . • . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . 149 A. The Legal Reforms . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . • . . . .. 149 B. The Land Titling Program. . . . • . . . . . . . . . • • . • . • . . . . . . . . • . • . . . . . . . .. 150 C. Costs and Benefits of Land Titling . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . .. ISO m. Conclusions and Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 9. TIlE TROPICS . . • . . . . . . . . . . . . • . . . . . . . . . . . . • . • . . • . . . . . . . . . . . . . . . . . . .• 152 I. Physical Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 II. Principal Conclusions about Comparative Advantage. . . . . . . . . . . . . . . . . . • . • . . . .. 153 m. Limiting Factors . . . . . . . . . . . • . • . . • . . . . . • . . . • . . • • . . . . . . . • • . • . . . . ., 156 IV. Recommendations . . . . . . . • . . • . • • • • . • . . . . . . . . . • . . • . • . . . • . • . . . . • . .. 159 10. SOIL CONSERVATION. . . . . . . . . . . . • . . . . . . • . . • . . . • . . . . . • . • . • . . . . . . . . . .. 164 I. General Patterns of Land Degradation Caused by Erosion . . . . • . . . . . . . . . . . . . . . . . 164 II. Patterns of Agricultural Soil Erosion . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . 166 m.Productivity Impact of Erosion . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 168 IV. Economic Analysis of Soil Conservation . . . . . . . . . . . . . . . . . . . • . . . . • . . . . . . .• 169 A. Soil Conservation Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . " 169 B. Effects of Soil Conservation Techniques . . . . . . . . . . • . • . . . . . . . . . . . . . . . . . . 170 C. Cost-Benefit Analysis . . • . • . . • . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . .. 170 1. Methodological and Data Issues ••••••••••••.••.•..••••••.•••.•..• 171 2. Plan of the Analysis . . . . . . • . . • . • • . • . • . . • . . . . • . . . . . . . . . . . . . . . . 172 3. Costs of Erosion Control Techniques . . . . . • . • . . . . . . • . . . . . . . . . . • . . . .. 174 TABLE OF CONTENTS (Continued) Page No. 4. Results for the Tropics . . . . • . • . . • . . . . • . . . . . . . . . . . . . . • . . . . . . • . .. 176 S. Results for the Central Rainfed Areas • • • • . • • . • • • . • • . • • • • . • • • • • • • . . .• 176 6. Sensitivity Analysis • . . . . . . . . . . • • . . . . . • . . . • . • • . . • . • . . . . . • • . . •. 178 D. Is a Land Reserve More Efficient Than Erosion Control Practices? • • • . • . . • • . • . . •. 180 E. Comparison to Other Results .•.. • . . • . . • . . • • . • • • • • . • • • • • . . . . . • . . • .. 181 1. Estimates of the Benefits of Erosion Control in Mexico . . . . . • • • . • . . • . . • . . .. 182 V. Policy and Institutional Issues . • . . . . . . • • . • • . . • . . . . . • • . • • • • . . . . . . . . . • .. 183 A. Soil Conservation Policies . . . . • . . . . . . . . . . . . . . . . . . . . . • . . • . . . . . • . . .. 183 B. Current Soil Conservation Policy . . . . . • . • . . . • . • . . . . . . . • . • . . . . . . . . . • . 184 C. Why have farmers not adopted better soil conservation practices? . . . . . . • . . • . • • . .. 18S D. Are Public Erosion Control Measures Iustified? . . . . . • . • . . • . . • . . . . . . . • . . . . 187  Index of Tables, Figures, Boxes and Annexes Tables Table 1.1 Evolution of the Mexican Economy: 1950-92 . . . . . . . . . . . . . . . . . . . . . . .. 4 Table 1.2 Potential Land Use, Actual Cropped Area, and Value of Output, by Region. . . .. 7 Table 1.3 Total, Agriculture, and Nonagriculture GDP: 1950-1992 . . . . . . . . . . . . . . . .. 9 Table 1.4 Annual Average Agricultural GDP Growth and Labor Share . . . . . . . . . . . . . . 10 Table 2.1.A Sources of Change in the Value of Output for Maize, Wheat, Sorghum Barley, Rice, Soybeans, and Safflower: 1950-92 . . . . . . . . . . . . . . . . . . . . . . 13 Table 2.1.B Sources of Change in the Value of Output for Cotton, Beans, Chickpea Sugar, Coffee, Cocoa, and Alfalfa: 1950-92 . . . . . . . . . . . . . . . . . . . . . . . . 14 Table 2.2 Domestic Price Decomposition for Five Crops . . . . . . . . . . . . . . . . . . . . . . . 16 Table 2.3 Nominal Protection Coefficients for Nine Major Crops . . . . . . . . . . . . . . . . . . 17 Table 2.4 Potential Transfers to Agricultural Producers for Nine Major Crops . . . . . . . . . . 20 Table 2.5 Ratio of Domestic to International Price Variability . . . . . . . . . . . . . . . . . . . . 21 Table 2.6 Restrictiveness Indices for Five Major Crops . . . . . . . . . . . . . . . . . . . . . . . . 24 Table 2.7 World Prices, NPCs, and Transfers to Agriculture: Urea and Phosphate Fertilizers, 1965-92 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Table 2.8 Input and Output Price Elasticities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Table 2.9 Changes in Inputs and Outputs In Three Scenarios . . . . . . . . . . . . . . . . . . . . . 29 Table 2.10 Period Averages of Agricultural Exports . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Table 2.11 Percent of 1990 Exports Affected by NAFT A . . . . . . . . . . . . . . . . . . . . . . . 31 Table 2.12 Pre- and Post-NAFTA Trade Policy for 12 Major Crops . . . . . . . . . . . . . . . . . 31 Table 2.13 Welfare Effects and Efficiency Gains from Maize Price Liberalization . . . . . . . . 33 Table 2.14 Savings from Alternative PROCAMPO Payment PoHcies . . . . . . . . . . . . . . . . 35 Table 3.1. Long-Run Market Integration Results. . . . . . . . . . . . . . . . . . . . . . . . . . .. 50 Table 3.2 Number of Weeks Required for 90 percent of the Adjustment. . . . . . . . . . . .. 51 Table 3.3 Percent of Adjustment Taking Place within two Weeks. . . . . . . . . . . . . . . .. 52 Table 3.4 Average Percentages of Adjustment . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 54 Table 4.1 Planted Areas with Irrigation and Public Spending on Irrigation, 1950-1992 . . . . . 58 Table 5.1 Agricultural Research Scientists and Expenditures, 1982 - 1992 . . . . . . . . . . . . 70 Table 5.2 Value of Crops and Needed Federal Research Spending for Central and Southern States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 Table 5.3 Extension Agencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Table 5.4. Extension Service Reforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 Table 6.1 Cost to Government of Rural Financial System . . . . . . . . . . . . . . . . . . . . . . . 93 Table 7.1. Government Expenditures for Agriculture in Industrial and Developing Countries 112 Table 7.2 Summary of Rural Development Expenditures by Institution. . . . . . . . . . . . .. 117 Table 7.3 Expenditures and Revenue Sharing for Aguascalientes . . . . . . . . . . . . . . . . . 120 Table 7.4 Growth in Municipal Expenditures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 Table 9.1 Main Characteristics of Four Tropical States . . . . . . . . . . . . . . . . . . . . . .. 153 Table 9.2 Land Use and Tenure in four Tropical States . . . . . . . . . . . . . . . . . . . . . .. 154 Table 9.3 Current and Expected Profitability of Crops in Four Tropical States . . . . . . . .. 155 Table 10.1 Land Degradation Caused by Soil Erosion from Wind and Water . . . . . . . . . . 165 Table 10.2 Severity of Soil Erosion in 1982 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 167 Table 10.3 Soil Conservation Techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 171 Table 10.4 Costs of Soil Conservation Techniques and Soil Erodibility Indices . . . . . . . . . 173 Table 10.5 Values of Key Parameters for Economic Analysis . . . . . . . . . . . . . . . . . . .. 175 Table 10.6 Net Economic Benefits of Soil Conservation Techniques at Social Discount Rate of 12 % . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 177 Table 10.7 Effect of Discount Rate on Private Profitability at Target Soil Loss of 5 mt ... , 178 Table 10.8 Hypothetical Subsidies and Extension Costs to Achieve Annual Erosion Loss of 1 mt per Hectare with Group A" or fiB" Techniques. . . . . . . . . . . .. If 180 Table 10.9 Net Benefits per Ton of Soil Saved by Soil Conservation Practices . . . . . . . . . 181 Table 10.10 Public Activities in Soil Conservation, 1950-1982 . . . . . . . . . . . . . . . . . . .. 184 Figures Figure 2.1.a NPCs for Maize, Wheat, and Barley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Figure 2.1. b NPCs for Sorghum, Rice, and Soybeans . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Figure 2.1.c. NPCs for Sugar, Coffee, and Cocoa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Figure 3.1.a MiHed Rice -- Prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Figure 3.1.b Milled Rice -- Price Differential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Figure 3.2.a Traditional Beans -- Prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Figure 3.2.b Traditional Beans -- Price Differential . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Figure 3.3.a White Beans -- Prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Figure 3.3.b White Beans -- Price Differential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Figure 3.4.a Adjustment Paths for Ecatepec-Guadalajara . . . . . . . . . . . . . . . . . . . . . . . . . 53 Figure 3.4.b Adjustment Paths for Federal District-Ecatepec . . . . . . . . . . . . . . . . . . . . . . . 53 Figure 6.1 Real Increase in Outstanding Credit to the Private Sector . . . . . . . . . . . . . . . . 90 Figure 6.2 Agricultural Financing: Outstanding Agricultural Portfolio . . . . . . . . . . . . . . . . 90 Figure 6.3 Flow of Funds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 Figure 6.4 Government Support to the Rural Financial System ... . . . . . . . . . . . . . . . . . 93 Figure 6.5 Government Cost per Dollar Lent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Figure 7.1 Federal and Rural Development Expenditures. . . . . . . . . . . . . . . . . . . . . .. 113 Figure 7.2 Summary of Rural Development Expenditures by Institution . . . . . . . . . . . . .. 114 Figure 7.3 Rural Development Expenditure by Program . . . . . . . . . . . . . . . . . . . . . . . 116 Figure 7.4 Rural Development Expenditure by Category . . . . . . . . . . . . . . . . . . . . . .. 119 Boxes Box 3.1 How Integrated are Mexican and US Grain Markets? . . . . . . . . . . . .. . . . . . . 38 Box 7.1 The Planning Process . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . .. 124 Box 7.2 The Role of COPLADE . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....... , 125 Box 7.3 Revenue-Sharing. . . . . . . . . . .. . . . . .. . . . . . . . . . . . . . . . . . . . . . .. 131 Annexes Annex 1: Crop Production and Price Data Annex 2: Fiscal Costs of Rural Financial Institutions Annex 3: Fiscal Data for Rural Development Appendix 1: Rural Development Expenditures by Institution Appendix 2: Rural Development Expenditure Data Appendix 3: Regression Results for the Distribution by State of Solidarity, Revenue Sharing and SARH Expenditures Annex 4: Integration of Mexican Markets Annex 5: Soil Conservation Analytics  MEXICO AGRICULTURAL SECTOR MEMORANDUM EXECUTIVE SUMMARY 1. Agriculture is the weakest sector of the Mexican economy . Average productivity per worker is less than half what it is in other sectors, incomes are correspondingly low, and growth has been weak for more than 20 years. Recent refonns - in the price setting mechanism, the role of government, the liberty of enterpdse, and farmers' rights over land and water - have so far failed to improve productivity and to revive output growth equal to that of population or of the whole economy. What has happened and what should be done? I. Background 2. In the early 1990s, the agricultural Gross Domestic Product (GOP) accounted for 7-8 percent oftotal GOP (Table 1.3). t The share had fallen steadily from 18 percent in 1950. The growth of the sector has also declined. While exhibiting an average annual growth of 5 percent in the early 196Os, it dropped below 3 percent in the period 1965-70. After a partial recovery in the 1971-76 period, it dropped to below 1 percent from 1977 to 1982 and has recovered only weakly since. I Within agriculture, subsectoral shares have changed little in the past 40 years. Crops typically provide about 60 percent of the sector's output, livestock about one-third, and forestry, fishing, and hunting the rest. Within the crop subsector in 1992, cereals accounted for 48 percent, oilseeds for 5 \ percent, fruits, vegetables, and legumes for 28 percent, sugar cane for 5 percent, fibers for 4 percent, and coffee and cocoa for 6 percent. \J 3. There is an irrigated and commercial class of farmers using modern technology, achieving high yields, competing on international markets, and earning high incomes. There is a class of rainfed farmers, who produce mainly for subsistence, achieve yields well below those with irrigation, do not generally compete on international markets, and earn subsistence incomes, in many instances by working as seasonal laborers outside their own farms. 4. There are sharp regional differences in patterns of output. Irrigated crops provide 54 percent of the national value of output, but are 80 percent of regional value in the north, 50 percent of regional value in the center, and 10 percent in the south. The north grows nearly all the wheat, and most of the temperate vegetables, oiJseeds, cotton, and irrigated maize. The center grows much of the rainfed maize, beans, and barley and irrigated crops (sorghum, maize, vegetables) are found in the high valley known as the Bajio. Rainfed maize, perennial crops (sugarcane, coffee, cocoa, banana, and others) and livestock dominate in the south. 5. The states of Baja California, Sinaloa, Chihuahua, Durango, Coahuila, and Tamaulipas, comprising most of what is the Northwest, North, and Northeast (Table 1.2), have about 3.1 million hectares under irrigation, or about 57 percent of the national irrigated area. They produce roughly one-third of the value of annual and perennial crop output (Mex$18.0 billion of all output, rain fed and irrigated combined). The value of crop output per hectare with irrigation is approximately thrice 1. All references to Tables in the Executive Summary are to Tables in the main text.  ii what it is under rainfed conditions. 6. The southern states, consisting chiefly of the Center Gulf, Pacific South and Peninsula regions (Table 1.2), have about 5 million hectares under cultivation, of which only 0.28 million have irrigation. The irrigated area is just five percent of the national irrigated area. The value of output per hectare is about N$2,400. The southern states use fewer chemical, mechanical and biological inputs per hectare and receive less public agricultural credit and public investments (see Chapter 6 on rural finance and Chapter 7 on public expenditure). The specific agricultural development problems of four southern, largely tropical, states - Oaxaca, Veracruz, Tabasco, and Chiapas - are discussed in Chapter 9. II. Chier Sectoral Issues 7. The chief sectoral issues related to growth are: a. Growth, Economic Incentives and Public Policy (Chapter 2) b. Commercialization (Chapter 3) c. Irrigation and Water Markets (Chapter 4) d. Agricultural Research, Intellectual Property, and Extension (Chapter 5) e. Rural Finance (Chapter 6) f. Fiscal Policies for the Rural Sector (Chapter 7) g. The Land Market (Chapter 8) h. The Tropics (Chapter 9) i. Soil Conservation (Chapter 10) A. Growth, Economic Incentives and Public Policy (Chapter 2) 8. Reforms in economic incentives have been profound. The Government has, since 1985: (a) joined the organization for the General Agreement on Trade and Tariffs (GATT), the North American Free Trade Agreement (NAFTA), and the Organization for Economic Cooperation and Development (OECD); (b) eliminated non-tariff barriers to agricultural trade, except those on maize and beans; (c) eliminated guarantee prices on 10 of 12 crops that had them in 1989 (exceptions are again maize and beans); (d) stopped fertilizer subsidies and privatized the state fertilizer distribution I monopoly; (e) ended the role of the national basic foods company (CONASUPO) in the marketing of grains except for maize and beans; (t) shifted some general food subsidies to targeted ones; and (g) begun the National Farm Modernization Program (pROCAMPO), a program of trade liberalization and farm income support that will end remaining price distortions in agriculture (chiefly those affecting maize, wheat, and beans) and provide transitional income support to producers that does not interfere with markets. 9. The effects of these reforms on growth, income distribution, and the level of private sector activity have been mixed. Growth. The rate of growth may have been affected by the reforms; it was 3.2 percent from 1976-82, 0.5 percent from 1983-88, and 0.8 percent from 1989-92. There is some evidence, presented in Chapter 3, that the profitability of some major commodities fell after 1987, when many reforms were taking place.  iii b. Income distribution. The price policy changes after 1990 - high guarantee price of maize, direct payments to wheat producers, lower producer prices for coffee and cocoa - have tended to favor irrigated farming against rainfed and tropical, thereby widening the existing disparities.' c. Private sector activity. Recent reforms permit a broader role for the private sector. A parastatal, the Mexican Fertilizer Company (FERTlMEX), formerly distributed much of the fertilizer used in Mexico, and this activity has been transferred to the private sector. CONASUPO used to handle much of wheat and sorghum harvest and that is now handled by the private sector. Much of CONASUPO's retail network has been privatized. The new Seed Law (July, 1991) has begun to promote private sector seed development (see Chapter 5). The new Intellectual Property Law should promote private technology generation and transfer more broadly throughout the sector. 10. Recommendations to improve growth a. Reduce the cost of PROCAMPO. The cost of PROCAMPO is too high and will cause permanent funding problems for other appropriate public investments, notably in irrigation and agricultural research, unless it is reduced. To reduce the cost, the Government of Mexico (GOM) could: i. give a smaller payment per hectare beginning in the definitive phase of PROCAMPO on October I, 1994; 1 ii. purchase the rights to the full value of PROCAMPO payments at a steep discount from eligible producers, thereby reducing the present value of the payments over the 15 year period; iii. tax PROCAMPO payments to fund public agricultural research and the capit costs of irrigation and other infrastructure investment. b. Privatize the commodity marketing functions of CONASUPO unrelated to the distribution of subsidized food. c. Replace all general consumer food subsidies with subsidies targeted to the poor, as CONASUPO (maize) and LICONSA (milk) are now doing; and d. Eliminate subsidies to large-scale maize millers. e. Attempts to establish domestic futures markets in farm commodities should be stopped because Mexican futures traders can use markets in the United States; f. Interventions in specific commodity markets should be limited even if they appear to be badly integrated (e.g., as shown in Chapter 3). Some of the poor integration is the legacy of CONASUPO; the necessary measure to end CONASUPO's influence is to liquidate its commercial functions. Better market integration will require: the end of 2. This conclusion is based on the gross value of output, not the net one, and would be affected by subtracting the increased cost of irrigation water to producers resulting from reforms in cost-recovery and higher electricity rates to irrigation.  iv guarantee prices; the end of general consumer price intervention; cheaper information3 and cheaper financial intermediation; and g. Interventions to stop the decline in maize production also should be discouraged. All the relevant studies related to NAFTA predict a fall in maize production and in land and labor allocated to maize. While those predictions are probably exaggerated, because they are based on the assumption of no technical change in the supply response, their general tendency is true. The GOM has done the right thing - prepared a transitional income support program (pROCAMPO) - but it should go no further in supporting maize production. B. Commercialization (Chapter 3) 11. The Government is changing the commercialization of major grains, oil seeds and beans. Previously, the Government owned agency of CONASUPO was heavily involved in purchasing storing and reselling a substantial part of the commodities. Under the proposed system, most of the commercialization activities will be handled by the private sector. The main strengths of the proposed system are: (a) non-distortionary production and consumption incentives structure; (b) significant reduction of the government's role in the marketing of agricultural products; (c) opening up of the trade regime; and (d) more transparent determination of agricultural prices. 12. The proposed system has provisions for risk management through the establishment of formal forward markets. These markets, along with the evolution of informal forward markets (after the proposed system starts functioning) and the use of US futures markets, may provide adequate risk sharing for market agents. 13. With regard to futures markets, the spot markets should be given some time to function so as to allow the private sector to determine the appropriateness of domestic futures markets. It is not clear that the Mexican government has a comparative advantage in initiating or managing futures markets. The government can facilitate and improve links between Mexican and US futures prices by removing trade barriers and infrastructure bottlenecks in commercialization (e.g., storage, communications, transportation, and finance). 14. The main recommendations are to: Not create a Commodity Futures Exchange. If there is a need for such an exchange, the private sector will demand it and subsequently will undertake its creation and support its operation. b. To improve infrastructure for spot markets, because they are vital for the new commercialization scheme. c. To educate farmers in how prices are determined in a competitive environment and the difficulties of the new marketing system. 3. The growth in telephone service is an important element in this.  v d. Monitor and supervise the spot markets in order to avert the development of local monopolies, but the Government should not intervene. C. Irrigation and Water Markets (Chapter 4) 15. Much of the weak growth of the sector can be attributed to the slackening expansion of -\ irrigated agriculture. It is unlikely that irrigated area can be expanded as rapidly as it has in the past ~~, (e.g., in the 1950s and 196Os) because the cheapest surface water sites (mainly in the north) are already in use, because of the rising economic and environmental costs of groundwater, and because of the imminent strong competition from non-agricultural uses of water. Future gains from irrigated agriculture have to come from greater water use efficiency. 16. The Government has made bold reforms affecting irrigated agriculture. Reforms have been to: (a) begin to raise the cost of water (investment and operations) to producers in public irrigation districts (IDs) through aggressive cost-recovery, a measure that will eventually improve production efficiency; (b) begin to transfer IDs to water users organizations (WUOs); (c) apply stricter economic criteria to investments, both new and rehabilitation of existing structures; (d) apply stricter environmental criteria to those investments; (e) emphasize rehabilitation and consolidation of existing irrigation facilities; (f) change the national land tenure law to facilitate outside investment, notably contract farming; and (g) change the national water law and implementing regulations to permit the development of water markets. Reforms in agricultural pricing in 1993-94 (and the abolition of government controls on the cropping pattern in 1992) will also help to make irrigated agriculture more efficient. The eventual elimination of producer subsidies for maize, wheat and beans will make them uncompetitive as irrigated crops, and thereby raise the relative returns to other crops, mainly fruits, vegetables, and forages that are more economically efficient with irrigation. 17. Despite the nature and extent of policy reforms affecting irrigated agriculture, and despite the benefits of NAFrA in improved market access for the products of irrigated agriculture, the outcome of the reforms is inherently limited in its effects on sectoral output. Though the reforms should make irrigated agriculture more efficient, chiefly in the allocative sense and in the technical sense with respect to water, the physical limitations on the new area with irrigation and the rising competition for water from non-farm uses will mean that efficiency gains will not have major effects on the sector growth rate. A further barrier to growth from irrigation is the low efficiency of some exportable crops, despite much lower wages in Mexico than in the United States, related to a lack of competitive pressure in the past, to low water prices, and to inadequate research. 18. Recommendations to bolster efficiency in irrigated agriculture are: a. Investment planning. The National Water Commission's (CNA's) investment planning continues to be hampered by uncertainty about the budget. One strategy to overcome this is to work with an indicative five-year budget with specific annual targets; more generally, ways to encourage rational investment planning and limit political interference and improve bureaucratic handling of budget issues need to be identified. b. Cost recovery. Impose full cost recovery on old and new projects alike. There is no] good equity reason to charge higher prices to new users who are likely to live in previously under-provided areas, and to charge less to old users who have resided in areas that historically have been better provided.  vi Electricity tariff. Increase the farm electricity tariff to the long-run marginal cost, thereby stopping the over-pumping of groundwater. d. Role of the WUOs. Expand the role of the WUOs to fund research on irrigated agriculture as part of the privatization of the agricultural research system in the north (see paras. 20.c and 20.d). e. Let water markets work. Because of infrastructure costs, imperfect information, and third party effects, markets cannot solve all the water allocation problems in Mexico. However, if the necessary investments - notably in water metering infrastructure by the private sector and information to producers about the law by the public and private sectors - are made, water markets can improve the efficiency of water use and allocation. CNA's role should be to provide overall water resources planning and information and to facilitate decentralized water management through markets and water users organizations (WUOs). i. Transfer the administration of the water law to the WUOs, thereby eliminating the role of CNA except as it affects water trades involving non-farm users. ii. Make registration the only requirement for water trades within IDs. The Government should not block reallocations of water to new crops (e.g., away from basic cereals and oil seeds to fruits and vegetables). iii. Devise a quick and public approval process for trades among and outside IDs and urban areas. iv. Create institutions to distribute water among IDs and urban areas, such as joint companies. v. Allow individual water users to trade water outside the IDs as long as there are no significant third party effects. The government should not interfere with water trades that reduce cropland as these are practically inevitable. vi. Devise a public process for allocating water from surplus concessions and during scarcities. In the case of surpluses, sales by open bids, as suggested by the water law, should be used where possible. During droughts, though human and urban uses have priority over farming, allocation rules still need to be developed for agricultural water. These rules must be created in consultation with users before drought. The critical elements are full participation by water users and allocation rules that fit the cropping and other conditions in the region. f. Allow other reforms to work on irrigated agriculture. Inappropriate "compensatory" programs could limit the effectiveness of reforms affecting irrigated agriculture. Since possibil ities include: (a) the delayed introduction of the agreed PROCAMPO program and pressure under it for larger payments, on the grounds that irrigated farmers have lost the most in absolute terms from free trade in farm products; (b) special credit programs; and (c) emergency relief on the grounds of exceptional market conditions, such as unexpectedly large food grain imports.  vii D. Agricultural Research, Intellectual Property and Extension (Chapter S) 19. Agricultural Research. Limited resources have had very serious negative effects on the public agricultural research system. While greater private (including foreign) research will occur with NAFfA and perhaps with tighter protection of plant varieties and other intellectual property, private research can never be a full substitute for public research. To fulfill its role, the public agricultural research system needs more resources. Whatever its level of resources, however, the system must become more efficient. The major changes in relative prices (fertilizer in 1987-1990), maize, barley, and beans (1994-95), water (after 1991), electricity (from May, 1990 until January, 1993) can only do so much to promote better efficiency in resource use on farm. Existing institutions - the National Institute for Forestry, Agriculture, and Livestock Research (lNIF AP), CNA, the private sector and producers' organizations - still must generate new techniques to exploit Mexico's comparative advantage in labor-intensive methods. 20. The main recommendations are. a. Develop a strategic plan with specific long-term funding targets. b. Maintain those targets. c. Give priority to the rainfed and tropical areas and privatize most of the research facilities in the north, if those funding targets cannot provide for reasonable programs in the irrigated and rainfed areas.' d. Grant seed money to a few (new or existing) private institutions that will take over the privatized research facilities in the north. This seed money would be granted: i. competitively; ii. to national or foreign institutions; iii. for five years, after which time the privatized institutions would be independent of core federal support, but could still compete for grants. The institutions would then rely on state governments, producers organizations, other donations, and their own commercial activities (e.g., consulting and farming). e. Do the same for the research activities of the former parastatals in coffee, sugarcane, and cocoa, irrespective of what is done with the research facil ities of the north. 21. Intellectual property. The protection of intellectual property in Mexico is now much stronger with recent legislative changes. Remaining desirable reforms would be to: a. adjust INIFAP's seed distribution policies to plant variety protection (pVP), when the latter becomes law; b. clarify the rights of INIF AP breeders under PVP; and 4. The privatization contracts would need to include clauses to maintain seasonal research sites for public plant breeding programs targeted to the south and center, however.  viii c. change intellectual property laws to permit protection of genetic engineering with natural and modified genes. 22. Extension. Limited resources have also seriously restrained the effectiveness of the public extension service. Although the trend towards private technical assistance is desirable, it should be put in context of the particular needs of the different resources and experiences of producers. As with research, a strategic plan should be prepared for extension. This would include: a. Assisting small producers with commercial potential through Government-financed extension that is provided by private agents, including that supplied by producers' organizations. The approach of the Trust Fund for Shared Risk (FIR CO) and the Trust Fund for Agriculture (FIRA) - in which extension is subsidized on a declining basis for a determinate period - has succeeded. However, because it is tied to the availability of investments (FIRCO) or commercial credit (FlRA), it does not apply to groups that do not receive such investments or who are marginally creditworthy. b. Obliging commercial producers to pay for all technical assistance. c. Restrict the direct role for public extension to training and accreditation of private extension agents, apart from the financing role. E. Rural Finance (Chapter 6) 23. The cost effectiveness of public rural credit and the permanence of rural financial institutions are doubtful. The development of effective financial markets serving small farmers requires several changes in public policies. Remaining credit subsidies to poor farmers should be better targeted 24. The rural credit reforms of the late 1980s had the explicit objective of redUCing the cost to Government of the rural financial system. This was accomplished. 25. Much of the remaining public support to rural credit institutions does not yet accrue to small farmers with viable investments. The segmentation of rural borrowers into the clients of three different agencies--the very poor with little or no commercial potential for the National Solidarity Program (pRONASOL), the poor with commercial potential for the National Rural Bank (BANRURAL), and all others for the Trust Funds for Agriculture (FlRA)-- made the system more transparent. But Government funds channelled through FIRA and BANRURAL now reach fewer poor producers and accrue disproportionately to the better-off farmers, particularly in the case of FlRA. The share of low-income producer discounts in FIRA's annual lending decreased from about 50 percent in 1988 to 18 percent in 1991. Reportedly, BANRURAL has been unable to serve its new target population. With reduced government transfers, the need to increase loan recoveries accentuated, and management moved to serve those larger producers able to provide collateral. Ensure the institutional self-sufficiency of FIRA and BANRURAL through financial self- sufficiency. 26. FIRA needs to become financially self-sufficient, except for a portion of its costs of reimbursing transaction costs to commercial banks and for technical assistance. To this end, FlRA  ix should: a. discount subloans at a rate equal to its cost of funds plus a spread to cover its costs of financial intermediation. To this end, FJRA needs to introduce cost accounting procedures to establish the cost of its different services; b. discontinue its complementary loan guarantee program, as this program has not been] successful in increasing commercial bank lending to PBIs. c. simplify its interest rate structure and, while interest rates remain controlled, reduce the interest rate subsidy and its rediscounting margins to encourage commercial bank financing; 27. Moreover, to better serve the needs of rural borrowers, FIRA should reorient its emphasis on credit and technical assistance, away from basic grains toward a broader range of activities, and favor discounting investment loans over short-term loans. 28. To induce increased commercial bank lending for low-income farmers, FIRA should continue reimbursing banks for the transaction costs of bringing new, low-income farmers into the system. The reason for this is that commercial banks lack adequate incentives to lend to poor producers due to the high transaction costs of lending to small borrowers with risky projects. However, FIRA needs to modify its contributions to the administrative expenses of commercial banks for Productores en Desarrollo P to cover the reimbursement of transaction costs only for low-income producers. 29. Technical assistance is FIRA's most effective mechanism to support increased agricultural production through adoption of improved farming practices. In view of the partial public goods nature of this service, the GOM should continue to pay at least part of the costs of extension. 30. Finally, Government transfers to FJRA should be made more transparent. FIRA should pay the Treasury Bills (CETES) rates on all its outstanding loans with BANXJCO. While interest rates remain controlled and below market, Government should absorb the costs of FIRA's rediscount margins to commercial banks through transparent budget transfers. 31. BANRURAL needs to be weaned from Government transfers. BANRURAL has moved closer to the desirable objective of financing only low-income producers with commercial potential. However, it is still subject to political interference and is not yet a financially viable institution. Its loan recovery rates are still low and its costs are high compared to the commercial banks. Additional reforms are needed for BANRURAL to be useful to the rural sector. To this end, BANRURAL should: a. continue current efforts to analyze loans carefully, evaluate the likelihood of repayment objectively, and design and implement sound policies to mobilize savings; b. continue cutting operational costs by closing money-losing branches, reducing its work force and improving management; S. In 1992, FIRA changed its classification of borrowers to Productores en Desarrollo 1 and 11. This classification applies only to qualification for ARA's complementary services to credit (guarantees, technical assistance, transaction cost reimbursement) but not to preferential interest rates on credit. Productores en Desarrollo I are those producers with annual net family incomes of less than 3,000 times the daily minimum wage. Productores en Desarrollo II are all other producers.  x c. design and implement a system of incentives where employees are rewarded for loans recovered and resources mobilized, and borrowers are encouraged to repay; d. increase its efforts to attract savings in rural areas, for example by paying competitive interest rates on deposits and issuing long-term savings instruments with small penalties for early withdrawals, and aggressively promoting small savings as the basis for financing future consumption and production needs; e. test innovative ways to serve the PBI clientele, for example by accepting and promoting collateral substitutes such as guarantee bonds (avales), group joint liability, character references, etc.; and f. stop financing larger borrowers who could gain access and be better served by private financial intermediaries. To improve the efficiency or rural financial markets and restore the incentives to commercial banks or lending to small producers (PBIs), the level and structure or interest rates should be simplified and PBI rates increased relative to larger producers (OPS), followed by a freeing of interest rates. 32. Interest rate subsidization continues to reduce the resources available to PBls and encourages rent-seeking behavior, accentuating the need to eliminate interest rate subsidies. Moreover, Government controlled interest rate policies for a rather small group of better-off producers (compared to PBls who make up the bulk of the rural population) who receive the bulk of FlRA's and possibly BANRURAL's funds, leads to unfair competition. Commercial banks cannot compete with the on-lending interest rates set on FlRA and BANRURAL funds as these rates do not reflect the costs and risks of lending in agriculture. ~ 3. Interest rates subsidies should be phased out. The ultimate target, however, will be to liberalize on-lending interest rates to agriculture, to allow FIRA, BANRURAL, and commercial banks to set freely their lending rates on all their loans. i Abolish insurance premium subsidies to promote competition between private I agricultural insurance and the Mexican Agricultural Insurance Company L_ (AGROASEMEX) 34. Experience with agricultural insurance in Mexico has been poor. The existence of insurance did not stabilize rural income, notwithstanding its large coverage. Although there was no explicit subsidy on insurance premiums, the high probability of getting compensation was a massive subsidy to producers. An important step was closing the National Agricultural Insurance Company (ANAGSA) in 1990. AGROASEMEX is now more cautious in providing coverage and accepting claims, but the insurance subsidy is still (an explicit) 30 percent ad valorem rate. A.,s there is not a good economic reason to s sidize a ricul insuranc , the agricultural insurance system needs to deve op on a commercially soun basis which can only e achieved by letting firms compete for customers without subsidizing insurance premia. F. Fiscal Policies for the Rural Sector (Chapter 7) 35. Federal expenditure reductions were a major component of the stabilization and adjustment programs of the 1980s. During the de la Madrid administration, the total federal budget declined by  xi 5 percent, and between 1989 and 1992, it dropped by 19 percent, reflecting a further tightening of fiscal policy under President Salinas. Rural development expenditures (excluding health and education) also decreased: dropping by 38 percent between 1983 and 1988, and by 11 percent during the first four years of the Salinas Administration to reach N$21.3 billion (US$7.1 billion) in 1992. Despite the expenditure reductions of the 1980s, the relative level of rural development spending in Mexico remains higher than in some other developing and industrial countries. The share of rural development expenditures in the federal budget equaled the share of agriculture in GDP in 1992 (Le., a ratio of 1.0), while the ratio in a sample of industrial and developing countries is 0.6 and 0.4 respectively. The share of rural development expenditures of agriculture GDP in Mexico was 39 percent, compared to 21 and 12 percent for a sample of industrial and developing countries . 36. The composition of rural development expenditures has also changed since 1983, shifting away from market interventions and recurrent costs towards investment and public goods; although the latters' share of total expenditures remains low and risks to drop further in light of the projected costs of the PROCAMPO income support program. The share of rural development expenditures allocated to market interventions (credit, fertilizer and marketing subsidies) has decreased from 55 percent to 37 percent between 1983 and 1992, while the share of resources allocated to public goods J and land titling has risen from 27 percent to 40 percent over the same period. The overall level of recurrent costs has declined steadily since the 1980s, dropping from 75 percent of total rural rc.P""""'>~/ development expenditures in 1985, to 55 percent in 1992. Decreasing subsidy expenditures by the ~_~VA development finance institutions (DFls), CONASUPO and FERTIMEX are the key factors behind the ~I decline in recurrent costs. Recurrent costs also declined for the Secretariat of Fisheries (SEPESCA) and the Secretariat of Agrarian Reform (SRA) since 1985, reflecting decreasing transfers to dependent agencies (the majority of which have been privatized or liquidated) and lower operating costs, mainly due to personnel reductions. 37. Recommendations to improve the overall level and distribution or resources to: a. Avoid increasing rural development expenditures, which are already relatively high compared to other developing and industrialized countries; b. Decrease the cost of PROCAMPO or the share of subsidies and recurrent costs in the budget will rise significantly. Even if macroeconomic conditions permit a transfer of ~ Pf..O~' resources from other sectors to support PROCAMPO, the GOM should avoid this, as "A- ,..... ". there are already, overall, adequate resources allocated to the sector. The GOM should also avoid reducing expenditures for public goods as this would diminish the ability of the sector to adapt to the changing economic conditions brought on by PROCAMPO and NAFTA. 38. Efficiency or Rural Development Expenditures. The efficiency of rural expenditures in Mexico could be improved significantly. Four sets of factors undermine the efficiency of rural expenditures. First, the ~ence of a well-defined rural development strategr has led to an expenditure program that does not necessarily reflect the development priorities of the sector, nor include the most efficient investments. Second, efficiency gains from decentralization are undermined by an inadequate assignment of expenditures among the different levels of government, a lack of clear distribution criteria for Solidarity resources, and by insufficient revenue sources for local and state governments to carry-out their expanded responsibilities. Third, the absence of multi-year budgets undermines investment planning as well as the availability of resources to support operations and maintenance activities. Fourth, the budget process is very centralized and controlled by the Ministry of Finance and Public Credit (SHCP), and as a result the budget does not always reflect the  xii expenditure priorities of the sector. 39. Recommendations to improve the efficiency or rural development expenditures are to: a. increase coordination on a program-by-program basis to avoid overlap and contradictions among the different programs in the sector. A consistent review of functions and responsibilities of the different agencies involved in rural development policy might allow further useful consolidation of activities in particular agencies; b. develop and apply a standard and rigorous project evaluation method (which includes environmental criteria) and ensure that there are adequate resources in SHCP and the National Ecology Institute of the Ministry of Social Development (lNE/SEDESOL) to review project evaluations; c. transform the State Planning Committee (COPLADE) into a consultative group and limit its ability to allocate resources geographically and among different programs to allow local governments, who have greater accountability to the population, distribute resources; d. evaluate the assignment of expenditure responsibilities at the federal, state and local government levels as many services are not clearly assigned to a given level, leading to their under-provision; e. continue the GOM policy of decentralization of rural development activities and expenditures to state and local governments; and equally as important, ensure that they have adequate resources to carry-out their expanded responsibilities; f. increase resource flow to state and municipal governments by matching grant programs in which eligibility and funding level respond to well established objectives and transparent criteria, or by increasing the revenue generation capacity of the municipalities, in order to permit them implement their expanded responsibilities efficiency and fully; g. assign expenditure responsibilities (e.g .• state. municipal) for capital investments and recurrent expenditures to the same level of government to ensure that adequate funds are available to cover recurrent costs; h. develop a long-term budget and investment planning framework to permit planning for recurrent and capital budgets and to facilitate fiscal policy; i. reform the current budget classification system so that maintenance appears under a separate new "chapter" in the recurrent cost budget; j. reduce the ability to introduce extraordinary budgets once the regular budget approval cycle has been closed, since these can significantly alter the composition of the federal budget; and k. expand the authority of municipalities in budget planning and expenditures by transferring the decision making power to the municipalities for capital investments for those goods and services they provide. This process should be carried-out  xiii gradually taking into account the overall implementation capacity of each municipality . 40. Poverty. Expenditures by Secretariat of Agriculture and Water Resources (SARH), the DFls, CONASUPO, as well as the revenue sharing and Solidarity programs, are briefly reviewed to assess if they reach the rural poor. While a comprehensive evaluation of the impact of each program on low income rural families is beyond the scope of this study, the analysis does suggest that many of the GOM's investment and subsidy programs are not effectively reaching the rural poor. 41. Recommendations to improve the impact of rural development expenditures on the rural poor are to: a. evaluate the revenue sharing program, not only to increase the overall amount of resources transferred to the states, but also to ensure that poorer states and municipalities are favored in the distribution process; b. distribute Solidarity resources to the states and within the states to municipalities using explicit rules and criteria to increase the equity of the program.; c. reconsider the role of FIRA as a development bank given the high proportion of its J~ lending directed towards wealthier producers in the sector; d. expand public irrigation and related infrastructure investments in rainfed areas; and e. reinforce SARH's practice of supporting larger delegations in poorer states that have weaker capacities to carry-out rural development activities. 42. Flscal incentives for private investment. The agricultural sector in Mexico has traditionally received favorable treatment in the tax laws, which has encouraged more resources to flow into the sector than is economically efficient. While fiscal reforms in the Salinas administration have changed agricultural taxation, the tax structure still favors the sector. Moreover, the tax structure also favors capital investments, which suppresses the demand for rural labor, and disproportionately benefits larger farmers. 43. Recommendations to create more neutral incentives for private investment are to: a. incorporate large agricultural businesses into the general income tax regime; and apply the general eligibility rule of the simplified regime (annual proceeds of under N$600 million) to every sector of the economy, including agriculture; b. correct the anti-labor bias in the Gross Assets Tax by including all types of assets in the base and by eliminating or reducing the 100 percent tax credit for capital equipment; c. incorporate the agricultural sector in the general VAT regime by eliminating the widespread use of zero rating except for exports, as in the general case. Basic foodstuffs could be exempt from the VAT. Small farmers should remain exempt through the application of the general small business rule; and d. increase the effort to support state governments in updating the cadastre so that  xiv assessed values are closer to market values under the property tax to provide municipalities with increased revenue. G. The Land Market (Chapter 8) 44. The reform of the Agrarian Law is a major achievement. In the long term, that reform will better incentives to invest and improve the efficiency of farm land allocation. However, because productivity differences between previously restricted land classes and other, less restricted classes, appear not to have been very great, the short to medium term productivity effects of the agrarian reform are not expected to be very important. 45. Recommendations to improve the emciency or the land market are to make the land titling program more effective through: a. Cost Effectiveness. Some of the surveying work should be open to competitive bidding. Less precise and costly technologies may be more suited for certain areas where the benefits from titling will be limited. They also may be more accessible to private surveyors to document future changes in land holdings. b. Specificity or Program Design. The high cost of the surveying technology, the complexity of the registration procedures and the focus of the communications program are most appropriate for the wealthier ejidos. The program should be reviewed to reflect the needs of ·special· ejidos, such as those in marginal rainfed, forest, coastal and urban areas; and c. Institutional capacity. The Agrarian Attorney General's office needs to have more capacity to mark boundaries and to arbitrate tenure conflicts, especially among indigenous groups and squatters. Similarly, the capacity of the National Agrarian Registry (Rejistro Agrario Nacional) and the State Public Registries need to be upgraded. H. The Tropics (Chapter 9) 46. Income per capita in the tropical states of Mexico is below the national average. However, raising income in the tropics presents enormous environmental, economic and social problems, however. Environmental problems are a high rate of deforestation, resulting soil erosion and destruction of such environmental values as biodiversity; cultivation on fragile soils and steep slopes, resulting in further soil erosion and land degradation; intensive use of agrochemicals on such crops as cocoa and banana, resulting in groundwater contamination and destruction of wildlife. Economic problems are low prices for some of the main commercial crops (e.g., coffee and cocoa), the uncompetitiveness of some of the sugar mills (they are important employers, notably in Veracruz), the uncompetitiveness of much of rainfed agriculture in the tropics against foreign competition, and the development of the tropical forest products industry without destroying nonmarket environmental values. Social problems arise from the fact that a good part of the tropical population is indigenous and segregated from much of Mexican life. 47. Recommendations to make agricultural growth in the tropics more vigorous and sustainable are:  xv a. Realistic decentralization. The withdrawal of some of SARH's presence in the tropical states has not been compensated by an equal strengthening of state capacity. It should be recognized that the transition period to a fully decentralized system could be much longer than expected and that it may be necessary to continue providing some public agricultural services from the federal level for some time until the state government, local governments and private sector can do so. b. Institutional development. The process of de-concentration and/or decentralization which is being implemented by the GOM needs further support in order to consolidate the mostly normative role that federal bodies should have and the executive role expected of state, municipal and communal agricultural institutions. Furthermore, there is a strong need to improve interinstitutional coordination at the state level on rural development issues. The lack of regional coordination could be tackled through the creation of a regional board, or a similar type of forum, for discussion of relevant state policies and programs. Eventually, a technical unit subordinated to this board could be made responsible for preparing recommendations adapted to the study area. i. In the short run, due to the difficulty of clearly distinguishing between groups of producers, PRONASOL's support to productive activities should continue. It must be considered, however, that the continuation of crop-linked, and not family-linked, programs could reduce incentives for more rational land use which pricing reforms and PROCAMPO have been designed to foster. Therefore, PRONASOL support to productive projects should be delinked from specific types of activities (e.g., maize) and tied exclusively to poverty criteria. c. The environment. Costly soil erosion in the tropics is a fact of life; a continued public extension effort to stop erosion is therefore necessary. Technical solutions for degraded soils exist but are not applied because farmers do not know of them. Pilot activities on selected small catchments have to be supported by appropriate strengthening of agricultural research and technology transfer. d. Research. SARH's restructuring has disrupted public technology generation and transfer. Tropical crops and agroforestry research programs for the region should be restructured. There is an urgent need to maintain genetic materials that was held by institutions since dismantled by the federal government (e.g., The Mexican Coffee Institute and the Mexican Institute for Sugar Improvement). i. There are few potential crop production technologies that provide competitive returns at world prices. There are a number of topics (forestry, integrated pest management, for example) that warrant further public research because their cost and risk make them unlikely to interest the very limited private research capacity in Mexico. ii. Specific research requirements could be integrated into adequately supported tropical research programs conducted by INIFAP. Additionally, the region possesses a valuable set of technological institutions which could also develop specific research lines on a contract basis. This system is already being used in the region,  xvi although on a limited scale. e. Extension. The GOM clearly intends to withdraw SARH from public extension. Nevertheless, several programs conducted either directly by SARH, CNA, and FIRCO provide producers with wholly or partially subsidized technical assistance. To avoid sending different messages to growers about paid technical assistance, either through individuals or firms, these actions and programs should be made mutually compatible and clearly defined. The success of technology transfer for livestock in Veracruz indicates that relatively simple, inexpensive and efficient extension systems can be developed. Technology transfer programs for the large number of small producers with a heterogeneous agricultural potential could be developed and supported. Extension activities could be delivered (or financed) by the state or municipal government; the model developed by FIRCO seems to have some key elements which could be adjusted to the requirements of the different groups of producers. Additionally, programs directed towards upgrading the extension skills of private technical assistants could be supported. f. nrrastructure. The two key elements of infrastructure are rural roads and irrigation/drainage. There are several areas in Chiapas and Oaxaca where producers O with potential are far as 70 km from even a dirt road. Commercial agriculture cannot be developed under those circumstances. Priority must be given to the rehabilitation of the existing rural road network and to the development of all-weather dirt roads. i. As for irrigation/drainage schemes, CNA would have to give priority not to the development of new infrastructure but to the completion of on-going work and rehabilitation of districts and units, always coupled with the establishment of WUOs for operating and maintaining works. Transferring irrigation responsibility to WUOS will still need temporary government-supported technical assistance. g. Coffee and cocoa. A coffee and cocoa rehabilitation program should be supported by the GOM. Both crops are mostly in the hands of small farmers and have a significant social importance, in addition to contributing to the stability of the tropical environment. Elements of PRONASOL's coffee rehabilitation program could be adapted to other areas and crops. h. Sugarcane. i. Social security payments. A sunset provision should be put in the next Cane Decree to place the social security payments for sugar workers on the same basis as payments for other agricultural labor . ii. Uncompetitive mills. The Government has classified mills according to their degree of competitiveness. Some of the least competitive privatized mills have high unpaid, and perhaps unpayable debts, leading to further pressure on the Government for support to keep the mills open. The Government should let mills close, rather than keeping them open indefinitely and thereby losing money indefinitely, and give generous severance payments to the workers. iii. Research and extension. The liquidation of the Mexican Institute for Sugar  xvii Cane Improvement (IMPA) has eliminated public research in sugar cane. In principle, the sugar mills have agreed to assume the responsibility for sugar cane research and extension, but have not done so. iv. Economic incentives. The producer price of cane in Mexico will be above the world price and below the US price. The post-NAFrA price of cane is more than sufficient incentive for cane production and the Government should give no additional support to cane production or milling. I. Soil Conservation (Chapter 10) 48. Soil erosion is one of the three chief environmental problems in Mexican agriculture. 6 The Government has recently taken important steps that lessen erosion. Firstz.,.it has solidified land ri2h..!,S 7 ~ w~ and set the foundation for a land market; these st!IDs are expected to increase investment in land, and . 1,..oOA'1 "more specifically. in land improvements that will conserve sol!: Second, it stopped subsidizing ("Yb~ ""VI~ fertilizers in 1990 and this will give farmers greater incentive to fight erosion. Third, subsidized n n$<.. "'If 8 credit for land clearing has been stopped and stricter procedures to authorize land clearing exist. Fourth, the proposed elimination of protection to maize and bean production - two highly erosive crops - will stop some cultivation on some marginal lands. 49. Economic analysis of soil conservation investments indicates that many would be profitable, so lack of profitable practices, even at high social discount rates, is not an insuperable barrier. easures are needed to extend those ractices to poor producers who will not adopt them purely on 50. ------ the basis 0 market effects. Recommendations to reduce the costs of soil erosion. - -..~.-~ a. Give clear institutional responsibility to one agency for soil conservation, except research on the issue, which should remain with INIFAP. This has not been done since the disappearance of the General Directorate for Soil and Water Conservation (DGSA) in 1985. SARH would have two roles in soil conservation: scientific and extension. The former would be maintained in INIFAP. The latter would be done by strengthening centralized support for soil conservation. b. Restructure the legal basis of PROCAMPO to use its funds to purchase erodible land for public reserves. If properly targeted to areas of severe erosion and low production value, a substantial area could be bought. A related idea would be to give payments to reserve land from production. Land purchases and payments for leaving land idle are alternatives to subsidies or technical assistance for soil conservation. If techniques which are both economically and fmancially profitable are available, then they produce much higher net benefits than a land reserve under the same conditions. If the only techniques available are not economically or financially profitable, then it would be better to pay farmers to do nothing with the land because the costs of reserve payments (including supervision costs) are less than the subsidies required to make farmers adopt such techniques. 6. The others are deforestation and groundwater loss. A Bank sector study on natural resources and forestry is reviewing the former; future sector work will study the latter. Only loil erosion and conservation are treated in the present paper.  xviii c. Target its extension program to severely affected areas exclusively. It is economically beneficial to promote appropriate techniques through the extension channels. For small and medium scale farmers this would be done through the public extension service. For larger and richer farmers it could be done through private extension, whether contracted individually or through producers' associations. SARH's extension role would be to provide technical assistance through the rural development districts (DDRs) and in collaboration with FIRA and FIReO. • d. Improve knowledge of soil erosion and conservation. • i. Productivity effects. A major remaining area for research is the productivity effects of erosion and conservation, plus the costs and benefits of erosion control for different crops and farming situations. ii. Off-farm and on-farm effects. The sizes of on-farm and off-farm effects are a major gap in knowledge because they determine the subsidies needed to elicit optimal on-farm soil conservation investments. The additional research required to estimate those effects may be relatively cheap, however, because the magnitude of erosive deposits in waterways can be estimated with mathematical models developed elsewhere. It is unlikely that further information about those relative magnitudes would change erosion control techniques, but it might affect the target areas and the public actions to control erosion. iii. Affected regions and activities. The most severely affected regions and crops are well-known, not only from field studies in Mexico, but from analogous research carried out in similar agricultural situations throughout the world. Therefore, further descriptive studies of, for example, the quantities of erosion loss per state or even per DDR, should not be apriority. One major unknown is the effect of livestock on soil erosion and the benefits of soil conservation practices in grazing areas. e. General subsidies to promote soil conservation are not justified economically except in unusual circumstances. Those circumstances are the very few sites at which the economic benefits of soil conservation are positive but private benefits are negative. At such sites, it will be cheaper to pay farmers to reserve the erodible land from production or to buy the land. f. Construction of public works on-farm for soil conservation are justified infrequently. Public works are ineffective because farmers lack incentives to maintain public works constructed. Rich farmers can corrupt public agencies and capture a disproportionate share of works and maintenance. Soil conservation practices are divisible and simple enough that private agents can build them with equal qUality. Some of the failure of public soil conservation activities in Mexico has been due to targeting investments wrongly. More sustainable conservation investments can be achieved by targeting them more efficiently through a clear plan to share costs and responsibilities. While public financing, as subsidies to producers or to construction frrrns, may sometimes be necessary to promote some techniques. the initial investments are done more efficiently by competitive private agents.  xix g. Provision ot incentives through public financial institutions tor soil conservation is inefficient. This will lead to excessive credit flows to dubious soil conservation investments as a means of getting subsidies. FIRA would provide commercial finance to richer producers who would contract extension privately, but would not transfer erosion control subsidies. The main reason for not channeling subsidies through FIRA is to avoid them where erosion is light. Co-generation of projects is an important self-selection mechanism to maintain soil conservation structures.  1. INTRODUCTION I. I Agriculture is the weakest sector of the Mexican economy. Agricultural productivity and income per worker are much lower than in industry and services. The growth of productivity and income in agriculture has been, and continues to be, weaker than in competing sectors. The chief result of the feeble productivity growth is widespread rural poverty. The projected effects of the North American Free Trade Agreement (NAFTA), though positive for the Mexican economy as a whole, may reduce agricultural income overall, widen the disparity with other sectors and promote even more rapid emigration from rural areas. 1.2 This Memo relates the performance of Mexican agriculture from 1950 through 1992 in an effort to explain: (a) the reasons for its weak performance, and in particular the slowdown in growth since the late 1960s; (b) the policy reforms since the mid-1980s; (c) the effects of those policy changes on sector growth; and (d) policy options to restore future growth. Most of the historical information is organized in six year periods covered by each Mexican administration (sexenio) since 1950 (e.g., Table 1.1). 1.3 A few universal factors determine agricultural growth in any country. They are: (a) natural resource endowments; (b) economic incentives, as determined by world prices, public policy, the macroeconomic environment and the transaction costs in accessing markets; (c) the agricultural technology available to farmers, as determined by the methods generated domestically and those generated abroad and then imported; (d) irrigation, as one of the principal types of agricultural investment and as the motor of Mexican agricultural growth for many years; (e) the performance of rural financial markets; (f) public spending on the farm sector; (g) how the costs of natural resources are accounted, though the latter is only a recent consideration; and (h) a land tenure system that provides security of access to land, allows land transactions and rental contracts, permits land to serve as a collateral on loans, and values investments in land improvements. The Memo treats those factors as follows. 7 1.4 Macroeconomic Environment. Section II of this chapter briefly discusses the macroeconomic developments of the Mexican economy as well as the major changes in public policy after the structural adjustment of the mid 1980s. 1.5 Natural resource endowments. Section III of this Chapter outlines the structure of Mexican agriculture and its resource endowments - Chapter 9 discusses the particular problems of the tropics and Chapter 10 reviews soil conservation problems. 1.6 Economic incentives and public policy. Chapter 2 reviews transfers from the government and consumers to producers through trade, price, and official purchase policy. Key points are: (a) the size of historical transfers from price policy and indirectly through the exchange rate and the domestic terms of trade; (b) abolition of trade/price protection to major crops and its replacement with delinked income supports in a program known as National Farm Modernization Program (pROCAMPO); (c) effects of the period of adjustment after the mid-1980s on Mexican agriculture; and (d) the projected effects of NAFT A and PROCAMPO. 7. The Memo refers to the conclusions of previous sector work where relevant, such as Mexico - Agricultural Sector Memo, Report No. 7609-ME, April 21, 1989; Mexico - Policy Notes on Agriculture, Food, and Rural Development, Report No. 8144-ME, February 1. 1990.  2 1.7 Commercialization issues (Chapter 3). With the removal of the National Basic Foods Company (CONASUPO) from trade and commercialization activities, the Government plans to assist farmers by creating spot markets (and possibly at a latter stage forward and futures markets). Chapter 3 elaborates on the new commercialization system by outlining the requirements and the potential problems associated with the spot markets. 1.8 Irrigation and water markets (Chapter 4). The Memo reviews the performance of irrigated agriculture, its role in growth, and the barriers to continued growth. Key points are the analysis of the evolution of public spending on irrigation investments and the legal and administrative changes that may permit the development of efficient water markets. This Chapter relies on Bank sector work and project preparation documents (Mexico - Irrigation and Drainage Sector Project Staff Appraisal Report and On-Farm and Minor Irrigation Networks Project Staff Appraisal Report). 1.9 Agricultural technology: research, intellectual property, and extension (Chapter 5). Growth of crop, livestock, and forest output comes from increases in area, yield per unit of area, shifts to higher value activities, and greater cropping intensity. Economic limitations on the supply of land have made it difficult for more than 25 years to expand the cultivated area and increases in area now have little remaining role in growth. Gains in yields, shifts in the cropping pattern, and more intensive cultivation are the main sources of growth, but depend on using more physical inputs (e.g., fertilizer and machines) and labor per unit of output, or on obtaining more productivity from each unit of input. Using more inputs per unit of output is only profitable if the price of inputs falls or if the price of outputs rises. Higher productivity depends on technical change, which in turn is a function of technology generation and transfer through research and extension. Key points in the Memo are a review of the performance of the Mexican agricultural research system; the possibility of greater imported technologies; and the consequences of recent legal and regulatory changes on research products, either domestic or foreign. Chapter 4 relies on previous Bank sector work (Mexico - Agricultural Technology Sector Review, Report 9297-ME, August 20, 1992). 1.10 Rural finance (Chapter 6). The analysis of rural financial markets is limited to an updating of Bank experience with rural financial operations and to an analysis of the costs of government subsidies to its agricultural development banks, insurance funds, and other trust funds. 1.11 Fiscal Policies for the Rural Sector. (Chapter 7), This includes a review of public spending on agriculture since 1982, analysis of the directed antipoverty program (National Solidarity Program, PRONASOL), and taxation questions. 1.12 The farm land market (Chapter 8). The land-tenure system in Mexico was developed after the revolution and subsequently embodied in the Constitution in a special restricted land class known as the ejido. Under this system, 53 percent of Mexican land was distributed to peasants, called ejidatarios. The ejido system was associated with several inefficiencies. The most important are listed below: (a) lack of access to credit due to lack of collateral usable by commercial banks and inadequate access to National Rural Bank (BANRURAL) credit Oimited to fixed amounts per hectare that correspond to extensive corn techaology at best); (b) restrictions on land use imposed by the Government or by collective rules (e.g., obligation to open the land in crops to livestock at a fixed date for all, preventing individual innovations that require a longer cropping season; (c) cooperative failures; and (d) no legal right to lease the land or to hire labor, which is particularly ineffective for households heavily vested with migration, or with no access to credit. Thus, farmers had no incentive to make major improvements in the land; on the contrary, in the absence of property rights they had an incentive to overexploit the soil beyond what would have been socially optimal. In 1991, the Government recognizing the inefficiencies of such tenurial arrangements, announced the end of  3 the existing land program and proposed a modification of the Constitution to permit the sale and rental of ejido land. In 1992 the proper legal framework was created to regulate the privatization of the ejido lands and to ensure property rights. Under the new system the ejidatarios have security of tenure and the option to remain on communal property or to become individual proprietors. Furthermore, in addition to being allowed to lease out their farm, they can enter into joint ventures with other ejidatarios or the private sector (local or even foreign). 1.13 The tropics (Chapter 9). The poorest states of Mexico are largely tropical and present difficult economic, social and environmental problems. An FAO/CP mission recently reviewed four states - Chiapas, Oaxaca, Veracruz, and Tabasco - in the humid tropics. An edited summary of its findings appears as a chapter in the Memo. 1.14 Soil conservation (Chapter 10). The executive summary of a companion piece of sector work on soil conservation is presented in this chapter. 1.15 What the Memo does not discuss. The Memo excludes some questions about the rural sector. It says nothing about rural non-farm employment questions, labor markets, or the capacity of public sector institutions. The first two would require information that can only be gathered in a detailed and costly research project. The capacity of public sector institutions is more an issue of project preparation and has been deferred to that (hypothetical) phase. The Memo does not discuss questions relating to indigenous peoples. An independent Resource Conservation and Forest Sector Review (Report No. 13114-ME) treats most of the environmental questions, except soil conservation. I. The Overall Economic Environment and Policy Rerorms 1.16 Prior to the debt crisis of 1982, Mexico enjoyed a period of sustained high economic growth and price stability. Between 1940-80 the GDP grew at an annual rate of 6.2 percent. However, behind high growth lay a deterioration in fundamentals. Conservative fiscal policies were able to keep inflation in single digit levels until 1972; but when replaced by more expansionist ones, they led to double-digit inflation and eventually to the collapse of the exchange rate in 1976. Expanded public borrowing, high world interest rates and the sharp decline in the price of oil in 1982 led to the debt crisis. Economic performance after the debt crisis was poor. For example, average GDP growth during 1983-88 was 0.2 percent while the average inflation was 86.4 percent (see Table 1.1). 1.17 In the mid-1980s, Mexico began to reform its economic policies away from state intervention and toward a market basis. The reform consisted of three major aspects: (a) fiscal adjustment; (b) trade liberalization; and (c) deregulation and legal changes. 1.18 Fiscal adjustment has been remarkable, with the main changes coming from the expenditure side. Non-interest expenditures, which had been at almost 35 percent of GDP in 1981, were by 1992 reduced to 20.5 percent. Capital expenditures were cut in half over the same period. While the cuts brought about increased efficiency of Government operations by eHminating most of the dubious large projects of the late 1970s, expenditures on maintenance and the social sectors were also reduced sharply. Total interest payments had represented an important expenditure component, but were reduced sharply by 1990-91 with the reduction of the debt burden. From 1982-1990 interest payments ranged from 12 to 20 percent of GDP, while in December 1992 they were only 3.9 percent of GDP. Following the renewed confidence brought about by the continuity and growing success of structural reforms (which led to lower interest rates), plus as a consequence of the steadily falling  4 debt to GOP ratio (due to strong fiscal accounts), domestic interest payments fell to 5.5 percent of GOP in 1991. 1.19 Tax Reforms. Although increased total government revenue contributed only 3 percentage points of GOP to fiscal adjustment, the efficiency of the tax system has been substantially improved. A major tax reform was initiated in 1987. The reform expanded the tax base, simplified the tax system, reduced rates, and modernized tax administration. Personal and corporate tax rates have been brought in line with international levels. The highest personal tax rate and the flat corporate tax rate are now 35 percent, compared to 60.5 percent and 42 percent in 1986. The Value Added Tax (VAT) had four rates, from zero percent to 20 percent, with a base rate of 25 percent, but in November 1991 a uniform rate of 10 percent was introduced for all sectors. In early 1990 the administration of the VAT, which until then had been assigned to the states, was taken over by the federal authorities. The excise tax on gasoline, as well as gasoline prices, were sharply increased in 1990-92. 1.20 Trade liberalization. In 1982, tariffs were high with a maximum of one hundred percent. Furthermore, quantitative import restrictions were used as the main vehicle of protection. In addition to the base tariffs, the Government used the option of applying surcharges of up to fifty percent of the base to provide temporary protection to industries whose non-tariff barriers (NTBs) were eliminated. After the reforms introduced in December 1987, tariffs were reduced to five levels, the maximum tariff was set at 20 percent, and official reference prices were abolished. Domestic production covered by import licenses settled at about 20 percent in 1988 and the production-weighted tariff was reduced to about 11 percent. An exception to the liberalization, as of 1989, was the retention of import licenses on about 40 percent of the value of agricultural products though this later fell. Table 1.1 Evolution of the Mexican Economy: 1950-92 (period averages) Current Annual growth Annual change account deficit Public deficit Period of GDP (%) in CPI (%) (as % of GDP) (as % of GDP) 1950-58 6.7 8.3 2.5 n.a. 1959-64 7.2 3.6 2.2 n.a. 1965-70 6.2 3.9 2.1 1.7ai 1971-76 6.0 14.1 3.8 3.9 1977-82 6.6 14.1 4.3 6.8 1983-88 0.2 86.4 0.1 11.2 1989-92 3.8 27.2 3.7 n.a. 1/refers to 1966-70 period. Source: World Bank, LA2NR Mexico Agricultural Database. 1.21 Deregulation and legal reforms. Although major steps towards deregulation and legal reforms have been taken for some time, the most important steps towards private sector participation in the economy were taken by the previous (Miguel de la Madrid, 1982-1988) and current (Carlos  5 Salinas de Gortari, 1989-present) administrations. 8 Deregulation bas been extended to industry and transportation. The national telephone system was privatized. A cumbersome transportation licensing system has been streamlined and the sector has been opened to competition. In December 1992, a law was passed that permits private ownership of electricity generating plants. A program to reform the public enterprises that will remain in public hands is being implemented. Enterprises targeted for major restructuring include rail services and food distribution. Although oil production will continue to be a Government monopoly, foreign participation in the petrochemical sector is being extended as the list of petrochemicals reserved for the public sector have been very significantly narrowed. 1.22 As a result of the fiscal measures, the overall fiscal balance went from a deficit of 15.6 percent of GDP in 1982 to a surplus of 1.6 percent of GDP in 1992. The recovery of growth, however, was still modest. GDP was almost unchanged between 1982 and 1988, implying a decline in per capita terms of 11 percent over that period. During the period 1989-91, growth in total GDP recovered to 3.7 percent annually. 1.23 Despite the successful implementation of economic reforms, economic growth has slowed since mid-1991. In 1992, GDP grew only at an annual rate of 2.8 percent. Estimated GDP growth for 1993 was only 0.4 percent according to the initial estimates released by the National Institute for Statistics, Geography, and Information (INEGI). According to the most recent estimates of INEGI, GDP growth was only 0.3 percent in the second quarter, a negative 0.8 percent in the third quarter, and a negative 0.1 percent in the fourth quarter. Given that the average annual population growth is about 1.9 percent, no per capita real GDP growth is expected to be observed in the near future. The trade accounts have worsened markedly, with a current account deficit of $22.8 billion in 1992 (Le., 6.9 percent of GDP). On the positive side, however, inflation in 1993 dropped to single digits (8.1 percent, based on the December 1992 to December 1993 change in the consumer price index). 1.24 Poverty. Official estimates indicate that about 20 percent of the population is still below the poverty line of $350/year. Income inequality is severe and may have gotten worse as a consequence of the economic crisis of the 1980s. The Government is attacking poverty through public expenditure programs targeted to the poor. PRONASOL, which was initiated in 1988, targets health, education, nutrition, housing, infrastructure and other projects to the poor in disadvantaged areas. 1I 1.25 Productivity. Productivity growth in Mexico should have been fostered by the structural reforms, notably trade liberalization and deregulation. Productivity has indeed picked up recently, but remains low and has not offset the loss of competitiveness brought about by the real peso appreciation. Gross labor productivity in manufacturing increased at an estimated 3.9 percent cumulative annual rate over 1987-92. Similarly, total factor productivity (TFP) growth has responded positively to reform with big asymmetries across sectors. Estimates for TFP growth in manufacturing fall in the 3 to 5 percent range over 1985-90. However, trends of differential gross labor productivity growth in manufacturing over longer periods do not produce grounds for strong optimism. Over the 1980-92 period, Mexico's labor productivity growth has been less than developed countries like the 8. The official shift to liberalism took place in 1985 when the administration of President de la Madrid announced that Mexico would seek membership in the General Agreement on Trade and Tariffs (GATT) and that it would liberalize trade policy. 9. PROCAMPO, which is going to replace the current subsidy scheme, may have some income redistribution effeets as farmers who have produced a number of eligible crops in the past three years will be receiving a N$350 per hectare for the ten years starting 1994 (sec Table 2 in text for information on the transition period).  6 U.S. and Japan, and far less than in some of Mexico's competitors in the U.S. market, e.g., Korea and Taiwan. 1.26 Exchange Rate. The nominal exchange rate remains an important ingredient of the Economic Solidarity Pact (PECE). Measured against U.S. wholesale prices, much of the increase in competitiveness that occurred during the 1986-87 depreciation has been lost. In October 1992, Mexico made the exchange rate more flexible by widening the existing band and raiSing the preannounced rate of devaluation of the band ceiling. The widening of the band will: (a) enable the market to determine in a more efficient manner the appropriate level of the exchange rate; (b) reduce the volatility of capital flows, thus facilitating monetary policy; and (c) reduce the risk of abrupt changes in the exchange rate regime. On the other hand, the fact that the rate is only partially flexible (a maximum 4.6 percent annual devaluation of the upper limit of the band), while limiting some of the above advantages, will reduce the fear that inflation might rise. The market reaction to the October 1992 exchange rate reform was favorable; following the announcement, peso- denominated interest rates declined and the exchange rate appreciated. Contractual wage revisions in the private sector, a better indicator of medium-term devaluation and inflation expectations, declined sharply from an average of 15.4 percent in the third quarter of 1992 to less than 10 percent in the last two months of 1993. 10 1.27 Because of fears of inflation and of damaging policy credibility, exchange rate policy was to set a daily nominal devaluation of the official rate. However, these daily nominal devaluations were not large enough to offset Mexican inflation so the real exchange rate appreciated. In October 1992, the exchange rate policy was adjusted to widen the existing band and to raise the rate of elevation of the band ceiling. The new regime is supposed to enable the market to determine more efficiently the appropriate exchange rate and it will also reduce the potential of abrupt changes in the exchange rate. Bank sector work on the exchange rate (World Bank, Report No. 1191, 1993) argued that widening the band was an appropriate step because it allowed a larger role for market forces in determining the parity, and eased fears of a very large devaluation while giving somewhat better incentives to producers of tradable goods. II. Natural Resource Endowment 1.28 Mexico's agriculture is heterogenous. There is an irrigated and commercial class of farmers using modern technology, achieving high yields, competing on international markets, and earning high incomes. There is a class of rainfed fanners, who produce mainly for subsistence, achieve yields well below those with irrigation, do not generally compete on international markets, and earn subsistence incomes, in many instances by working as seasonal laborers outside their own farms. 1.29 The Ministry of Agriculture and Water Resources (SARH) has classified land by aptitude for crops, livestock, forestry, and other uses, including bodies of water, urban zones, and unproductive areas (SARH, 1987b) as shown in Table 1.2. A more detailed breakdown of the capacity of the roughly 33.4 million hectares classed as apt for cropping shows that only 8.2 million hectares are in the "very high" or "high" potential classes, with another 9.2 million hectares in the average class. The sum of the very high, high, and average classes (17.4 million hectares) is in fact less than the average area of annual and perennial crops in Mexico (19.4 million hectares in Table 1.2). 10. This was, in part, due to a "guideline" ceiling on contractual wage revisions. The combination of exchange rate, wage, monetary, and fiscal policy has probably played a more important role.  7 1.30 Mexican agriculture has five climates, defined mainly by average temperature and length of growing season as determined by the availability of water. The climates are: (a) Arid. This climate is found chiefly in Sonora, Sinaloa, Baja California Norte, and Baja California Sur; adding Nayarit state, this is the northwest region of Table 1.2. Annual rainfall is less than 200 mm and summer temperatures are so hot that crop production is impossible without irrigation. Where there is irrigation, the main winter crops are wheat, fruits, forages, and vegetables; the main summer crops are oilseeds. sorghum, vegetables, and rice. Cotton was important in the past but has fallen to almost nothing. The average 1991-1992 cropped area in those states was about 1.82 million irrigated hectares and 0.55 million rainfed hectares. The northwest, north, and northeast states in Table 1.2, have about 3.09 million hectares under irrigation. or about 57 percent of the national irrigated area. They produce roughly one-third of the value of annual and perennial crop output (N$18.0 billion of all output, rainfed and irrigated combined). The value of crop output per hectare with irrigation is approximately twice what is under rainfed conditions. There is little or no potential for expansion of rainfed crop land or for conversion of pasture land without irrigation. Expansion of irrigation is limited by water scarcity and by salt encroachment on coastal aquifers used as sources of groundwater for farming, and the scarcity of water is likely to rise as urban areas develop and as water markets become more active (see Chapter 4 on water markets). Table 1.2 Potential Land Use, Actual Cropped Area, and Value of Output, by Region Best potential use Aetual cropped area Value or output (miWOIlS or hectares) (miDiollS or hectares) (billiollS or 1992 N$) Reciou Crops Livestock Foresti")' Other Total Irrigated Rainred Irrigated Rainred i North_Sf 3.77 21.52 4.68 11.48 41.44 1.82 0.55 9.51 1.32 North 2.84 20.61 12.25, 0.97 36.67 0.49 1.08 I 2.68 0.74 Northeasf 5.09 22.02 L25 1.24 29.60 0.78 1.10 2.57 1.18 ! Celller North 3.49 9.62 1.11 0.12 14.35 0.32 1.80 2.82 1.34 Cenler Pacific 4.11 4.07 5.47 0.90 14.55 0.67 1.77 4.09 4.75 Center 4.S0 S.48 2.52 0.25 12.76 I 1.09 2.93 5.82 4.90 Celller Gulf 5.00 2.57 I.4S 0.73 9.75 0.07 1.32 0.41 4.10 Federal Districf 0.01 0.03 0.05 0.06 0.15 0.00 0.Q3 0.00 0.25 Pacific South 2.60 9.78 10.33 0.59 23.30 0.17 2.63 1.06 5.53 Peninsula 1.96 3.22 7.97 1.02 14.17 0.04 0.76 0.17 0.65 Total ~OTES: 33.38 98.93 47.15 17.85 196.72 States b'y regIOn are Northwest: HaJa Califorma NorU and Soutt!, NaY 5.44 13.97 I ant, Sonora, Sma oa; North: Chihuahua 29. 13 1 24.771 Durango; Northeast: Coahuila, Tamaulipas. Nuevo Leon; North Center: Aguascalientes. Zacatecas, San Luis Potosi; Center PacirJc: Colima, Michoacan. Jalisco; Center: Guanajuato. Queretaro. Hidalgo. Tlaxcala. Puebla. Mexico. Morelos; Center Gulf: Tabasco. Veracruz; PaculC South: Chiapas, Oaxaca. Guerrero; Peninsula: Campeche. Yucatan, Quintana Roo. SOURCE: SARH. 1987b and SARH. Anuario Estadistico Agricola. various years. (b) Temperate. This climate is found in the center of Mexico. usually exceeding 1,500 meters above sea level; it is typical of much of the Center North and the Center in Table 1.2. Rainfall is between 500 and 1,000 mm in the summer. The cropping  8 system is mixed, with some areas having a lot of irrigation (e.g., parts of Guanajuato and Michoacan) amounting to about 1.41 million hectares and others having almost none (e.g., most of Hidalgo and Oaxaca), amounting to about 4.72 million hectares. Irrigated summer crops are sorghum, sugar cane, fruits, vegetables, forages, and maize; irrigated winter crops are sorghum and vegetables. Rainfed summer crops are sorghum, maize, and beans. (c) Arid-semiarid. This climate is found in Chihuahua, Durango, Coahuila, San Luis Potosi, Nuevo Leon, Tamaulipas, and Zacatecas (chiefly the north and northeast regions in Table 1.2). Rainfall is typically from 100-400 mm. Temperatures are cooler than in the arid climate so there is more rainfed cropping. The average 1991-92 rainfed area was 2.18 million hectares; the average irrigated area was 1.26 million hectares. The main summer crops are maize, beans, sorghum, cotton, potatoes and forages; winter crops are wheat, maize, other cereals, and forages. Water scarcity restricts the expansion of crop land, which has been almost nil in this climate in the last 25 years. Semiarid forestry and extensive livestock production much of Chihuahua and Durango. (d) Dry tropics. This climate is found along the Gulf and Pacific coasts of central Mexico and is the least extensive of the five, being spread among the Center Pacific, Center Gulf and Pacific South regions depicted in Table 1.2. There is little irrigation. Rainfed summer crops are maize, beans, soybeans, citrus, and pastures. Livestock production is a major activity (e) Humid tropics. This climate is found in the southern states of Chiapas, Oaxaca, Guerrero, Tabasco, Campeche, Yucatan, and Quintana Roo plus much of Veracruz. It is hot and humid all year. Rainfall usually exceeds 1,500 mm. Annual crops are varied and include maize, beans, soybeans, and rice. Perennial crops are important, notably sugarcane, coffee, cocoa, banana, mango, and others. The Pacific South and Peninsula states in Table 1.2 have only about 0.2 million hectares of irrigation and about 3.4 million hectares of rainfed farming, with much larger areas in pastures and forests. The southern states, consisting chiefly of the Center Gulf, Pacific South and Peninsula regions in Table 1.2, have about 5 miUion hectares under cultivation, of which only 0.28 million have irrigation. The irrigated area is just five percent of the national irrigated area. The value of output per hectare is N$2,OOO or just two-thirds of the national average. The southern states use fewer chemical, mechanical and biological inputs per hectare and receive less public agricultural credit (see Chapter 6 on rural finance and Chapter 7 on public expenditure). The specific agricultural development problems of four southern, largely tropical, states - Oaxaca, Veracruz, Tabasco, and Chiapas - are discussed in Chapter 9 of the Memo. III. History Structure of Mexican Agriculture 1.31 In the early 1950s agricultural GDP accounted for about 20 percent of total GOP (fable 1.3). Since then, the share has fallen to 7 percent to 8 percent in the 1990s. The growth of the sector has also a fallen. While exhibiting an average annual growth of 5 percent in the early 196Os, it dropped below 3 percent in the period 1965-76. After a partial recovery in the 1977-82 period, it dropped to historic lows of below 1 percent. Within agriculture, subsectoral shares have changed little in the past 40 years. Crops typically provide about 60 percent of the sector's output, livestock about one-  9 third, and forestry, fishing, and hunting the rest (SARH, April, 1993: p. 153). Within the crop subsector in 1992, cereals accounted for 48 percent, oilseeds for 5 percent, fruits, vegetables, and legumes for 28 percent, sugar cane for 5 percent, fibers for 4 percent, and coffee and cocoa for 6 percent. 1.32 Input use. Chemical fertilizers and machinery have been replacing agricultural labor since 1950. Fertilizer use grew nearly 50 times from 1950 to 1987 and about 20 times per unit of harvested land. Total tractor use grew by 90 percent between 1965 and 1985 and rose 30 percent per unit of harvested land. While the total agricultural labor force grew by about 26 percent from 1950 to 1985, it fell by roughly 45 percent per hectare. A strong stimulus to fertilizer use was heavy government subsidies (see Chapter 2). The agriculture of the northern regions uses more chemical, mechanical, and biological inputs than does that of the south and has typically received more public agricultural credit than did the other states. 1.33 Price policies were burdened with conflicting rural and urban poverty alleviation objectives. Agricultural trade was highly regulated. Small farmers and ejidatarios (see para. 1.12) were subject to excessive regulation especially for credit and land use. While Government expenditures for food and agriculture were excessive, allocation for infrastructure, maintenance, research and extension were insufficient. Targeted food subsidies were not available to poor rural populations. Pervasive presence of loss-making parastatals in the production and distribution of agricultural inputs, as well as in production storage, and marketing of consumer goods, and in the purchase, processing and distribution of agricultural commodities. Table 1.3 Total, Agriculture, and Nonagriculture GDP: 1950-1992 (millions of 1992 N$) Agricultural GDP Period Non-Ag. Total GDP Crops Livestock Forestry Fisheries Total GDP 1950-58 19,077 10,122 1,237 2,213 30,657 117,651 148,307 1959-64 26,071 12,202 1,601 508 40,382 197,074 237,457 1965-70 30,631 14,588 1,886 580 47,684 307,881 355,566 I 1971-76 33,362 20,545 2,659 987 57,553 450,080 507,633 1977-82 39,136 21,183 3,153 2,217 65,690 676,716 742,406 1983-89 44,545 21,231 3,643 2,958 72,378 768,324 850,702 1989-92 39,399 27,675 4,448 3,664 75,186 844,320 919,505 1992 53,090 26,045 4,425 3,511 86,046 932,390 1,018,436 Source: World Bank, LA2NR Mexican Agriculture Database. 1.34 On the input side, the government subsidized electricity, water, credit, and fertilizer. On the output side the decentralized agency of CONASUPO (under the Secretariat of Trade and Industrial Development--SECOFI), which was the key player of agricultural price policies, handled purchasing, processing, wholesaling and retailing for much (usually between 15 percent and 50 percent of total  10 production) of maize, beans, wheat, rice, sorghum, milk, and oilseeds. Until 1985, CONASUPO was the only legal importer of these products. Since then, private processors (traders excluded) have been allowed to import all the above products, except beans and powdered milk. As an example, in 1990 CONASUPO purchased 16 percent and 10 percent of the total production of maize and beans while the associated 1991 figures were 23 percent and 28 percent respectively. As prices received by farmers are above world prices, CONASUPO's deficits are subsidized by the government. For example, the net official transfers to CONASUPO were 3.6 billion NP or 18 percent of total rural development expenditures in 1992. 1.35 Growth from 1940 to 1965. Agricultural growth was slow until 1930 because of revolutionary disturbances, but became rapid after 1945. The sector's growth rate was 4.6 percent from 1940-64 while overall GDP growth was about 6.5 percent. Crops grew at a rate of 5.0 percent annually, livestock output at 2.6 percent, and forestry at 3.9 percent. 1.36 Area expansion, in part through public investment in irrigation, was initially important in raising production. After 1950 this expansion was associated with greater wheat yields (the result of research producing higher-yielding varieties and of increased fertilizer use on those varieties), but not with significantly higher yields of other crops. Cultivation also spread to areas that had not been cropped before or had been pastures. There were some changes in regional specialization, associated with development of high-yielding varieties (HYVs) of wheat in the north Pacific region. Table 1.4 Annual Average Agricultural GDP Growth and Labor Share Agricultural GDP Agricultural Labor Productivity Period Ratio % of Total Growth rate Thousands % of Total 1950-58 20.8 4.90 na na na 1959-64 17.1 3.80 5,867 50 0.17 1965-70 13.5 2.14 5,243 42 0.20 1971-76 11.4 2.62 4,593 1a 32 0.24 1977-82 9.0 3.67 5,272 28 0.24 1983-88 8.6 0.52 5,981 27 0.24 1989-92 6.9 0.75 5,844 25 0.22 The low figure for the 1971-76 agricultural labor is due to a change in the way the Mexican authorities reported the data. 01 Source: World Bank, Mexico Agricultural Database. 1.37 Crop shares. The share of fibers (cotton and henequen) has plummeted and shares of forages and oilseeds have grown. Little change appears for cereals and grain legumes. There has been a rise in sorghum with a corresponding faB in maize, probably reflecting substitution of the new higher-  11 yielding sorghum for maize in some areas. Growth of area and/or yields has generally been independent of changes in cropping patterns. 1.38 Up to the end of the 196Os, Ardito-Barletta (1971) summarized the history of productivity growth this way: (a) changes in the location of production raised productivity by adding new irrigated lands and new lands in the humid tropics; (b) irrigation, fertilizer, seeds, and machines explain 25 to 50 percent of productivity gains from 1940 to 1962; and (c) individual crop gains caused changes in productivity, not changes in cropping pattern; in fact, had individual crop yields not grown, aggregate yields would have fallen as the result of shifts to a cropping pattern with lower-yielding crops. What is important from his findings is that even when Mexican agriculture was growing more rapidly, it was not growing from produGtivity increases but from more land and variable inputs. 1.39 1965 to the present. Beginning in the mid-1960s, agricultural growth began to fall. This drop was associated with slower expansion of total cropped area, as shown in Table 1.2. The aggregate area growth rate was 0.3 percent from 1966 to 1985, compared with 2.8 percent from 1950 to 1965, while the rate of irrigated area expansion fell from 3,4 percent to 2.1 percent. 1,40 The other important characteristic of past expansion of Mexican agriculture was rapid mechanization as labor for higher wages elsewhere. The active labor force in agriculture fell gradually from slightly more than 6 million workers in the early 1960s, to 4.6 million in the 1970s and rose again to 6 million in the 1980s/90s. The agricultural labor force, as a percentage of the total active labor force, fell steadily from an average of SO percent in 1960-64 period to an average of 25 percent in 1989-91. The productivity of agricultural labor has remained poor. The SO percent of the labor force in agriculture produced, on average, 15 percent of GDP in 1960-64 while 25 percent of the labor force now produces 7.s percent of GD P. Considering the ratio of the agricultural per worker GDP over nonagricultural per worker GDP (last rew-of table 1.4), between 1971 to 1988 has remained constant at 0.24, i.e., the per worker GDP of the agricultural sector is 24 percent of the per worker GDP of the rest of the economy. lye '1~ (J;~'" //I~ I, " ~v /1,'" "1,!?"ri ,.- .' '~\v ,- \~'" '1-'"  12 2. GROWm, ECONOMIC INCENTIVES, AND PUBLIC POLICY 2.1 This chapter analyzes the economic incentives given to Mexican producers as a result of the interaction between world prices and Government policies. Section I decomposes the growth of the value of output for the 14 most important crops for the period 1950 to 1992 (organized by sexenio). Section II reports and discusses the nominal protection coefficients (NPCs) for these crops; it also reports the transfers made to producers through output subsidies (or taxes); finally it elaborates on the price stabilization issue. Notice, however, that beans, despite its importance, is a notable absence from this level of analysis; this is the case because there is no standardized quality of beans traded at international level and hence the lack of an international price. Section III discusses the direct and indirect effects of output price distortions; it also briefly elaborates on the effects of the exchange rate on agriculture. Section IV reports estimates of transfers through fertilizer subsidies. Section V analyzes the effects of recent policy reforms on the profitability of four major crops (Le., maize, wheat, sorghum, and beans). Section VI discusses NAFTA and PROCAMPO. I. Decomposition of Growth of Output Value 2.2 Causes of slackening growth were identified by analyzing changes in the value of output for 14 major crops from 1950 to 1992 (Tables 2.1.A and 2.1.B). The total value of production (V) is defined as the product of crop area (denoted as A), output per hectare (denoted as q), and real producer price (denoted as p); i.e., V = (A*q*p). The growth rate in the value of output can be . decomposed into the contributions of changes in total cropped area, changes in yield, and changes in real price, as: In(V,) - In(V,.,) = [In(A,) - In(At-1)] + [In(Qr) - In(Q,.,)] + [In(p,) - In(Pt-1)]. where In(.) denotes the logarithm of the respective variable. 2.3 Aggregate area expansion was most important until the end of the 1960s. For maize and beans, the most widely cultivated crops, area expansion has been weak in the recent periods. The area expansion of sorghum has decreased consistently, as has that of soybean, safflower, chickpea, cotton, and cocoa. The only crops for which area expansion maintained its contribution to the growth of total value were coffee and sugarcane. 2.4 With respect to yields, the picture is mixed. The contributions of yield increase in maize and sorghum have shown no consistent trend, though those of wheat and beans have fallen consistently. The contributions of sugarcane, coffee, and cocoa yields may have increased. While yields grew more quickly for some crops after 1965, those gains were insufficient to counter the failure of area to grow. Of the five major crops producing faster yield growth in the latter period - maize, barley, rice, sugar, and cocoa - only barley grew more rapidly on aggregate. Of the seven crops with slower yield growth, wheat, sorghum, soybeans, beans, and alfalfa grew more rapidly in the latter period. Changes in the cropping pattern from less to more productive crops were not important. The national cropping pattern, as measured by the shares of different crops in area, changed only slightly from 1950 to 1964 (World Bank, Mexico - Agricultural Technology Sector Review, Report No. 9297, 1992, p. 11).  13 Table 2.1.A Sources of Change in the Value of Output for Maize, Wheat, Sorghum, Barley, Rice, Soybeans, and Samower: 1950-92 Maize Wheat Sorghum Barley Rice Soybeans Satnower Value 47.4 34.8 na 20.9 52.4 na na 1950-58 Area 24.6 8.8 na -4.8 28.0 na na to 1959-64 Yield 22.0 45.8 na 4.7 12.8 na na Price 0.8 -19.8 na 21.0 11.6 na na Value 22.6 8.7 162.1 15.3 20.4 163.1 na 1959-64 Area 16.2 -2.3 150.5 5.5 9.7 173.0 147.1 10 1965-70 Yield 15.0 36.7 26.4 18.2 16.5 -5.2 13.0 Price -8.7 -25.7 -14.8 -8.4 -5.8 -4.7 na Value -16.0 -4.7 65.8 38.4 29.6 102.5 70.2 1965-70 Area -6.9 -10.7 52.3 6.6 15.0 97.2 65.2 10 Yield 4.5 22.9 6.3 45.1 7.4 -7.1 -0.9 1971-76 Price -13.7 -16.8 7.2 -13.3 7.2 12.4 5.8 Value 28.4 6.6 19.4 21.0 -28.8 0.6 14.0 1971-76 Area -5.9 5.8 20.1 5.5 -13.8 19.7 51.9 10 Yield 30.8 14.4 13.7 19.9 20.4 1.2 -20.7 1977-82 Price 3.6 -13.6 -14.5 -4.4 -13.8 -20.4 -17.2 Value 20.0 37.9 31.5 19.6 12.8 38.0 -95.4 1977-82 Area 3.6 29.0 16.2 6.3 0.3 22.1 -44.1 10 Yield 6.8 8.1 1.3 5.0 6.2 1.3 -34.1 1983-89 Price 9.6 0.8 14.0 8.3 6.3 14.6 -17.1 Value I -6.7 -8.1 40.9 13.1 -89.4 19.8 -86.0 1983-88 Area 3.2 -4.0 -9.2 2.5 -37.0 -3.8 -54.7 10 Yield 11.1 -2.2 1.3 -5.8 -20.4 10.2 -12.6 1989-92 Price -21.0 -1.8 33.0 -9.8 -27.5 26.2 -18.6 Value 52.3 29.6 210.1 74.0 19.5 226.3 na Area 7.7 16.9 123.2 16.7 2.7 146.6 118,4 1950-76 10 Yield 55.8 64.6 66.2 66.8 41.0 44.2 10.1 1977-92 Price -11.2 -51.9 20.8 -9.4 -24.2 35.4 na Source: Calculated from World Bank, LA2NR Mexico Agricultural Database.  14 Table 2.1.B Sources or Change in the Value or Output ror Cotton, Beans, Chickpea, Sugar, Coffee, Cocoa, and Alralra: 19S()"92 J~Beam Chickpea Sugar Coffee Cocoa Alralra Value 29.8 75.1 42.1 62.3 21.6 52.8 71.7 1950-58 Area -8.5 24.7 12.7 47.0 39.6 65.5 42.2 to Yield 32.2 24.3 11.6 7.7 1.6 10.1 8.2 1959-64 Price 6.1 26.2 17.8 7.7 -19.5 -22.8 31.3 Value -8.9 21.4 16.1 36.2 18.2 -15.7 41.7 1959-64 Area -25.1 17.9 29.9 31.8 11.4 4.9 25.1 to Yield 23.4 14.3 3.9 10.7 26.3 -4.5 18.5 1965-70 Price -7.2 -10.8 -17.7 -6.3 -19.5 -16.0 -2.0 Value -15.0 22.8 14.8 -1.7 19.9 37.0 23.9 1965-70 Area -44.0 -1.2 8.1 -2.7 6.4 8.2 39.3 to Yield 9.7 14.2 4.8 5.0 4.7 17.1 9.9 1971-76 Price 19.3 9.8 41.9 4.0 8.9 11.7 -25.2 Value -31.9 -6.6 50.4 4.3 18.8 41.6 22.1 1971-76 Area -17.6 0.8 5.1 10.6 15.8 -17.4 21.2 to Yield 9.3 6.3 -3.7 -8.3 -5.3 25.3 10.9 1977-82 Price -23.6 -13.6 48.9 2.0 8.3 33.7 -10.0 Value -55.0 8.2 -84.8 68.2 47.9 -42.0 2.7 1977-82 Area -37.2 6.2 -46.4 -13.4 26.1 7.2 8.8 to Yield 4.9 -6.0 4.7 19.1 35.4 15.3 -24.8 1983-89 Price -22.7 8.0 -43.0 62.4 -13.6 -64.5 18.7 Value 17.5 7.6 22.0 50.5 -165.8 -24.0 na I 1983-88 Area -29.7 -8.9 -0.4 32.3 -3.8 10.5 na to Yield -20.7 6.3 45.9 -19.0 121.9 -8.1 na 1989-92 Price 68.0 10.2 -23.6 37.2 -284.0 -26.5 na Value -70.2 54.1 42.1 99.3 57.3 47.4 52.7 1950-76 Area -100.0 10.8 4.7 36.9 52.1 13.6 68.1 to Yield 34.3 32.3 7.9 10.3 30.7 49.1 11.7 1977-92 Price -4.4 10.8 29.5 52.0 -25.5 -15.3 -27.1 Note: Cotton refers to cotton seed and sugar refers to sugar cane. Source: Calculated from World Bank. LA2NR Mexico Agricultural Database.  15 2.5 Taking the total value of output of the 14 crops for the specified periods, more than 80 percent of output growth in the first two periods was due to the expansion of cultivated area. In the last four periods, the share of area growth in output growth was not greater than one-third, and was much less in three of the four periods. Though the contribution of value per hectare was relatively higher in the last four periods, it was not high enough to maintain output growth at the same level it was from 1950 to 1964. 2.6 Since irrigated area grew more rapidly than total area and achieved higher yields per unit of land, some of the gain attributable to yields reflects a shift of land from rainfed to irrigated. Area expansion was more important in maize, beans, sugar cane, and coffee while yield growth was more important in cotton and wheat. The last two were relatively new crops, wheat being largely unknown before higher-yielding varieties were introduced in the 1950s. 2.7 Input use. Econometric studies, summarized in World Bank (1992), have shown that factor productivity caused little of the previous agricultural growth and nearly all was due to input and primary factor use. In other words, growth came not from more productivity, but from more inputs. 2.8 The contribution of irrigation. Irrigation was a major source of the early post-World War II expansion. Productivity indices for 37 crops from 1946 to 1962 show that cropped area within irrigated districts grew at a compound rate of 8.4 percent and yields at 3.6 percent. Outside irrigated districts, area grew at 1.3 percent and yields at 1.8 percent. The effect of better water supply was reinforced by lower input prices within irrigated districts, apparently because of economies of scale in distribution and better road access. The growth of irrigated area has slowed since the mid-1960s. The estimated annual increment was 198,000 hectares from 1950 to 1965 and fell to 126,000 from 1966 to 1985. Preliminary indications are that it fell to about 84,000 hectares during 1986-88. As important, the growth rate of the yield index on public irrigation districts fell from seven percent before 1965 to less than two percent thereafter. ll. Nominal Protection, Transfers and Price Stability 2.9 Decomposition of changes in prices received by farmers. The issue of importance is the contributing factors of domestic price variability. Domestic price received by producers can be decomposed as: pd = pw e NPC, where, pi denotes the real domestic price of the commodity (Le .• nominal price deflated by the Mexican CPI); pW denotes the real world price of the commodity (Le., nominal world price deflated by the U.S. CPI); and e denotes the real MexicanlU.S. exchange rate (Le., nominal exchange rate multiplied by the ratio of the U.S. over the Mexican CPI). NPC denotes the nominal protection coefficient and explains the part that is not explained by world price and exchange rate. Taking growth rates on both sides of the above expression gives: In(pt) - 'n(p'~1) = [In(Pt) - In(p'~l)] + [In(e,) - In(er _1)] + [In(NPC,) - In(NPCt _1 )]. The three terms in the square brackets of the above expression give the relative changes (from period t-1 to period t) in contribution to domestic price of world price, exchange rate and the NPC.  16 2.10 Table 2.2 presents results regarding the decomposition of the growth of domestic prices due to changes in world prices, exchange rate, and level of protection for five major commodities (Le., maize, rice, sorghum, soybeans, and wheat). Table 2.2 Domestic Price Decomposition ror Five Crops I Maize Rice Sorghum Wheat Soybeans Domestic price -9.3 -6.8 -8.9 -4.6 -22.5 1961-64 World price -6.6 -10.4 -5.3 -6.8 -17.0 to Exchange rate -6.8 -6.8 -6.8 -6.8 -6.8 1965-70 Protection 4.0 11.0 3.1 8.9 1.7 Domestic price -11.9 4.1 3.1 9.5 -17.1 1965-70 World price 27.7 28.6 26.0 30.0 35.3 to 1971-76 Exchange rate -12.8 -12.8 -12.8 -12.8 -12.8 Protection -21.8 11.8 -10.2 -7.8 -39.7 Domestic price 1.8 -22.9 -11.9 -15.5 -10.8 1971-76 World price -30.0 -27.6 -28.0 -19.0 -26.7 to Exchange rate 0.0 0.0 0.0 0.0 0.0 1977-82 Protection 31.6 4.7 16.1 3.6 15.9 Domestic price 9.0 4.9 13.2 11.5 -0.1 1977-82 World price -36.8 -46.7 -41.2 -45.9 -42.2 to Exchange rate 40.5 40.5 40.5 40.5 40.5 1983-88 Protection 5.4 11.2 13.9 16.9 1.6 Domestic price -21.5 -26.8 -31.6 -26.5 -23.5 1983-88 World price -18.6 -9.7 -14.7 -17.6 -12.3 to Exchange rate -17.4 -17.4 -17.4 -17.4 -17.4 1989-92 Protection 14.5 0.3 0.4 8.5 6.2 NOTES: Because of rounding errors the figures may not match exactly. Note that protection (i.e., the NPCs) have been calculated over average prices and exchange rate. Therefore, they may differ from the NPCs reported in table 2.3 below. SOURCE: Calculated from World Bank, LA2NR Mexico Agricultural Database. 2.11 A simple partial measure of resources transferred by policy distortions is the NPC, defined as the ratio of domestic to world price for a given good. The domestic price is typically the average annual market price received by the producer. The world price is defined as an import or export  17 parity price, adjusted for relevant transactions costs and converted to domestic currency at the market exchange rate. Table 2.3 summarizes average NPCs for the chief commodities: maize, wheat, rice, barley, sorghum, soybean, coffee, cocoa, and sugar cane. In 1992, those crops accounted for 75 percent of the planted area and for 48 percent of the value of all crop (annual and perennial) output in domestic prices. Figures 2.1.a, 2.1.b, and 2.l.c depict the NPCs for six annual and three permanent crops. Table 2.3 Nominal Protection Coefficients for Nine Major Crops (Annual Values Averaged by Sexenio) I I Maize Wheat Sorghwn Barley Rice Soybeans Sugar Coffee Cocoa 1950-58 0.85 1.06 na 0.57 na na 0.45 0.75 0.66 1959-64 1.16 1.18 0.99 0.90 0.45 0.88 0.62 1.20 0.91 1965-70 1.24 1.15 0.88 1.04 0.51 0.95 1.19 1.08 0.95 1971-76 1.05 0.86 0.87 0.81 0.49 0.92 0.42 1.00 0.87 1977-82 1.36 0.93 0.99 0.95 0.48 0.93 0.64 0.73 0.79 I 1983-88 1.45 0.92 1.14 1.02 0.52 1.08 1.17 0.81 0.52 1989-92 1.72 1.19 1.15 1.46 0.53 1.16 1.19 0.46 1.06 1950-70 1.05 l.ll 0.94 0.80 0.40 0.92 0.71 0.97 0.82 1971-92 1.37 0.95 1.03 1.02 0.50 1.01 0.83 0.77 0.79 NOTES: -na - indicates that the respective figure is not available. Because of rounding errors, the figures m,? not add up exactly. Sugar, coffee and cocoa refer to primary products. ,~ ). i<J"'. / ; SOURCE: World Bank, LA2NR Mexican Agricultural Database. tl/" ,'J l\, .> . ~ J~ A}'1 b~ . ~ r f' rJ' \' v;;t <.t; 2.12 Policy was neutral, with one exception, in the earlier period. The ¢cception was rice, which was taxed at a rate of more than 50 percent. Policy grew more protectionist from 1970 to 1988 forJ X maize, sorghum, barley, and soybeans and less so for wheat. 11 There was little change in the protection to rice, which remained heavily taxed. The permanent crops - sugar cane, coffee, and cocoa - were heavily taxed in all periods. JX 2.13 Resource Transfers. The resources transferred into or out of agriculture for a particular commodity are calculated by valuing the production of the crop at the difference between domestic '\ and world price, properly adjusted by transportation costs, as follows: Transfers = q1.(pw + ne - pd) where II is domestic price of the crop, p'" is the world price of the crop, T denotes transportation costs, and e is the exchange rate; if is the quantity produced. Total potential transfers are the sum of individual transfers. For example, during 1989-92, annual average potential transfers due to output price distortions were N$4,265 million (in 1992 pesos). One has to take into consideration the fact 11. Wheat producers began to receive direct transfers from the federal government in 1990, however.  18 Figure l.1.a: NPCsfor j1;laize, Wheat, an'! Barley 2.00 I--------------------------~ , ;. .. .. .. " .. . . .. . . . . . .' . 1.00 i • I,·" 0.50 1 I I i I· .'" ... Maize - - - - ""neat - - - Barle-I I • r 0.00 ..;-'_i_-:--,..-:---'--:-----:--:--~_,__'__:___:__:___:___:__:__:__,__:__:___:__:__i__:__,_.,_;_~_,_~-__, 1950 1955 1960 1965 1970 1975 1980 1985 1990 Figure 2.1.b: NPCs for Sorghum, Rice, and Soybeans 2 ---------------------------------------------------------------- • • • . . . .. Sorghum - - - - - - Rice - - - Sovbe:ms . I I 1.5- i 0.5 I I I o ~I---~_r~~---'-_:__:_~~.,_;__i_~~~_:_~~~~_i_~~~~_:_~- 1951 1956 1961 1966 1971 1976 1981 1986 1991  19 Figure 2.1.c: NPCsforSugar, Coffee, and Cocoa 2..:5'0 I I 2.00 I T I· .. ·· ... Sugar Coffee Cocoa : .", 0 1..:5'0 'I . ..... ....... " , 0 . . 1.00 T 0..:5'0 1. .... - -, - . I i .. 0' - , .. .. I -, ! 0.00 I i : ! 1950 1955 1960 1965 1970 1975 1980 1985 1990 •  20 that not all quantities produced are traded. This is especially the case for maize, where it is estimated that only ha1f the production is commercia1ized. Adjusting for non-commercia1ization, the actua1 transfers due to price distortions were about N$I,858 million (in 1992 pesos). In genera1, policy taxed exportables more than importables, as seen in the stiff tax on coffee and cocoa and the relatively mild subsidy to importable cereals except maize. Table 2.4 Potential Transfers to Agricultural Producers for Nine Major Crops (Annual Values Averaged by Sexenio: 1992 N$ millions) . so~.... " ~I H ....... Rice 1950-58 -796 67 na -107 na na -1,083 -367 -70 1959-64 1,023 278 10 19 -495 na -1,384 215 -24 1965-70 849 269 841 103 -466 -5 248 121 -24 1971-76 281 -435 -338 80 -715 -101 -4,669 83 -51 1977-82 2,699 -206 15 21 -566 -76 -1,831 -964 -177 1983-88 3,591 -253 471 6 -534 100 -1,969 -1,109 -232 1989-92 4,814 375 328 116 -219 155 -541 -771 8 1950-70 1971-92 f;358 2,846 -130 205 416 112 -76 5 -48 -508 na 19 -740 -2,253 -10 -690 -39 -113 NOTES: "na" indicates that the respective figure is not available. Because of rounding errors, the figures may not add up exactly. Sugar. coffee and cocoa refer to primary products. SOURCE: World Bank. LA2NR Mexican Agricultural Database. I 2.14 Regional effects. The regiona1 effects of protection harmed the distribution of income, for several reasons. The first is that coffee and cocoa are grown in the poorer southern states and are pure exportables, so there were no compensating income gains to producers who are a1so consumers \ of those commodities. A second is that poor producers of exportables are consumers of importable / cerea1s, mainly maize, which became more heavily subsidized to producers; poor rura1 producers of / exportables had little or no access to consumer food subsidies designed to compensate for the high ~ ' producer price of maize. A third reason is that exportables, and sugar which was both an exportable and an importable, are more labor intensive than the importable cerea1s and oilseeds; hence the xation on exportables reduced demand for labor in those crops. 2.15 Price stabilization. One possible justification for government interference in agricultura1 markets is to increase price stability and thereby to reduce risks to producers associated with price variability. Has policy in fact increased producer price stability? The question is specifica1ly formulated as: Were Mexican domestic producer prices more stable than world prices in the periods under analysis? We answered this question by taking the time series of rea1 domestic prices paid to producers of the nine chief commodities, and of world prices for the same commodities in the same periods. 2.16 There are severa1 possible measures of variability. A problem with price series is that they sometimes do not satisfy the statistica1 properties (i.e., stationarity) required to ca1culate standard statistical measures such as the mean and the variance. In order to compare domestic and world price  21 variability we used the Z-statistic, which is very similar to the standard deviation. If P, denotes the world price of a particular commodity at time t, the measure of variability, Z, is defined as follows: t 0.5 (Pr - Pr_l)2 Z = [ r-2 --n----".1--] The main difference from the variance is that instead of taking the deviation of p, from its mean value we take the deviation from its previous value (i.e., P,-J)' Notice, however, that even if P, is nonstationary, its first difference, (p, - P,-J)' is stationary, thus making the Z-statistic a valid measure of price variability. Table 2.5 Ratio or Domestic to International Price Variability Maize Wheat Sorghum· Barley Rice Soybeans" Cocoa 1950-70 0.66 0.96 1.15 1.03 0.21 0.62 0.38 1971-92 0.58 0.41 0.99 0.47 0.31 0.69 0.36 1950-92 0.62 0.58 1.07 0.59 0.27 0.76 0.37 NOTE: oJ For sorghum and soybeans the starting year is 1958 and 1960, respectively. SOURCE: Calculated from World Bank, LA2NR Mexico Agricultural Database. 2.17 The Z-statistic was calculated for both domestic and world prices of: maize. wheat. sorghum, barley, rice, soybeans, and cocoa. Results are reported in Table 2.5. A ratio less than 1.00 suggests that domestic prices have been more stable than world prices and vice-versa. On the average over the 1950-92 period, domestic prices have been more stable than world prices for all crops except sorghum. This suggests that fuJI liberalization, and hence fuller transmission of world prices to domestic, will make Mexican producer prices less stable for maize, wheat, barley, rice, soybeans, and cocoa. The appropriate policy responses to that are discussed in Chapter 3. III. Direct and Indirect Effects or Output Price Distortions 2.18 NPC is the directly observed component of policy distortion and as such does not capture distortions due to supply changes (Le., supply changes that would have occurred had producers faced world prices). Therefore, to fully account for distortions one should use a more general measure of protection. In this section, we construct an index which captures two additional distortions: supply changes and exchange rate misalignments. 2.19 Anderson and Bannister (1993) have introduced the Trade Restrictiveness Index (TRI) to extend the Producer Subsidy Equivalent (PSE) to include supply changes. While such an index would be ideal to be applied in the current model, because Mexican agriculture is characterized by many policy reversals (i.e., changes from tax to subsidy and vice-versa) we are not able to apply such an index. Instead we will apply an index which, while it is superior to the nominal protection index  22 since it accounts for changes in quantities through elasticities, it is not directly derived from an optimization model as the TRI. 2.20 We calculate the Restrictiveness Index (Rl) as follows. Let [I and p'" denote the domestic and world price of the commodity in question. Further, let if be the quantity produced when producers face prices [I. Now let q'" denote the quantity that would have been produced if producers faced prices p"'. Then, define the RI at period t as: From the components involved in the above expression we observe [I, p"', and if but not q"'. However, qW can be calculated if the supply elasticity, E" is known. 2.21 The percentage deviation of if from tf is equal to the percentage deviation of [I from p"', properly adjusted by the elasticity of supply, Le., (if - q"')/if = E.(pI- p"')/pd, Upon rearranging terms and solving for qW, the above expression yields: q'" = qd [1 _pd;d Pw Es] , If the world price of the commodity is equal to its domestic price (i.e" no restrictions), then if = q"', which in tum implies that RI = 1. If E, = 0, then the quantity of the commodity will be the same produced under the two prices; the latter scenario is equivalent to the nominal protection coefficient measure of restrictiveness. Values of RI greater than unity imply subsidy while values of RI less than unity imply tax. 2.22 For certain configura~ions of prices and supply elasticities, the RI as defined above may take negative values. For example, let f, = 2 and also let the domestic price exceed more than twice the corresponding international price; then the RI becomes negative. However, this is not a deficiency in part of the index, but simply reflects the fact that the supply curve intersects the price axis in the segment defined below the domestic price and above the world price. 2.23 A numerical example gives an intuitive interpretation of the RI. If the world price is $50/ton, the domestic price is $loo/ton (Le., a $50/ton subsidy), and production is 200 tons. That gives a protection index of 100150 = 2. Assume a supply elasticity of 0.6. While at domestic prices producers supply 200 tons, at world prices they will supply 200*(1-0.5*0.6) = 140. Therefore, the value of production under world prices would have been 140*50 = $7,000. Note that the value of production under domestic prices is 200*100 = $20,000. That implies a restrictiveness index of 20,00017,000 = 2.86, which is higher than the index of 2 derived earlier. 2.24 The index would be the ratio of the domestic to world prices, if quantities under domestic and world prices were equal. However, the value at world prices considers the quantity that producers would have produced if world prices prevailed in the domestic market. Therefore, the restrictiveness index depends not only on the price differential but also on the supply elasticity of the commodity in question. Hence, commodities with high supply elasticity present a high restrictiveness index, even if they exhibit small price differentials.  23 2.25 The next question relates to the exchange rate. To understand the extent to which the level of exchange rate affects the Rl, consider the following extension of the above example. Suppose that the exchange rate is overvalued, so that while the nominal ratio is N$3 to one SUS, in fact it should have been N$4 to one SUS. Then, the N$300/ton domestic price becomes $75/ton -- a $25/ton subsidy - which in tum implies that the domestic value of production equals 200*75 = $15,000. Therefore, = the restrictiveness index becomes 15,000/6,000 2.5. From the above example, one can see how sector specific as well as macroeconomic distortions may affect the level of protection of one commodity and consequently the performance of the sector. Below, we specify an econometric model which determines the degree of misalignment of N$. 2.26 Exchange Rate Misalignment. We evaluate the degree to which the N$ is misaligned with respect to the US$ by specifying the following econometric model (Edwards, 1989): 10g(REER) = p. + P,log(T01) + Pj?ESGDP + f. The definition of REER is given by the ratio of domestic CPI to the U.S. WPI, adjusted by the corresponding nominal exchange rate. TOT denotes the terms of trade and is defined as the ratio of export to import price index. Finally, RESGDP denotes the resource balance (Le., value of exports minus value of imports) to GDP ratio. TOT is expected to have a positive effect on REER as a worsening TOT call for depreciation of the equilibrium REER. On the other hand RESGDP is /") expected to have a negative effect on REER as a widening gap between exports and imports calls, ;1 again, for REER depreciation. 2.27 We estimated the model using annual observations from 1960 to 1992. The results are as follows: 10g(REER) = 1.40 + 0.38 10g(T01), - 2.96 RESGDP (111.6) (3.95) (-8.08) R2 = 0.80, DW = 1.38, ADF = -3.45, PP = -3.97. Numbers in parentheses denote I-values. DW is the Durbin-Watson statistic; ADF and PP are the Augmented Dickey-Fuller and Phillips-Perron tests for stationarity of the residual. The market integration section further elaborates on the estimation procedure of these type of models. 2.28 To see the effect of exchange rate, consider the case of maize. Its average 1991 world price was $107.2/mt and the domestic guarantee price was N$707.39/mt or $234.27/mt at the official exchange rate. For the domestic price to become equal to the US Gulf price a 54.22 percent reduction should have taken place. The corresponding prices for wheat $116.2/mt (world) and $194.05/mt (domestic) and sorghum were $105.3/mt and $142.3/mt. 2.29 Now consider Table 2.6 which reports sexenio averages of the following three measures of trade distortions: (a) nominal protection coefficient (NPC), which is the ratio of domestic to world price; (b) restrictiveness index (Rl); and (c) the generalized restrictiveness index (GRl), which also has taken into consideration exchange rate disequilibrium effects. The commodities examined are maize, rice, sorghum, soybeans, and wheat. 2.30 It was assumed that the exchange rate was in equilibrium in 1989. The equilibrium and·1 ."1~ market exchange rates were almost equal in 1985. There was a slight undervaluation in 1988 and a ,.l relatively large overv.aluation in 19~1. As a c~nsequence, one observ:s.a considerabl~ difference 0Y t1vJ between Rl and GR! 10 1991, mean10g that while producers were SUbSIdized through high output \\ J .\~., . V f ,,.1' \ !' l' ,.-" ,  24 \prices, at the same time they were taxed through the overvalued exchange rate. In what follows, '-same specifics regarding the five commodities examined will be given. 2.31 Maize. Domestic guarantee prices were compared to white maize U.S. Gulf fob prices. From the information on NPC based on guarantee and yellow fob price it can be seen that domestic production has been considerable protected, especially since 1976 with NPC ranging from 50 percent to 100 percent. The higbest NPCs are observed in 1983-88 and in 1989-92. When taking into consideration the output effect, the protection becomes even bigber. For example, in 1989-92 the index went from 1.92 to 1.97. However, if one adjusts for exchange rate disequilibrium, the index reduces to 1.97, still implying an almost 100 percent protection. Table 2.6 Restrictiveness Indices for Five Major Crops Maize Rice Sorgbum Soybeans Wheat NPC 1.33 0.46 1.00 0.99 1.27 Rl 1.39 0.29 0.99 0.99 1.36 1961-64 GRl 1.20 0.25 0.86 0.86 1.17 NPC 1.37 0.49 1.02 1.10 1.27 Rl 1.45 0.32 1.06 1.12 1.36 1965-70 GRl 1.23 0.28 0.90 0.95 1.16 NPC 1.14 0.47 0.96 1.04 0.91 Rl 1.17 0.31 0.94 1.05 0.89 1971-76 GRl 0.92 0.24 0.73 0.82 0.71 NPC 1.55 0.45 1.10 1.06 0.99 Rl 1.66 0.29 1.27 1.08 0.99 1977-82 GRl 1.38 0.24 1.05 0.90 0.83 NPC 1.69 0.51 1.32 1.28 1.02 Rl 1.84 0.35 2.21 1.34 1.02 1983-88 GRl 2.12 0.39 2.59 1.54 1.15 NPC 1.92 0.52 1.30 1.46 1.39 Rl 2.12 0.35 1.93 1.57 1.52 1989-92 GRl 1.97 0.33 1.81 1.46 1.41 NOTES: The supply elasticities with respect to price for maize, wheat, and soybeans are those reported in Table 2.8. For soybeans the elasticity of beans was used (equal to 0.227), The rice elasticity was set equal to 0.5, the average of the other four elasticities. Calculations are from World Bank, LA2NR Mexico Agricultural Database.  25 2.32 Rice. From the five commodities examined, rice is the only commodity which exhibiting taxation throughout the period examined. Consistently, Mexican rice producers were receiving half of what their U.S. counterparts were receiving, when one compares domestic price with the US Gulf fob. In particular, the NPC ranged from the lowest average of 0.45 in 1977-82, to the highest average of 0.51 in 1983-88. Accounting for production and exchange rate effects the lowest and highest values of the index become 0.24 and 0.39 in 1971-76 and 1983-88, respectively. 2.33 Sorghum. Until 1984 producer prices for sorghum were close to their world level counterparts (Le., U.S. Gulf, fob). Since 1985, however, producer prices have been, on the average, 40 percent higher. When one takes into account the output effect sorghum exhibits a restrictiveness index which goes from 1.32 to 2.21 in 1983-88 and to 1.59 when the exchange rate effect is taken into consideration. 2.34 Soybeans. Domestic soybean prices have closely followed world prices, although in the last seven years producers were receiving an average of 40 percent more, when compared to U.S. Gulf fob prices. As Table 2.6 indicates, the NPC exhibited an average of 1.46 in 1989-92. Accounting for the production effect the index increased to 1.57, but the exchange rate effect canceled the production effect so it went back to 1.46. Because of the low elasticity of supply, the Rl was higher by only a small margin. 2.35 Wheat. Comparing domestic wheat prices to the U.S. Gulf for prices reveals that wheat has not been greatly protected. In the period 1989-92 it exhibited very similar pattern to that of soybeans. The three indexes were 1.39, 1.52, and 1.41 in 1989-92. IV. Other Transfers 2.36 A more complete measure of resource transfers is given by the producer subsidy equivalent (PSE). The PSE is defined as the sum of transfers from all policy measures to a given crop. It measures, in principle, the net incentive effect of policies on farm inputs and outputs. The main obstacle to extending the PSE fully across commodities is the lack of data on subsidies/taxes to producers for inputs. A set of the most important inputs subject to government intervention includes water, fertilizer, finance, machines, non-fertilizer agricultural chemicals, seeds, and crop insurance. We have reasonably complete data only for fertilizers. 2.37 Fertilizer subsidies. The Mexican government subsidized fertilizers for many years. It has only been since 1990 that the domestic price has been equal to or greater than the border price equivalent. Table 2.7 shows the evolution of world prices of urea and phosphate, which are the main materials used. The NPC (Le., the ratio of domestic to world price) as well as the transfers to agriculture due to fertilizer price distortions are also reported. 2.38 The world price of fertilizer materials is reported in current US$ and refers to physical quantity of fertilizer. The domestic price is expressed in N$ and refers to Nand P content. To calculate the nominal protection coefficient, the world prices were converted to domestic by adjusting by the relevant nutrient content and adding international and domestic transportation costs. In particular, the fertilizer NPCs have been calculated as follows: NPC :; PN d and pt NPCp :; - - - - - - N (P;/0.45 + 1) e (P;/0.46 + 1) e  26 where JI and pW denote domestic and world prices; subscripts N and P denote nitrate and phosphate content fertilizer; T denotes the sum of international and domestic transportation costs; e is the nominal US/Mexico exchange rate. The transfers to/from agriculture have been calculated as: Transfers - [[ :':5 • +-p~] qN· [[ :';5 • +- p;] qp. qN and qp are the quantities of nitrate and phosphate fertilizer consumed. 2.39 Other transfers. Other transfers include water subsidies (discussed in Chapter 4), financial transfers (to be fully discussed in Chapter 6) and those for machinery, non-fertilizer chemicals, and seeds, whose values could not be estimated. It is not possible to quantify those transfers in a way that would allow a definitive calculation of a PSE. Transfers for machinery and seeds have been eliminated, however. Table 2.7 World Prices, NPCs, and Transfers to Agriculture: Urea and Phosphate Fertilizers, 1965-92 D 1965 i World Price (Current US$/mt) Urea 90.5 Phosphate 47.3 Urea 1.16 NPC Phosphate 1.62 Transfers to Agriculture (million 1992 N$) -208 1970 48.3 42.5 1.78 1.76 -638 1975 197.7 205.0 0.57 0.84 1,212 1980 221.9 178.0 0.61 0.89 974 1985 136.3 121.4 0.70 0.90 726 1986 107.0 121.2 0.63 0.66 1,184 1987 117.1 138.3 0.56 0.57 1,507 1988 155.0 158.4 0.61 0.69 1,013 1989 136.8 144.5 0.76 1.24 358 1990 157.0 131.8 1.07 1.59 -473 1991 172.0 133.1 0.95 1.34 -46 1992 140.3 120.7 na na na SOURCE: World prices of fertilizers are from the International Monetary Fund (IMP) (various issues). Domestic prices are from the FAO Fertilizer Yearbook (various issues). Calculations are from World Bank, LA2NR Mexico Agricultural Database. 2.40 Factor price effects or policies. We Jack good data on the factor price effects of government policies. The main effects have been on the opportunity cost of land, through the historical constitutional restrictions on the land market, and on wages, through the agricultural exemption from  27 p!~Q[sQcial~t~. The former lowered the opportunity cost of land by restricting its reallocation to other uses, either more productive agricultural ends or non-agricultural purposes. This lower opportunity cost of land was a transfer to producers that we cannot quantify. The social security exemption lowered the price of labor to those farmers large enough to hire workers and also constituted a transfer that we cannot quantify. v. Effects of Policy Reforms on Profitability of Individual Commodities 2.41 Bank sector work previously attempted to estimate a long run growth function for Mexican agriculture (see World Bank, LA2AG, "Impact of Policy Reforms on Agricultural Growth". October 1991). The main finding was that the deterioration in the agricultural terms of trade to a historicall low level in 1987 was the reason for the extremely poor performance of the sector in 1988 while steady improvement in the terms of trade since 1987 has enhanced sectoral growth. Furthermore, the J study found that considerable increases in agricultural credit disbursement in 1984 and 1988 had a strong impact on real agricultural GDP. 2.42 A more limited effort was undertaken to evaluate the effects of policy reforms on the profitability of individual commodities using a profit function approach ("Price Responsiveness and Structural Adjustment in Mexican Agriculture, LA2AG Mimeograph, September 1993). The II approach allows estimates of the elasticities of supply of the commodities, the elasticities of demand for factors and inputs used in production, and of the effects of policy reforms. The commodities examined were maize, beans, wheat, and sorghum. The inputs were labor and mineral fertilizer. The data cover 1961 through 1991 with a dummy variable for the period after 1986, to capture the effects of the reforms. We first present the elasticities (Table 2.8) and then the potential effects of price liberalization (Table 2.9). 2.43 The own-price supply elasticity of maize is 0.19 indicating that a 10 percent drop in the price of maize will lead to an almost 2 percent decline in supply of maize. All three cross-price supply elasticities are positive though very small in magnitude. The effect of the wage rate on the supply of maize is negligible. The price of fertilizer has the largest effect on the supply of maize; its value is - 0.323. The relatively low own-price supply elasticity of maize may reflect the fact that this commodity is mainly for home consumption, so that price changes affect only the quantities which are traded. ~ -:tr~ j'(~ i "'f~ j.)t..it-t./ jJ~'rsI., S "-3 S' H - ..,. ~able 2.8 Input and Output Price Elasticities ~ ~ () ,~ n ........... 111,,1l.. I-I'~""'" r ~"- Cl \ r-9 0_191 0'090 002R 0013 -0_232 0001 _WheaL 0.352 0.315 -0.019 -0.073 -0.577 -0.001 .1. St -0.598 -0.217 1.401 0.071 -0.640 -0.017 Beans 0.011 -0.087 0.143 0.227 -0.293 -0.002 Fertilizer 0.459 0.230 0.244 0.221 -1.025 -0.003 Tnhnr ..0 14\1 .0 OQQ o 7?1 o OR! ..0 11\7 ..0 014\ SOURCE: Baffes. 1993a.  28 2.44 Supply of wheat exhibited a somewhat higher own-price elasticity (0.315), implying that a 10 percent decline in the price will induce an almost 3.2 percent reduction in supply. A relatively high cross-price elasticity was observed with respect to maize (0.352). The cross-price elasticities with respect to sorghum and beans are negJigible as in the case of maize. The elasticity with respect to the price of fertilizer is -0.577, which is the second highest fertilizer elasticity observed among the four crops. Again as was the case with maize, the elasticity with respect to the wage rate was very low. 2.45 Beans exhibited an average own-price elasticity of supply of 1.401, which is the highest observed among the four crops. The price of sorghum as well as the wage rate did not show a significant effect on the supply of beans. However, a 10 percent increase in the price of maize wiJI induce a 6 percent reduction in the supply of beans as the value of -0.598 indicates. The elasticity with respect to the price of fertilizer is -0.640. As before, wage rate did not have a significant effect on the supply of beans. 2.46 Sorghum exhibited an own-price supply elasticity of 0.277. The prices of maize and wheat had no significant effect on the supply of sorghum as their elasticities were close to zero. Only the elasticity with respect to the price of beans is somewhat significant; its value is 0.143. The elasticity with respect to the price of fertilizer is -0.293. Wage rate had no effect on sorghum supply. 2.47 The prices of all four crops positively affect the demand for fertilizer. That is, an increase in the price of the four crops induces higher demand for fertilizer. The own-price demand elasticity for fertil izer is -1.025. The price elasticity with respect to wage rate is very low. The own-price elasticity of labor is very low (-0.015). The prices of maize, wheat, and fertilizer negatively affect the demand for labor while the price of beans and sorghum positively affect it. The effects of wheat and beans prices are negligible. 2.48 If one excludes the early period for beans, own-price supply elasticities for all crops exhibited relatively low values. The wage rate has a negligible effect on the choice of both output and input mix including the demand for labor itself. The own-price elasticity of fertilizer slightly exceeded unity; furthermore, all four input prices had a significant effect on the choice of the level of fertilizer. The demand for labor is negatively affected by the price of maize and fertilizer while it is positively affected by the price of sorghum. 2.49 FinaJly, an attempt was made to capture any effects resulted from the structural changes ~l cY) introduced in the mid-1980s. The results showed that there is a structural reduction in total profits ~~ / due to the changes, that is, the liberalization of trade policy reforms, combined with the elimination '~/ of heavy fertilizer subsidies reduced profits and output in the sector. ,~/ 2.50 What would happen if output prices move to international levels? Table 2.9 reports three simple simulations based on the elasticities presented in Table 2.8 (the first three rows refer to the hypothesized scenarios while the last three rows refer to simulation results). The following three scenarios are considered: (I) full liberalization of output prices is assumed implying a 70 percent reduction in the price of maize, a 30 percent reduction in the price of wheat, and a 15 percent reduction in the prices of sorghum and beans; (lJ) assumes partial liberalization on maize (35 percent price reduction) and full liberalization of the other three crops as in scenario I; and (III) assumes partial liberalization in all three crops, i.e., 35 percent reduction in the price of maize, 15 percent reduction in the price of wheat, and 10 percent reduction in the prices of sorghum and beans. It is assumed that no further change will take place in input prices (as the price of fertilizer is close to the  29 " vV,V'-lc- {, l.\." .,,1 .'.' 'j; l.)fi{.': J t t.u' JI .I L t n ((1' international levels while the wage rate, which is competitively determined, is not expectJ to decline because it is already so low). :) l l(~J 2.51 As simulation I shows, full liberalization of output prices will induce a 20 percent reduction in the supply of maize, a 37 percent reduction in the supply of wheat, a 29 percent reduction in the supply of sorghum, and a 9 percent reduction in the supply of beans. On the input side, the use of fertilizer will decline by 57 percent while the use of labor will go down by 10 percent. It is interesting to observe that under the full liberalization scenario, the supply of maize is not affected to a great extent while the use of fertilizer declines by more than half, a result which was expected given the low own-price supply elasticity of maize. The labor requirements of the sector will be reduced by 600,000 workers. 2.52 Under simulation II, while the price of maize declines by 35 percent (half of the reduction required to be equated to the international level) all other prices are assumed to move to their international level counterparts. As expected, the resulted changes in quantities are less than the ones observed in simulation I. Supply of maize will decline by 13 percent while the corresponding declines for wheat, sorghum, and beans are 23 percent, 18 percent, and 7 percent respectively. Fertilizer and labor use will decline by 34 percent and 8 percent, respectively. Table 2.9 Changes in Inputs and Outputs In Three Scenarios (percentage change from base value) Maize Wheat Sorghum Beans Fertilizer Labor I. -70 -30 -15 -15 0 0 Assumed II -35 -30 -15 -15 0 0 Scenario III. -35 -15 -10 -10 0 0 I. -20 -37 -29 -9 -57 -10 Simulation II. -13 -23 -18 -7. -34 -8 Results III. -10 -17 -15 -6 -29 -7 SOURCE: Baffes, 1993a. 2.53 In simulation III partial liberalization in all four crop prices is assumed. Under this scenario, supply of maize will decline by 10 percent while the supply of wheat, sorghum and beans will decline by 17 percent, 15 percent, and 6 percent respectively. Finally, fertilizer use will go down by 29 percent while labor use decline down by 7 percent. VI. The Effects of NAFT A and PROCAMPO 2.54 The NAFTA and PROCAMPO, the new rural income support program, will affect Mexican agriculture in many respects. Terms and expected effects of NAFTA are discussed in the following section while the subsequent section describes PROCAMPO.  30 A. NAFTA 2.55 Real agricultural exports have not grown in a sustained way in the last 40 years (fable 2.10). The average annual growth rate of Mexico's agricultural exports has not been statistically different from zero since 1950. The composition of exports has shifted from coffee, cotton, and sugar to coffee, fruits and vegetables, and livestock; cotton exports have become uncompetitive and have practically disappeared. Table 2.10 Period Averages of Agricultural Exports Share of commodity group in total agricultural Agricultural exports exports (percent) Annual Fruits, average Grains, vegetables, N$ 1992 growth rate legumes, tobacco, Coffee, Animal billion (percent) oilseeds cotton sugar products 1950-58 8.26 3.8 2.5 62.3 26.6 8.6 1959-64 8.71 2.8 4.1 51.6 28.4 15.9 1965-70 I 9.42 -2.0 11.6 46.2 26.1 16.2 1971-76 ' 8.14 1.0 1.8 44.8 34.0 19.3 1977-82 8.16 -1.5 1.1 44.6 28.9 9.4 1983-88 8.66 16.1 1.1 37.4 44.0 17.5 1989-92 7.19 -5.2 na na na na Source: World Bank, LA2NR Mexican Agricultural Database. 2.56 Mexican agricultural exports have varied between US$2 and 3 billion in the last decade and averaged 6 percent of total non-oil exports from 1990-92, compared to 7.7 percent of total exports from 1980-88. The value of agricultural exports was 26.9 percent of that of manufactured exports in 1980-88, a ratio which had declined steadily since 1980 (Nacional Financiera, La economia mexicana en cifras, 1990, p. 689). The 1990-92 averages (US$ million) were: total, 2,216, of which coffee, 320; vegetables, 490; cattle, 346 (Group Financiero Bancomer, Mexico - Key Indicators, 1993). Mexico's net agricultural exports to the United States by category ,(US$ million) were: grains and oilseeds, -358; livestock, meat and dairy 8; horticulture 95 (losting, p. 158; and SARH, EI sector '.,[ agropecuario en las negociaciones del tratado de Iibre comercio Estados Unidos - Mexico - Canada). _,j" US exports to Mexico are chiefly cereals (maize, sorghum, wheat) and oilseeds (soybeans and their " ~' derivatives). These relations have been reasonably stable for 20 years, with the exception of Mexican . "':/ rlive cattle exports which have risen after 1989 consequent to the lifting of the Mexican export tax. v'" J\ While the overvalued exchange rate has harmed agricultural exports, it cannot explain why ;; agricultural exports have grown much slower than non-oil manufactured exports, which presumably '-J • would have been affected in a similar fashion by the exchange rate. ...- ~ ,j~ 2.57 The NAFfA agreement between Mexico and the United States defines five categories of '~ goods according to the speed of liberalization (fable 2.11).  31 Table 2.11 Percent of 1990 Exports Affected by NAFT A II Mexico to US US to Mexi A Immediate liberalization 54 31 B Free within 5 years 9 17 C Free within ten years 23 32 C+ Free within 15 years 0 1 D Already liberalized 14 18 SOURCE: SARH, El sector agropecuario en las negociaciones del tratado de libre comercio Estados Unidos-Mexico- Canada. Table 2.12 Pre- and post-NAFrA Trade Policy for 12 Major Crops PRE-NAFTA STATUS NAFTA TARIFF & QUOTA STATUS QRs Tariff Duty Free Above Quota Tariff Maize. grain Pennit 0% 2.5 million tons -- 215 % tariff reduced by 24 % over High guarantee price 3 % annual increase the first 6 years, then gradually over 15 years reaching zero by year 15 Beans Pennit 0% 0.05 million tons .- 139 % tariff reduced by 24 % over High guarantee prices 3 % annual increase the first 6 years, then gradually over 15 years reaching zero by year 15 Sorghum No pennit 15 % seasonal No quota Immediate elimination of tariff Soybeans No pennit 15 % seasonal No quota Immediate reduction of tariff to 10 %, then phased out in 10 years -- reduction of dutiable season Wheat, grain Pennit 10% on durum wheat; No quota 15 % common tariff to be phased substantial direct out over 10 years payments to wheat producers Wheat, flour No permit 15% No quota Long-term liberalization Rice No pennit 20 % on brown & No quota Current tariffs phased out over 10 kernel milled rice & years 10 % on rough & broken rice Barley Pennit 5% 0.12 million tons -- 128% tariff gradually reduced High domestic agreed 5 % annual increase over the first 6 years, then price over 15 years reaching zero by year 10 Oats No pennit 10% No quota Tariff phased out over 10 years Cotton, raw No pennit 10% No quota Tariff phased out over 10 years & fiber Cottonseed No pennit 15% No quota Tariff phased out over 10 years Cake Safflower No pennit 10% seasonal No quota NOTES: The above quota tariffs for maize. beans. and barley are in accordance to the Uruguay round agreement of GAIT.  32 2.58 Table 2.12 summarizes the pre-and post-NAFTA trade policy arrangements for 12 major crops. r- -- 2.59 Effects of NAFfA. This section analyzes the possible effects of NAFTA. Given the associated reforms in Mexican agriculture, we may assume that the various "free trade plus Mexican liberalization" scenarios are the most probable. However, even those most probable scenarios do not incorporate a dynamic output response from technical change or additional investment within Mexico. The only existing study with such a dynamic response is that of Levy and van Wijnbergen (1991) which examined the welfare effects of combinations of free trade and complementary public investment policies. 2.60 The Levy-van Wijnbergen model separates farmers into four categories: (a) subsistence farmers; (b) landless rural workers; (c) rainfed farmers; and (d) farmers owing irrigated land. The model also makes the assumption that the government undertakes adjustment polices. such as additional investments in rural areas, principally in irrigation. Table 2.13 summarizes the welfare effects of price liberalization consistent with NAFT A. One critical assumption relates to the degree to which the U.S. will liberalize its imports of fruits and vegetables. The model assumes a base welfare level of 1.000. Simulation results indicate that, regardless the assumptions of the model farmers using irrigated land always gain. If compensation is given, subsistence farmers and irrigated farmers will gain. In terms of efficiency gains, it is expected that they will be U.S .$1.2 billion per year or almost 0.40 percent of 1990 GDP. 2.61 The consensus of available studies is that the Mexican economy will gain from NAFTA (e.g., Brown in Lustig, Bosworth, and Lawrence, 1992). The effects of NAFTA on Mexican agriculture have also been studied extensively and there is also a consensus about the direction of the effects on sectoral product, employment, and trade, if not about specific magnitudes. The agreement initially favors Mexico as more of its current products receive immediate liberalization (category A) or are free within five years (category B). Josling (1992) has reviewed two types of studies, general equilibrium and commodity, and the following is based on his account. 2.62 Various general equilibrium models have been applied to NAFT A. They share common defects of being aggregate, derived from earlier data and based on reported covenants of the treaty, which may have changed since the model results were developed and published. Another common difficulty is that the model writers were forced to develop scenarios based on assumptions about Mexican policy liberalization independent of NAFTA, with respect to guarantee prices, official commodities purchases, input subsidies, and credit allocations. The contrasting results of various models can therefore derive from differing levels of aggregation, data periods, and assumptions about Mexico's domestic policies.  33 Table 2.13 Welfare Effects and Efficiency Gains from Maize Price Liberalization Price Liberalization. Price Liberalization without Compensation with Compensation wlo export crops wI export crops wlo export crops wI export crops Welfare Level (the base value is 1.000) . Subsistence fanner 0.959 0.956 0.970 0.958 Landless worker 0.977 0.988 1.000 1.000 Rainfed fanner 0.953 0.951 0.932 0.938 Irrigated fanner 1.045 1.027 1.025 1.015 Annual Efficiency Gains (1989 US$ million) Resources left 1,241 961 166 376 SOURCE: S. Levy and S. van Wijnbergen "Labor Markets, Migration and Welfare: Agriculture in the North-American Free Trade Agreement," Journal of Developmelll Economics, 43(1994):263-278 (Tables 3 and S, pp. 269 and 273). 2.63 The first issue is the effect of the agreement on net exports from Mexico. A USDA study has. estimated the change in net exports between the US and Mexico by commodity group. The net effect i of the treaty is likely to be .negative on agricultural exports from Mexico to the US and, in any event, to be small in the context of Mexico's total exports. Josling reports one study of the net percentage changes in Mexico as: food grain production, -16.4 (biggest effect), -6.5 (smallest effect); feed grain production: -5.6, -3.2; fruits and vegetable exports: 24.7, 25.8. 2.64 Fruit and vegetable exports. Fruit and vegetables exports are an important group of commodities because Mexico's comparative advantage is believed to lie there. Even within the simplifying assumptions of comparative statistics, however, these are difficult goods to study because of variability in the commodities produced, production systems, seasonal effects, and consumer preferences in the two countries. Josling reported about a 25 percent expected gain in fruit and vegetables exports resulting from NAFfA. Cook summarized the evidence about this class of trade: \ concluded that Mexico has a clear cost advantage in fresh tomatoes, green peppers, squash, avocado~ and perhaps cucumbers. In other commodities, Mexico is at a minor cost advantage or a cost :, disadvantage. Despite Mexico'S apparent competitive advantage in these commodities, because of r lower wages, it does not necessarily achieve lower unit costs because of labor productivity, inefficie,t irrigation systems, and infrastructure costs. .J B. PROCAMPO ,'\ ,,\" /'1 r. 'fO /\,~ 1\'1'· \; ~/; " 2.65 On October 4, 1993, President Salinas announced a new support program for Mexican farming. The program, named PROCAMPO (programa Nacional de Modernizacion del Campo), has two components: agricultural trade liberalization and direct producer income support. The agency handling PROCAMPO is the Support Services for Agricultural Marketing company (ASERCA), ';.'[ which was created in 1991 as a decentralized agency of SARH with the mandate to take the necessary '/{ measures to adapt agricultural marketing to the new economic environment of liberalized prices. 2.66 The Government intends to free agricultural trade almost immediately and provide direct income supports to producers during a transition period in the agricultural trade liberalization  34 component which will be completed with a move to world prices for all farm commodities by October I, 1994. This would cause the domestic producer prices of some major crops - maize, beans, wheat, barley, and sorghum - to fall sharply, though to varying degrees. For maize, the most important crop, the guarantee price of N$750/ton that held throughout 1993 dropped to N$650/ton in the fall/winter season of 1993-94 and will go to N$600/ton in the spring/summer season of 1994; for beans, the next most important crop in area, the corresponding changes will be from N$2,100 to N$I,800 and then to N$I,600. 2.67 PROCAMPO includes a direct producer income support component to compensate farmers for some of the effects of the lower free trade prices. PROCAMPO will grant direct income support payments to farmers who have sown maize, soybeans, dry beans, sorghum, wheat, rice, cotton, barley, and safflower. All products benefit at present from government support programs. Farmers who have not historically planted areas to the above mentioned crops will not be eligible for payments under PROCAMPO. Payments will be granted on a per hectare basis and will be calculated as the average of the areas planted with eligible crops in the previous three years to December 1992. The income support will be roughly US$I13 (Le., N$350) per hectare, given to 3.3 million farmers in 1994, the transition year. Those farmers will receive a like amount in real terms for 10 years beginning on October 1, 1994. Beginning in the 11th year, the amount given will decrease linearly and become zero at the end of year 15. 2.68 Full implementation of PROCAMPO will take place in 1995. The transition period is needed to ensure the accuracy of the list of eligible producers and create the necessary bureaucracy to handle the payments. During the transition plan, a payment per hectare will be phased-in for maize, dry beans, wheat, soybeans, rice, cotton, and sorghum. The AutumnlWinter 1993-94 price will be N$330 per hectare; the Spring/Summer price will be N$350 per hectare. In AutumnlWinter 1994-95 full payments will be granted while the prices to be received by producers will be determined by market forces - presumably close to world levels. 2.69 Since the publication of most of the studies, the Mexican government has announced the PROCAMPO program. In terms of the general equilibrium studies, PROCAMPO implies full liberalization of Mexican policies and thereby resolves some of the variability in the projected results from those studies. PROCAMPO further introduces an element that was not anticipated in the NAFT A studies - compensation to producers losing from the move to free trade prices - thus posing the additional question of what is the effect of that compensation on the studies' implications. In addition, we now have more up-to-date information on the government's complementary investment policies, information that was essential in allowing Levy and van Wijnbergen to conclude that the transition to free trade prices could be welfare increasing for rural maize producers. 2.70 The Government contends that PROCAMPO has several advantages over trade protection. It: (a) is market oriented because it does not distort domestic prices; (b) reorients government agricultural support away from rich producers to poor ones; (c) is finite; (d) has a legal basis that is not discretionary; and (e) is consistent with NAFT A and the General Agreement on Trade and Tariffs ~I- (GATf). It does not encourage inefficient production with price incentives because it is totally , del inked from the quantity of current output. The most important consequence of PROCAMPO is that by providing direct payments to farmers, production decisions in terms of both output and input j-, mix are not distorted. It does not worsen the budget deficit by encouraging excess production of commodities that CONASUPO must then buy. PROCAMPO is more equitable than trade protection since it supports producers who received nothing from trade protection because they produced little or nothing for the market.  35 2.71 It is clear that PROCAMPO will be better in many respects than income support through trade protection, which was the old system. Most important is that the income support will be neutral; it will not send distorted signals regarding the production of specific crops. Second, subsistence farmers, because they consume most of their production, do not benefit from the guarantee price system which supports only commercialized output. Consequently, the poorest producers, who most need income support, do not receive it. PROCAMPO will give more income to subsistence farmers and, at the same time, the elimination of the guarantee prices will lower the prices of the maize and beans that they currently purchase. While larger farmers will still garner more of the total benefits of the trade policy, they will receive less in percentage terms than under the previous policy. Table 2.14 Savings from Alternative PROCAMPO Payment Policies Full Procampo 90% of Procampo 80% of Procampo 70% of Procampo Discount Payment - Payment - Payment - Payment - Rate 113 US$lha 102 US$/ha 90 US$lha 79 US$lha PANEL A: NPV of Procampo Payments (US$ million) 5% 11,982 10,784 9,585 8,387 10% 8,178 7,360 6,543 5,725 15% 6,450 5,805 5.160 4.515 20% 5,253 4.728 4,203 3,677 PANEL B: Ratio of Reduced Payment NPV to Full Payment NPV at 5% (percentages) 5% 100 90 80 70 10% 68 61 55 48 15% 54 48 43 38 20% 44 39 35 31 PANEL C: NPV of Savings with Alternative Payment Policies (US$ million) 5% 0 1,198 2,396 3,395 10% 3,804 4,621 5,439 6,257 15% 5,532 6.177 6,822 7.467 20% 6,729 7.254 7,779 8.305 SOURCE: Data taken from SARH. PROCAMPO: Vamos at grano para progresar. 1993. Calculations are in World Bank, LA2NR Mexican Agricultural Database. 2.72 It is expected that farmers with high yields (usually large commercial farmers in the Northern Mexico) will loose while farmers with low yields (usually medium size and subsistence farmers in the Southern Mexico) will gain. To see how losses and gains will be generated, consider the following example. The indifference yield for maize is 3.30 tons per hectare. It is calculated as 330/(750-650) = 330, where 330 is the fixed payment, 750 is the existing guarantee price of maize, while 650 is the  36 new price of maize, all in $N. That is, farmers with yields greater than 3.30 tons per hectare will loose while farmers with yields lower than 3.30 tons per hectare will gain. The corresponding indifference yields for beans and soybeans are 1.10 and 3.93 ton per hectare. These indifference yields correspond to FalllWinter 1993/94 period - which is a transition period. If one considers the world price of the commodities, rather than the transition period prices, then the indifference yields become much lower. For example, setting the world price of maize at N$450/ton, the indifference yield becomes 1.17 tons per hectare (calculated as 350/(750-450). However, this does not necessarily imply lower profitability as these calculations assume that farmers will keep producing maize. n C-Farmers have the choice of switching to more profitable crops. Therefore, the marginal loss in profitability due to lower maize prices (or any of the other eligible crops) may be more that compensated if one takes into account the N$350 per hectare payment and switching of crops, so that farmers, in fact, will better off. 2.73 Saving on PROCAMPO payments. PROCAMPO will add significantly to the Government of Mexico's (GOM) food and agriculture budget and it is therefore essential to consider ways to make it cheaper. Some simple simulations were performed of two possible economies in PROCAMPO payments. One is to simply reduce the average payment per hectare. The base payment is N$3501ha (US$1l3Iha at an exchange rate of N$3.1IUS$); the area on which payments are expected is 11.655 million hectares (of which maize is about 7.2 million hectares). Calculations were performed at 90 percent, 80 percent, and 70 percent of that base payment. A second way is to offer to buy producers payments at a steep time discount, thus lowering the NPV of the stream of PROCAMPO payments over the 15 year life of the program. 2.74 Table 2.14 shows the results of those calculations assuming that just 10 percent of the land is made available for buyout of the PROCAMPO payments. The savings at 90 percent, 80 percent, and 70 percent of the base paymentlha would obviously be proportional, but their magnitude is of interest. The savings, even assuming that the Government cut the base payment by only 10 percent per hectare, and that no farmers sold their PROCAMPO rights at a discount, would have a net present value of US$I,198 million. In annuity terms over a 15 year period, that sum would fund an annual program of research, extension, or marginal land buyouts of US$115 million. VII. Conclusions and Recommendations 2.75 Over time, producers are likely to completely alter the output mix, e.g., switching from production of maize and beans (which will no longer be profitable in many production regions) to fruits and vegetables. At the same time, it is possible that the "know-how" of producing alternative crops is not there. Therefore, the Government must transfer the existing technology and extension services where the demand for such services and technology is going to be. 2.76 The Government, through the guarantee price system, had practically isolated producers from exchange rate fluctuations (that is not to say that exchange rate policies did not affect agriculture). As \. V ' producers will be exposed to international markets from both importing inputs and exporting outputs, !"'fl~~, \ -exchange rate will affect them directly. To the extent possible, macroeconomic policies should not ~Iow for large misalignments of the real exchange rate. 2.77 Infrastructure regarding communications, transportation requirements, and storage facilities will be a vital element in the transformation of the sector since there is going to be substantial shifts in the distribution of production among regions.  37 3. COMMERCIALIZATION 3.1 This chapter discusses the previous commercialization system of major grains, oil seeds, and beans, analyzes current policies, and reviews the proposed commercialization system. The proposed commercialization system aims to remove the government from crop commercialization, eliminate subsidies and replace them with direct income transfers, open up the trade regime, further increase price transparency through the establishment of auction markets, and provide some instruments for risk sharing, such as the establishment of forward and, at some later stage, futures markets. Risk management under the proposed system is also discussed. Finally, we elaborate in some detail on market integration issues, since market integration is the most important precondition for well- functioning spot markets. 3.2 Until 1985, CONASUPO was the only importer of the main agricultural products (grains, oilseeds and beans). Starting in that year, private processors (traders were excluded) were allowed to import all products except beans and powdered milk. Imports were subject to import licensing requirements. Since 1985 the private sector has handled a large share of imports. Thus, even under the previous system, the private sector physically handled a large part of crop production and imports. The government set guaranteed prices for 12 main crops for each planting season. These prices were set to cover production costs for a variety of representative technologies and areas, with international prices serving only as a guide. Intermediate goods and consumer prices for basic foods and key inputs were also regulated. The system distinguished between urban and rural consumers - subsidizing the former and taxing the latter. CONASUPO bought crops at higher than border prices and sold them to urban consumers at lower than border prices, while rural consumers paid prices close to guaranteed prices. 3.3 Despite reforms after 1985, the Mexican Government still intervenes in agricultural marketing. The Government, through CONASUPO, buys maize and beans from farmers for subsequent processing, wholesaling and retailing. 12 Purchases of maize and beans are around 20 percent to 25 percent of output: given that on-farm consumption is around 15 percent to 30 percent, some 45 percent to 65 percent of production was marketed by the private sector. 13 3.4 Guaranteed prices were abolished in 1990 for most crops but, remain in effect for maize and beans with CONASUPO still being involved in marketing part of these crops.14 Maize, wheat, barley and beans are subject to import licenses. For sorghum, soybeans and rice import licenses have been abolished and replaced by a 15 percent seasonal tariff for sorghum and soybeans and an annual 10 percent tariff for rice. 12. CONASUPO's affiliates include two storage companies, BORUCONSA and ANDSA, accounting for over 50% of Mexico's basic food storage capacity; manufacturers of oils (ICONSA), corn flour (MICONSA). and dairy products (LICONSA); a wholesale food distribution network (IMPECSA), and nationwide chains of retail grocery stores (DICONSA). 13. For detailed information on production, commercialization and trade aspects of each commodity see, Mexico: Policy Notes on Agriculture, Food and Rural Development, Report No. 8144-ME, February I, 1990. 14. While the share of total maize production purchased by CONASUPO has somewhat increased compared to the 1980-88 average (from 17% to 22% in 1992) it has significantly reduced for beans (from 35% to about 8% in 1992).  38 3.5 Farmers and processors of maize face fixed prices within the season and therefore are not fully exposed to price volatility. The bearing of within-season price risk remains with the government because ASERCA's support payments for wheat, sorghum and soybeans do not fluctuate throughout the season as border prices change. For maize and beans, CONASUPO assumes all seasonal price risk. The annual price risk is borne by the farmers and processors - and to some extent by the government as agreement and guaranteed prices do take some time to reflect world price movements (see Box 3.1 below). The allocation of seasonal and annual price risk has not significantly changed between the previous and the current marketing systems. . .. . Box 3.1 How' IntegrateCJ are Mexican and US Grain Markets? .:" '." .... . Statistical analysis forl1l~e.<'Wheatisorghum, and soybeans reve~e4 that there exis~ 10ng-teIlll •.• co movement between MeXiclUlgUaranteeprices and US Gulf prices· over the period 1950;;1990. Such rerationsbip,bowevet/Wal)IlOr~videntbetween guarantee. and US Gulfpricespfbar1eyan# . rice. The averagetbnehtook. for tile Mexican prices to adjust to world price $hockswa:s;.~ii • years for maize; four years for sorghum; three years for soybeans; andfiveye3rsrorwbeat.N'otic¢ that for most of the period exa.rllinedhere, .guarantee prices were higherthanborderpd~es:.·.· . . I. The Proposed System of Commercialization 3.6 The Government has plans to liberalize the marketing of grains, oilseeds and beans beginning in the 1994/95 Fall-Winter season. It will no longer be involved in the purchasing and marketing of agricultural products. Farmers will sell their output on private markets at liberalized prices. It is expected, therefore, that prices received by the farmers for maize, beans and wheat will decline. IS Because farmers' incomes will drop significantly. ASERCA plans to compensate them by direct income transfers (see Chapter 2 on PROCAMPO). 3.7 To assist the transition to the new system, ASERCA plans to establish a national company (Fideicomiso Nacional para el Mercado de Granos y Oleaginosas (FINAGRO» located in Mexico City to operate spot and forward markets initially and possibly futures markets later. The plan also creates eight regional trust companies operating as branches of the national trust company. with delivery points inside each region. These regional trust companies will operate the physical and forward sales and purchases. 16 They will be connected via a computerized system to FINAGRO. The system will transmit information to all regional market locations with regard to prices, quantities, and characteristics of the commodities. It will also allow arbitrage and thereby reduce the risk of large local supply and demand imbalances. Each regional market will allow forward sales. At a later stage, there might be one futures market located in Mexico City and operated by FINAGRO. 15. The role of CONASUPO in the proposed system is not yet detennined. CONASUPO may still participate in the commercialization of crops as a private agent, having to cover its costs and make profits. It is likely, however, that the Ministry of Finance will no longer guarantee CONASUPO loans. 16. The number of products in each of the regional markets will vary depending on the region's production/consumption profile.  39 3.8 The trust companies will be private. Their members will be producers. producers' organizations, traders, agroindustry, storage companies, CONASUPO, credit unions, financial groups, banks, or other financial intermediaries. They will set minimum standards for the members regarding financial capacity and fulfilling certain characteristics in their activities. The trust companies will also handle a guarantee fund to deal with non-compliance in forward sales. The creation of spot markets, however, requires that certain preconditions are in place. The most important are listed below. 3.9 No government intervention. The Government cannot be involved in the marketing of crops directly or indirectly. It is understood that ASERCA's role will be limited in (a) supporting farmer incomes through direct transfers, in a manner that will not distort decisions regarding crop planting and marketing and (b) assist in the creation of spot markets. 3.10 Liquidity (i.e., many buyers and sellers). Liquidity will guarantee that prices are determined in a competitive environment. The regions where markets are to be established must attract a large number of buyers and sellers. At present we do not have good information on the size of each commodity market and the number of participants. We know that private agents have been actively involved in crop marketing. There are two concerns with regard to market activity for the period following market liberalization. First, the locations of the spot markets to be set up are based on historical production and consumption (buying) patterns. These are likely to change when prices are liberalized, which in turn may affect market liquidity. Second, there are large disparities among farmers with regard to their ability to market their crop. Larger farmers will not have much of a problem in adjusting, however, small farmers will. Given the large size of the physical contracts to be traded in the proposed spot markets, around 10 tons, small farmers can only access them through a cooperative marketing arrangement. 3.11 Inrrastructure. Market information, transportation, storage and financing systems have to be efficient and accessible in order to reduce marketing costs and make arbitrage between markets effective. However, infrastructure is a point of concern. Problems exist with the slow and expensive transportation of crops, the high cost of financing and bottlenecks in storage, which in turn will affect both the competitiveness of Mexican crops vis-~-vis imported crops and arbitrage among regions. With the planned market liberalization, possibilities for arbitrage are likely to increase. 3.12 Legal Basis. The existence of widely acceptable warehouse receipts, quality standards, property rights, enforceability of contracts, among others, are needed. An adequate legal basis is needed to support the creation of spot markets. Warehouse receipts (cenific'ados de deposito) exist in Mexico and are acceptable as collateral for loans, but CONASUPO, and the guarantee or agreed prices (precios de concenaci6n) limited their use in the past. Because close to 70 percent of all cargo moves by road in Mexico, competitive trucking services are vital. In July 1990, the Government removed route restrictions, eliminated mandatory use of cargo centers, released price controls and gradual liberalization of truck imports. No formal evaluation of the impact of deregulation in the trucking industry has been done. Transportation concerns, however, appear to be due mainly to infrastructure problems rather than to policies concerning the trucking industry. II. Risk Management Options 3.13 The proposed marketing system has a mechanism for risk sharing calJed deferred sales (venlas diferidas). Deferred sales are a form of forward sales and serve three essential functions: (a) share risk; (b) provide future price discovery; and (c) reduce price uncertainty.  40 3.14 Under the proposed system producers may choose to sell forward the harvested crop by putting it in a warehouse and obtaining warehouse receipts. The producers pay the costs of storage and finance until delivery. Forward prices are expected to differ from spot prices by (a) the sum of physical storage and inventory financing costs (costas de acarreo) and (b) price differences caused by shifts in demand and supply at the time of delivery. 3.15 On the buyers' side, processors who want to eliminate price risks can buy forward. Forward prices will be determined by the demand and supply of contracts for forward deliveries. The proposed mechanism will work as follows. When a processor finds a farmer willing to deliver on a specific date and the operation is closed, the processor deposits an initial margin in FINAGRO. The margin is 10 percent of the price when transaction is closed. The processor is required to make additions to this margin, called variation margins, if the actual price the next day is lower than the initial price by more than 2.5 percent. The idea behind the variation margins is to reduce risk of non-compliance. The system of initial and variation margins is the same as that in futures markets. Margins are required from the buyer (processor) but not from the seller (farmer). Farmers do not pay margins because they have effectively collateralized the transaction with their warehouse receipts. 3.16 The proposed system of deferred sales serves the important function of smoothing out price fluctuations within the year. Farmers, processors and traders can arbitrage price differences if prices between two time periods exceed storage and financial costs. 3.17 The proposed system of deferred sales provides less risk sharing than a true system of forward markets. The system gives more flexibility to buyers (processors) than sellers (farmers). Farmers need to have their crop in storage before they can sell forward, while buyers can buy forward any time they want to provided they post the necessary margin. An important function of a true forward market is to allow farmers to sell forward before planting to assure that they can reduce their price risks and hence guarantee some of their profits. This guarantee increases the certainty of the farmer's investment and also makes it possible to obtain commercial financing on better terms as banks perceive the farmer's investment to be less risky. 3.18 In a true system of forward markets, farmers can sell forward in the same way as the processor can buy forward. In the United States, forward contracts between grain farmers and elevators are legally binding and usualty no collateral is involved. Because of the novelty of the system in Mexico and because a few deals going sour could jeopardize the whole operation, collateral is advisable in the beginning. The collateral can be in the form of initial margins and variation margins thereafter, in the same way as for buyers, or farmers could put as collateral the income transfer from ASERCA (Le., PROCAMPO payment). The forward contract will be between the farmer(s) and the storage company, or even the processors. The counterpart in the contract hedges the assumed risk in the US futures/options markets. 3.19 Because there is usually some minimum quantity requirement regarding the size of pre-planting forward contracts, small farmers will need to organize into groups in the same way as when marketing the physical crop. Credit unions or cooperatives can serve this function. The pre-planting forward contract will allocate the price risk from the farmers to the traders or storage companies which, by handling larger volumes, can access the US futures/options markets. 3.20 An important issue in the use of pre-planting forward contracts is that of production risk. To illustrate this, consider the following: a farmer decides to sell forward the crop before planting. If yields drop significantly, the farmer'S position can become speculative because he has assumed effectively a short position. One option to fill the short position is to buy from the market an amount  41 equal to the shortfall and use it to meet the contract. This problem can magnify if the crop failure the farmer has experienced is widespread, causing prices to increase at the same time as more short producers enter the market. 3.21 In practical terms, no one hedges forward all of the crop. Hedging takes into consideration the probability of crop failures and adjusts the amount to be hedged to reflect it. Another way to handle the problem is to suggest a limit on how much can be sold forward before planting, based on historical yields. 3.22 Forward contracts can be incorporated in a package of pre-planting financing. For example, traders or warehouses may advance money to a farmer and request to be repaid at a date after harvest with a specific amount of the crop. This transaction incorporates an implicit forward price. These types of pre-planting forward contracts may take place informally under the proposed system. To what extent there is a need to formally standardize them and incorporate them into the marketing system should be determined by the private sector itself (farmers, traders, processors and/or warehouses). 17 In. Mexico and CommOdity Futures Markets 3.23 Another way to facilitate risk management and price discovery is to create a Commodity Futures Market (CFM). However, forward and spot markets often precede futures market and are not fungible as each is individualized to the particular market circumstances. Spot and forward markets also imply physical delivery, while in commodity futures markets physical delivery occurs rarely. Futures contracts are highly fungible financial instruments traded on organized exchanges so that contracts of the same maturity are perfect substitutes. 3.24 The demand for price insurance alone does not explain the existence of organized futures markets since forward cash contracts also provide price insurance. The additional and necessary element for a futures market is the demand for a fungible financial instrument traded in a liquid market. Hedgers and speculators are attracted to futures markets because they assist them to manage their portfolio risks. Portfolio management is based on specific relationships between individual futures prices and the returns of other financial instruments. Demand for such a portfolio management service must be identified when evaluating whether or not to establish a futures market. Two important preconditions for a CFM are the broad economic environment and specific futures contracts. 3.25 Broad economic environment. Most important are: (a) well functioning communications, storage, transportation, information processing systems; (b) developed commercial and financial sectors; (c) appropriate legal and regulatory framework; (d) no government intervention and existence of spot markets with publicly available information; (e) sound economic policies towards monetary management, foreign trade and creation of a viable currency; (t) enough viable trading firms and financial institutions who do sufficient market activity to join together in an exchange environment. A country cannot depend on producers and consumers alone for the viability of a CFM; and (g) a viable clearing house to establish the financial integrity of the market. 17. Note that informal fOlWard contracts can be legally binding. In the US, the telephone conversation between the farmer and the grain elevator is sufficient to cover legally the transaction. If it is necessary, however, the grain elevator may send a contract to the farmer to be signed.  42 3.26 Specific rutures contracts. The chief conditions for viable commodity futures contracts are: (a) high correlation between cash and futures prices (i.e., low basis risk); and (b) a trading volume large enough so that transaction costs will be small; and (c) widespread entry and exit from the market are possible without disrupting prices. 3.27 If those conditions exist, then creating a domestic futures exchange or using existing foreign exchanges is a matter of costs and benefits. The costs are those of establishing and operating the CFM; creating and fostering liquidity; and operating an appropriate regulatory environment. These costs must be compared with the costs of using external markets. such as the additional commodity price and currency risk and maintenance of capital accounts in a foreign currency. 3.28 The main benefit of a domestic CFM compared to a foreign is the lower basis risk. 18 Lower basis risk can occur if a domestic CFM futures contract more closely represents the local cash commodity (physical contract size, quality). Delivery will be at a nearby location. so the basis should be predictable and close to zero at contract maturity. 3.29 The use of existing foreign CFMs involves commodity price and exchange rate risk. Basis predictability is more costly in foreign markets because a commodity futures contract is usually defined for a cash commodity abroad with different characteristics. The underlying demand and supply conditions may differ between the two countries. Even where market fundamentals for the commodity in the two countries are the same, currency misalignment may create additional risk. 3.30 Foreign markets usually have the advantages of established rules and regulation liquidity. Higher liquidity may mean that basis risks will be lower with foreign markets than with internal markets, despite the dual price and currency risk. For example, corn and wheat are traded internationally, and contracts in Chicago are price reference points. These contracts reflect demand and supply conditions in the US, but they are also sensitive to the world grain situation because of the large role of US exports. When a country's currency is stable, hedging in existing markets may be a reasonable matter because of their higher liquidity and lower transaction costs thinly-traded domestic market can potentially have higher execution costs and more basis risk than the combined price and currency risks from existing foreign markets. 19 International futures contracts for such products are not likely to be successful, so countries can give more consideration to developing their own contracts in these situations. 3.31 With the liberalization of marketing for grains and oilseeds and a more open trade regime, Mexican prices for these commodities will move closely with those in the US. thus reducing price disparities between world and domestic prices.:>D Following liberalization, Mexico is expected to be in a significant production deficit and will depend on US imports to fiU the gap. Conversations with grain and oilseed traders indicated that they expect to see a close correlation between Mexican and US prices. They cited the example of Mexican cattle, coffee and cotton prices which, to a large extent, reflect movement in US prices. 18. Additional benefits are the improved price discovery, improved information transmission, improved credit systems, more accurate fOlWard prices, and possibly reduced transactions, entry and exist costs. 19. It is the case, however, that livestock are not highly traded internationally (especially across oceans), so their prices in various countries are more independent of each other. 20. For maize and barley, the correlation between Mexican and US prices will depend on whether the import quotas are binding under the NAFrA regime.  43 3.32 Arguments that the establishment of a CFM in Mexico could make such markets more accessible to Mexican farmers, particularly the smaller ones, are not well-founded. Futures markets are not designed for use by small farmers. Smaller farmers do not usually possess the know-how to use CFMs directly, and the transaction and monitoring costs are usually too high for them to access these markets. Smaller farmers usually find larger traders or warehouses that can pool the risk from a large number of small farmers and then hedge it in the exchanges. Thus, whether the CFM is located in Chicago or Mexico City it will be quite difficult for the smaller or medium-sized farmer to use it. 3.33 Exchange rate risk might prevent use of the US exchanges. However, there currently exists a forward market for interest rates in Mexico and by using it one can create a synthetic exchange rate hedge. The problem is that the current interest rate forward market can provide a coverage for up to only three months. Thus, there is a need to develop longer interest rate forward coverage and also exchange rate contracts to simplify hedging (Le., so that hedgers do not need to construct synthetic hedges and can use exchange rate forward contracts directly). The development of exchange and interest rate contracts is also important in the case of a Mexican CFM, since they can provide additional income to the exchange and also be very useful in cross-hedging operations between U.S. and Mexican CFM. 3.34 Copy-cat contracts (Le., contracts that duplicate existing ones) have failed. For example, Brazilian coffee and Malaysian cocoa producers use New York futures contracts for hedging purposes, despite the existence of domestic contracts. In September 1993, at UNCTAD's meeting of the Group of Experts on Commodity Exchanges, there was a warning against the creation of copy-cat contracts. Instead, it was noted that different formats of an exchange should be considered such as franchiSing of existing international trade contracts (Le., trading such contracts in local hours and clearing them on the exchange of origin). This causes illiquidity in the local futures markets and further inhibits their use. The experience so far of the establishment of CFMs in developing countries has not been encouraging. From recent observation of the 24 futures/options contracts listed in five major CFMs in developing countries, only eight had some liquidity. Also, even in developed countries, about two-thirds of the new futures/options contracts launched fail as traders prefer more liquid markets than markets with lower basis risk. For the case of Mexican grains and oilseeds, US futures markets provide very good liquidity and, for most cases, an acceptable basis risk as many countries use the US markets as a benchmark for their domestic grain and oilseed prices. 21 3.35 Transportation and storage costs are often cited as reasons to establish a domestic CFM in Mexico. The idea is that a domestic CFM will reduce the basis risk due to transportation and storage. However, if transportation, storage and inventory finance costs do not fluctuate considerably, the basis risk is not significantly affected. Further, even if they do fluctuate considerably, problems in transportation and storage are mainly internal. 22 That is, transportation and storage problems are likely to affect the internal arbitrage possibilities. In this case, regional price movements will deviate, including those in Mexico City, which is the point of futures contract 21. There is, however, the successful case of com and wheat futures contracts in the Budapest Commodity Exchange. These contracts were introduced in 1990 and their liquidity has been growing rapidly. Possible reasons for their success has been the size of the internal market, the fact that internal consumption and production are fairly well balanced, the long distance from the US markets, and exchange rate risks. 22. In many cases, it is easier and cheaper to transport grains and oilseeds to Mexico City from the US than from several regions in Mexico.  44 delivery, Zl implying significant internal and external basis risks. If regional prices do not move closely together, then representative prices for grains and oilseeds cannot be established, reducing significantly the viability of Mexican futures contracts for these commodities - the next section deals exclusively with the existence of a representative domestic price. 3.36 The creation of a CFM in Mexico should be the work of the private sector. It should not be set up by the government; certainly not before it is seen how well the spot and forward markets function. Z4 A significant factor in the decision to use a Mexican CFM should be the observed relative basis risk, i.e., the co-movement of price changes between regions in Mexico and between US and Mexican prices. However, internal and external basis risks for Mexico will not be known until after the spot and forward markets are established. The bottom line is that active trading, both hedging and speCUlative, is required for a successful CFM. Private firms and financial institutions must be committed to using the CFM. 3.37 At present, there are futures contracts for most of the commodities that Mexico plans to trade in regional spot markets. For com (yellow maize), soybeans, barley and wheat there are futures contracts in US exchanges with significant liquidity for up to 12 months. There are also futures contracts in New York, for coffee, sugar and cotton that could be appropriate for Mexico to use. There are no futures contracts for white maize, sorghum, and beans. However, sorghum and maize in the US are hedged using the com futures contract. For example, maize can be traded based on com plus a premium reflecting quality differences. If, for the case of Mexican maize, the premium is very volatile (mainly intra-year volatility) then the effectiveness of Chicago's com futures contract as a hedging instrument for Mexican maize may diminish. Based on US experience, this is not very likely. Similarly, sorghum can be traded based on com adding the appropriate premiums. For beans there is neither a futures contract nor a reference prices that can be established for trades. The lack of futures contracts for beans can be due to its lack of standardization and relative low volumes of trades, at least in the US. Finally, there exists a futures contract for rice in Minneapolis but it is not very liquid. 25 3.38 An important precondition, not only for forward and futures markets, but also for spot markets is that there exist possibilities for arbitrage. Stating it otherwise, a representative price must exist (subject to transportation costs). To see whether possibilities for arbitrage exist, we examine the degree of market integration in the existing markets (cemros de abastos). If these markets are well integrated, it implies that the proposed spot markets will have a high probability of success. If these markets are not integrated, the proposed spot markets will subject to many difficulties, unless the reasons of weak integration are identified and subsequently treated accordingly. IV. Integration or Existing Markets 3.39 Effective arbitrage is a direct consequence of market integration. In integrated markets localized scarcities are likely to be eliminated within short periods of time while in the absence of integration not only these scarcities will tend to persist, but governmental attempts to eliminate them will be costly (Ravallion, 1986). Furthermore, as stated in the previous section, effective arbitrage 23. The same argument can be applied if another location is chosen as the point of futures contract delivery. 24. The Government, however, may need to aid in the creation of a CFM if the private sector expresses the need and takes the initiative to establish it. 25. The heavy government intervention in Asia has inhibited the extensive use of rice futures.  45 (and hence market integration) is a condition for the viability of commodity exchanges (e.g., spot market, forward market, or a sophisticated futures exchange). 3.40 Tests of market integration use models that essentially make a "Law of One Price" (LOP) argument. Suppose that a market for a specific commodity consists of two locations, 1 and 2. If trade takes place, any imbalances in demand and supply will lead to regional price disparities and hence profit opportunities. Eventually, price fluctuations originating from location 1 will be transmitted to the price in location 2 and vice-versa, hence leading to realignment of prices and elimination of arbitrage opportunities. Naturally, equalization of prices is expected to take time as the speed at which supply and demand imbalances will be eliminated may be low, depending on infrastructure, distance, market structure, and non-price market regulations. 3.41 In this section, we examine the degree to which the markets of rice and two types of beans are integrated. Although these are not the only commodities subject to modernization, since they share the same marketing characteristics of both grains and oilseeds they can be used as indicators of the status of the Mexican markets in terms of integration. 3.42 The model is applied to the markets for milled rice (arrozpulido), traditional beans ifrijolflor de mayo), and white beans (haba). In the milled rice and traditional beans markets four destinations (consumption regions) were considered: Guadalajara, the Federal District (Mexico City), Ecatepec, and Torreon. Sinaloa is the source of milled rice and Zacatecas is the source of traditional beans. For the market of white beans the following four consumption regions were selected: Guadalajara, Federal District, Ecatepec, and Monterrey. White beans originate from the states of Puebla and Mexico. 3.43 Ecatepec and Federal District serve the northern and southern part of Mexico City, respectively and are 45 kilometers apart. Guadalajara is approximately 450 km west of Mexico City. Torreon and Monterrey are about 1,500 km to the north of Mexico City and approximately 300 km apart. The prices in the consumption regions are determined in local wholesale markets supposedly reflecting supply and demand conditions. Notice that at this level of trade government plays no role so the reported prices are the ones that retailers pay to wholesalers under competitive conditions. 3.44 The price series to be analyzed were obtained from Anuario Estad(stico de Granos y Semillas, published by SNIM (Servicio Nacional de Informaci6n de Mercados), a price monitoring agency of the Ministry of Trade and Industry (SECOFJ). The series are weekly for 1991-92 and include 104 observations. Prices refer to products of high quality, quoted in the morning of each Wednesday at the wholesale markets (centros de abastos). The price series was deflated using the consumer price index (CPI) published by the Banco de Mexico. Figures 3.1.a, 3.2.a, and 3.3.a depict the price series regarding milled rice at Federal District and Torreon, traditional beans at Guadalajara and Ecatepec, and white beans at Guadalajara and Ecatepec, respectively. A. Market Integration in the Long Run 3.45 This section discusses results regarding long-run market integration. The results regarding mined rice (panel A, Table 3.1) are as follows: If one excludes the ADF test for Federal District- Ecatepec, all three tests indicated that the regions of Federal District, Ecatepec, and Guadalajara are  46 integrated in the long run. However, Torreon appears to be totally segmented from the central locations, as its price difference with these locations in no case was stationary.1fj 3.46 Long run market integration results for the market of traditional beans are presented in panel B of Table 3.1. Guadalajara-Ecatepec as well as Federal District-Ecatepec exhibited strong evidence of long-run integration. However, in the Federal District-Guadalajara case. only one test supported long-run integration. In contrast to the milled rice market, the Torreon-Ecatepec differential for the traditional beans market presented some evidence of long-run integration. 3.47 Finally, the market of white beans presented a rather clear picture and similar to that of milled rice (panel C, Table 3.1). Federal District, Guadalajara, and Ecatepec are integrated in the long run as all respective price differentials exhibited stationarity at the 5 percent level of significance. However. Monterrey is not integrated with the locations of the center as its price was stationary. 3.48 Figures 3.1.b, 3.2.b, 3.3.b depict price differentials reflecting three distinct patterns in terms of the existence and the degree of market integration; the dotted line corresponds to the respective means. Specifically, Figure 3.1.b depicts the milled rice price differential for Federal District- Torreon - not exhibiting long-run integration. Figure 3.2.b depicts the traditional beans price differential for Guadalajara-Ecatepec - exhibiting long-run integration but low speed of adjustment (the next section deals exclusively with the adjustment issue). Finally, figure 3.3.b depicts the white beans price differential for Guadalajara-Ecatepec - exhibiting long-run market integration as well as high adjustment speed. 26. A series is said to be stationary if its mean and variance exist. Otherwise it is non-stationary. One can induce stationarity in a series by taking first differences.  47 Figure 3.1.a: Milled Rice - Prices 3.00 r---------------------------------. 2.50 - - - - F. District \ }t - - - - - - Torreon c:.o 2.00 S)( ~ 1.50 1.00 0.50 +--+-+--+--1---1--1---1---11---1----.1---1---1--1---1--1---1--+---+--+--+---: II 21 31 41 51 61 71 81 91 101 weeks Figure 3.1.b: Milled Rice - Price Differential 0.60 . , . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - . 0.30 0.00 -0.30 -0.60 -0.90 -1.20 +---+--+---!---+---+--+---t--I----.!---I--I---I--+---+--+--+--+--+-+--+----' 11 21 31 41 51 61 71 81 91 101 weeks  48 Figure 3.2.a: Traditional Beans - Prices 4.00 ,--------------~-----~--_n----~--___. 3.80 - - - - Guada.lajara··_········..···_-- Ecatepec 3.60 3040 co ~ 3.20 ..... :< ~ 3.00 2.80 2.60 2040 2.20 +--+--+---+--+---+--+---+--f--f----i--+--+--+---I--!--f--;--i---;--i----l 11 21 31 41 '1 61 71 81 91 101 weeks Figure 3.2.h: Traditional Beans - Price Differential 0.60 0.40 -0.20 -0.40 -0.60 -0.80 .1.00 -I---+--I--+--+-+--+--I----II---+-+---+--+--+---+--+--+---+-+--+--+--' 11 21 31 41 61 71 '1 81 91 101 weeks  49 Figure 3.3. a: White Beans - Prices 5.10 ; - - - - - - - - - - - - - - - - - - - - - - - - - - -_ _ _---. 4.60 ---Ecaupec - - - - - Guadalajara - ~ 4.10 ~ ~ C.l :::;: 3.60 3.10 2.60 +--+-+--+-+--+----1----f---+-+---+--f----i--+--I---+--I---+--+-+--1--.J 11 21 31 41 51 61 71 81 91 101 weeks Figure 3.3.b: White Beans - Price Differential 0.80 . , . . . - - - . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - , 0_60 0.4C 1 I ::: 1--- -------------vi4----- - ------IlJ1;0~~vt--- ~.20 ~_80 +-.!......t---i----i--+---+--t--t--+---+--+--I-----'---+---+--+--I---+----;--+--+--' I 11 21 31 41 51 61 71 81 91 101 weeks  50 Table 3.1. Long-Run Market Integration Results I I DF PANEL A: Milled Rice I ADF I PP I DISTRICl - GUADALAJARA -3.82<1b -4.34<1b -3.8QC F. DISTRICl - TORREON -1.58 -1.98 -1.92 F. DISTRICl - ECATEPEC -3.49<1b -2.76 -3.37<1b GUADALAJARA - TORREON -1.62 -2.06 -1.94 GUADALAJARA - ECATEPEC -4.()()<Ib -4.62<1b -4.07<1b TORREON - ECATEPEC -1.39 -1.88 -1.71 PANEL B: Traditional Beans F. DISTRICl - GUADALAJARA -2.40 -3.09<Ib -2.71 F. DISTRICl - TORREON -2.34 -2.67 -2.29 F. DISTRICl - ECATEPEC -3.89<1b -3.67<1b -3.79<1b GUADALAJARA - TORREON -2.44 -2.66 -2.60 GUADALAJARA - ECATEPEC -3.42<1b -4.1S<Ib -3.69<1b TORREON - ECATEPEC -3.0?<lb -2.93<1b -3.07<1b PANEL C: White Beans iF. DISTRICl - GUADALAJARA -4.3S<Ib -4. 6 S<Ib -4.41 <Ib F. DISTRICl - ECATEPEC -3.84<1b -3.98<1b -3.79<1b GUADALAJARA - ECATEPEC -5.11<1b -3.97<1b -S.28<1b NOTES: @ indicates stationary price differential at the 5 percent level of significance, in tum implying that the markets as defined by the respective locations are integrated in the long run. As the price of white beans in Monterrey was stationary (all the other prices were non-stationary), it is concluded that Monterrey is not integrated with the other locations. B. Quantifying the Long Run 3.49 The next question in this is: How long is the long run? Results - based on an extension of the model which tested for long-run market integration - are reported in Table 3.2.  51 Table 3.2 Number of Weeks Required for 90 percent of the Adjustment Milled Rice Traditional Beans White Beans 1. RDmDUcr-GUADAWU~ 22 31 7 2. GUADAWUARA - F. DISDUcr 22 - 13 3. ECATEPEC - GUADAWUARA 15 - 5 4. GUADAWUARA - ECATEPEC 12 16 - S. F. DISDUcr - ECATEPEC 13 - - 6. ECATEPEC - F. DISDUcr - 12 9 7. TORREON - ECATEPEC ns 20 na 8. ECATEPEC - TORREON os - na NOTES: The figures give the number of periods (weeks) required for 90 percent of the adjustment to take place. For example, if the price of rice in Ecatepec increases by 20 percent, it will take 12 weeks for the price of rice in Guadalajara to increase by 18 percent (row 4). (-) indicates that the error correction term of the dynamic specification was not significant at the 5 percent level; (ns) indicates that the corresponding price differential was not stationary; (na) indicates that data for the specific market were not available. 3.50 As the first column indicates, despite the fact that the first stage indicated strong long-run market integration for Federal District-Guadalajara, the second stage of the model gave a rather long adjustment period in both directions (22 weeks) - this model will also be referred to as error- correction model. The Ecatepec-Guadalajara model indicated a shorter adjustment period, giving 15 and 12 weeks. The Federal District-Ecatepec model, while showing a relatively quick adjustment when Federal District was the dependent variable (13 weeks), in the reverse direction the error correction term was not significant. Ecatepec-Torreon was not integrated in the long run, consequently we did not proceed with the second stage model. 3.51 In the traditional bean market (second column, Table 3.2), the Federal District-Guadalajara differential was marginally stationary, a fact reflected in the dynamic specification where in one direction the error-correction coefficient was not significant while in the reverse direction the model implied an extremely slow adjustment (31 weeks). On the other hand, the results indicated quicker adjustment for Ecatepec-Guadalajara (16 weeks); however, the error correction term was not significant in the reverse direction. The error correction term of Federal District-Ecatepec direction was not significant either while the reverse direction gave an adjustment period of 12 weeks. Finally, the Torreon-Ecatepec model, while showing an adjustment period of 20 weeks in one direction, it did not yield significant error-correction term in the reverse direction. 3.52 The third commodity examined was white beans (third column, Table 3.2). The differential between Federal District and Guadalajara exhibited strong stationarity. The dynamic model gave adjustment periods of 13 and 7 weeks. The Ecatepec-Guadalajara model, while giving insignificant error correction term when Ecatepec was the independent variable, the reverse direction exhibited the quicker adjustment observed among all three markets examined (5 weeks). FinaIty, the Federal District-Ecatepec model exhibited high adjustment from Federal District to Ecatepec (9 weeks) but  52 insignificant error correction term in the reverse direction. To summarize, contrary to the long-run market integration results, the adjustment periods crucially depend on the commodity: while prices in the market of white beans present relatively quick adjustment, the markets of milled rice and traditional beans exhibited slow adjustment. 3.53 For a more practical point of view one may be interested in the percentage of adjustment which takes place within a shorter period of time. Table 3.3 reports results for a two week-period. To summarize: while the market of white beans indicated that relatively high percentages of adjustment are attained within the two-week period, the other two markets did not indicate so. Table 3.3 Percent or Adjustment Taking Place within two Weeks I I Milled Rice Traditional Beans White Beans I. F. DISTRICT - GUADALAJARA 12% 13% 69% 2. GUADALAJARA - F. DISTRICT 8% - 49% 3. ECA TEPEC - GUADALAJARA 36% - 71% 4. GUADALAJARA - ECATEPEC 49% 32% - S. F. DISTRICT - ECATEPEC 39% - - 6. ECATEPEC - F. DISTRICT - 39% 57% 7. TORREON - ECATEPEC ns 23% na 8. ECATEPEC - TORREON ns - na NOTES: The figures give the percentage of adjustment which will take place within a period of two weeks. For example, for any rise in the price of milled rice in Ecatepec, 49 percent of the price change in Guadalajara will occur within two weeks (row 4). For other notes see Table 3.2. 3.54 To give a more concrete picture of the adjustment mechanism, Figures 3.4.a and 3.4.b depict the adjustment paths for Ecatepec-Guadalajara and Federal District-Ecatepec for all three commodities for up to 25 periods. The horizontal axis represents periods while the vertical axis gives the respective adjustment (expressed in percentage units),  53 Figure 3. 4. a: Adjustment Paths for Ecatepec - Guadalajara 100% •• ",. ___ M"' ___ ..- .' ---' - -_---io I - ~ -------.....---- ....... ....... E 80% / / -.,...--.,...-- --- ...... --- -- /- <1) E 60°A, ; '!Zi ~c:I '- . I ,," ;- _/ ",.""."... i 40% !;: - - 0 .i .! ,,' '#. .: // i ;//'/..{ 20% I V/~ W.Bam . - - - - - - M. Rice -----.---~ T. Bam I '" 3 7 9 11 13 IS 17 19 21 23 2S weeks Figure 3.4.b: Adjustment Paths for Federal District - Ecatepec 100% ,_. -- .--.~.-- .....-.-.. ---........--.--..--.-.--.-....."'"'.."''r 80% ......... ----_ ....... -...."....-, -= ." .'. ~ 60% c '!Zi ~<: - '0 40% '#. 20% .....- .. - ......... M. Rite - - - - W. Beans -----.---- T. Beans 0% 3 7 9 11 13 IS 17 19 21 23 2S weeks  54 Table 3.4 Average Percentages of Adjustment Milled Rice Traditional Beans White Beans F. DISTRICT-GUADALAJARA 12% 13% 69% GUADALAJARA-ECATEPEC 49% 32% 71% 51% ECATEPEC-F. DISTRICT 39% 39% 57% 45% AVERAGE 33% 28% 66% NOTES: The percentages of the first three rows and columns are taken from Table 4 and they correspond to the direction exhibiting the highest rate. The last row and last column report averages over rows and columns, respectively. The averages have been rounded to the nearest integer. 3.55 Although it is beyond the scope of this section to examine the reasons behind the differences in the adjustment rates, the results merit some further discussion. If distance (or infrastructure) was a problem, one would expect higher adjustment rates between Federal District and Ecatepec than Ecatepec-Guadalajara and Federal District-Guadalajara. Contrary to expectations, however, taking the averages over destinations (Table 3.4) gives the following picture: Federal District-Ecatepec averaged 45 percent, Guadalajara-Ecatepec averaged 50 percent while Federal District-Ecatepec averaged 31 percent. 3.56 On the other hand, one would expect that, since the commodities examined share the same characteristics in terms of transportation and storage requirements, they would exhibit similar adjustment rates among markets 27 • Again, contrary to expectations, while the markets of milled rice and traditional beans averaged 33 percent and 28 percent respectively, the market of white beans gave an average of 66 percent. Therefore reasons beyond distance and infrastructure may be responsible for the differences found among adjustment rates. 3.57 While long-run integration was found in proximate markets (e.g., Federal District, Ecatepec, and Guadalajara) it was not seen in distant markets (e.g., Torreon and Monterrey), indicating that long-run market integration is spatial and not commodity specific. For the markets that were integrated in the long run, the dynamic specification indicated that price changes are transmitted with low speed for miJIed rice and traditional beans and with a much higher speed for white beans, indicating the speed of adjustment depends on the commodity. From a policy viewpoint, the creation of forward or futures markets in Mexico seems premature, given that one of the main preconditions, i.e., well functioning spot markets, is not always satisfied. 27. Given that 70% of all cargo in Mexico moves by road, the trucking industry may be a point of concern. Until 1989, the trucking industry was heavily regulated through price controls, route restrictions, and other restrictions. Consequently, in addition to uncompetitive prices, the services were extremely unreliable. In luly 1989, however, the government removed almost aU restrictions so the industry is supposed to operate on a competitive environment. Therefore, policy induced distortions in the transportation industry should not be a point of concern as the sample used in the analysis covers the 1991-92 period, weU after the deregulation took place.  55 V. Conclusions and Recommendations 3.58 To conclude, the proposed system of commercialization represents an important improvement over the previous and current systems. The main strengths of the proposed system are: (a) non- distortionary production and consumption incentives structure; (b) significant reduction of the government's role in the marketing of agricultural products; (c) opening up of the trade regime; and (d) more transparent determination of agricultural prices. 3.59 The proposed system has provisions for risk management through the establishment of formal forward markets. These markets along with the evolution of informal forward markets (after the proposed system starts functioning) and the use of US futures markets may provide the necessary risk sharing to market participants. 3.60 With regard to the futures market in Mexico, the spot markets should be given some time to function; then the private sector will determine the appropriateness of domestic futures market. It is not clear that the Mexican government has a comparative advantage in initiating and/or managing futures markets. The government can facilitate and improve conditions that could link Mexican agriculture prices with US agriculture futures prices and that could support the development of spot markets by removing trade barriers and infrastructure bottlenecks associated with the crop commercialization (e.g., storage, communications, transportation, and financing). 3.61 The market integration tests performed above indicated that even in the existing markets (Le., centros de abastos), integration is not very strong. While such result is expected for location separated by large distances, it is not expected for locations as close as Federal District and Ecatepec. 3.62 Some recommendations are: (a) A Commodity Futures Exchange should not be created. If there is a need for such an exchange, the private sector will demand it and subsequently will undertake its creation and support its operation. (b) To the extent possible, it should be ensured that the necessary infrastructure regarding the creation of spot markets exists, as such markets are a vital element of the proposed commercialization scheme. (c) Given that the Government has been involved in the marketing of major crops for most of the century, knowledge of how are prices determined in a competitive environment is sometimes lacking. While this lack of knowledge may seem trivial, it is likely to affect the way that small and medium farmers will handle the marketing of their crop. Hence, opportunities from experienced traders to exploit small farmers may arise. A way of educating the farmers regarding the difficulties of the new marketing system must be identified, in case that such exploitation becomes a serious problem. (d) After the spot markets have been established, the Government should monitor them (but not intervene) in order to avert the development of local monopoJies.  56 4. IRRIGATION AND WATER MARKETS 4.1 Irrigation has propelled Mexican agriculture since World War II but has confronted serious problems in the last twenty years. The problems have included a slowing rate of area expansion, due to rising water scarcity; low return to many public irrigation investments; inattention to the serious environmental costs of irrigated agriculture, including aquifer depletion and salt encroachment; and inadequate maintenance of works, causing, inter alia, low conveyance efficiencies throughout many surface irrigation networks. This Chapter reviews the main issues affecting the efficiency of irrigated agriculture. Because the Bank has long been involved in Mexican irrigation, there is a rich body of previous sector and lending work to use as sources. 2I I. Irrigated Agriculture 4.2 Roughly 5.2 million hectares now have irrigation capacity.2\! Of that area, approximately 3.2 million is available into IDs and another 2.0 million hectares in IUs. The IDs were built on public authority and were, until recently owned and managed by the main government irrigation agency, the National Water Commission (CNA). The IUs were built by government, but are owned and operated privately, typically have smaller command area, and usually use tubewells, not surface water sources. 4.3 The states of Sonora, Sinaloa, Baja California Norte, Tamaulipas, and Guanajuato have the most valuable irrigated farming (as shown in Table 1.2) while the states of Tabasco, Campeche, Guerrero have the least. In the more arid irrigated areas, such as most of the northwest, crop production would be impossible without irrigation. Irrigated crops include such high-value commodities as tomato, green peppers, squash, cucumber, broccoli and strawberries all of which are prominent exports. In addition, they provide more employment per unit of output than do most of the rain fed field crops. Crop yields with irrigation are much higher than yields of the same crops without irrigation. Prominent examples are maize (average 1991-92 yield, 3.91 mtlha with irrigation, and 1.84 rainfed), sorghum (5.53, 2.78), wheat (4.66, 1.83), and beans (1.35, 0.48). 4.4 While the value of agricultural production grew slowly during the 1980s, irrigated agriculture's share of the value of production grew from 45 percent to 55 percent, while total cropped area fell by 18 percent and the irrigated area remained about constant. In 1991, the per hectare value of production and yield in irrigated areas were about 175 percent and 150 percent, respectively, higher than the value of production and yields of rainfed areas. Irrigated production accounted for 28. World Bank and FAO/CP, Review of Conditions for Encouraging Private Investment in Irrigation, 1993 (Report No. 11774-ME); World Bank, Irrigation and Drainage Sector Project, Staff Appraisal Report, 1992 (Report No. 9779-ME); World Bank, On-Farm and Minor Irrigation Networks Improvement Project, Staff Appraisal Report, 1994 (Report No. 12280-ME). 29. The definition of irrigation capacity is the potential to supply additional water upon demand to crops; this excludes crops grown with residual moisture, flood-recession crops, or crops receiving single irrigations (riego de punto) at planting time, as is the practice in parts of the country. Some 5.44 million hectares were planted with irrigation on average from 1989 through 1992. The Staff Appraisal Report (SAR) for the Irrigation and Drainage Sector Project (Loan 3419-ME; Report 9779-ME) refers to 6 million hectares, of which 3.2 million are in Irrigation Districts (IDs) and 2.8 are in Irrigation Units (IUs), The discrepancy in the total area is due to the difference between double--cropped areas and those that theoretically have irrigation capacity; some areas with irrigation capacity lie faUow and othenl are not irrigated for lack of water.  57 about 70 percent of agricultural exports. The main irrigated crops were maize (21 percent of the value of output in domestic producer prices), wheat (18 percent), sorghum (13 percent), beans (6 percent), safflower (3 percent), and soybean and sugar cane (6 percent each). The remaining 27 percent includes forages, seeds, fruits, and vegetables. II. Problems of Irrigated Agriculture 4.5 The chief problem of irrigated agriculture has been that the rate of irrigation development has slowed in the last decade, after rapid development from 1950-58 and 1971-82 (fable 1.1). The average rate of irrigation expansion was about 3.4 percent in the 1950s and rose to 3.8 percent from 1971-87 because of the growth in the early 1970s. Informal estimates of the rate since 1987 indicate that it is less 100,000 hectares per year, or at most 1.7 percent annually, and was negative in 1990 and 1992. A. Slowing Expansion of Irrigation 4.6 What are the causes of the slowing rate of new irrigation development? For the IDs, which are publicly financed, the proximate cause is less government spending, as seen in Table 4.1. A second cause is rising per cost per hectare, indicating that the cheapest projects had been financed and the new ones were, of necessity, put in more costly sites. At the same time that the rate of new IDs fell, the additions to the IUs also diminished, suggesting that they incurred the same rising costs per unit of new development. 30 4.7 A second cause that cannot be quantified exactly is rising environmental costs, though the general picture is clear. Though the national groundwater balance is positive overall as extraction is only 70 percent of natural recharge, the balance is negative in many regions. Much of the renewable volume remains unused in the wetter and less developed areas in the south; development of irrigated agriculture is not profitable there and the cost of transferring water from the surplus south to the deficit north is prohibitive. Most water use therefore occurs in arid areas, where the net recharge is negative and underground storage is consequently endangered. The aquifers in the north are over- exploited; groundwater levels are falling, in some cases by as much as 1.2 meters per year, and there has already been some sea-water intrusion in the northwest coastal aquifers. The over-exploitation of groundwater has accelerated water table depletion, resulting in additional pumping costs and decreased well discharges. Studies to provide fuller information on groundwater supply and demand are needed (see para. 4.59). 4.8 Expansion of the irrigated area ended with the financial crisis in 1982 and the consequent sharp reduction in public investment in irrigation. Public investment in the sector fell from US$3,600 million in 1981 to less than US$230 million in 1990. The economic crisis also affected maintenance of existing infrastructure. Hence, lack of funds cut the rate of growth of irrigated areas and adversely affected their productivity by reducing maintenance. 4.9 Expansion of irrigation confronts technical barriers in addition to the financial ones. The major irrigation works - large and small dams plus major groundwater facilities - have mainly been developed in the north of Mexico. Surface water is now increasingly scarce in the north because the most economical storage structures have already been built. Further significant expansion of the irrigated area in northern Mexico is therefore of doubtful economic and environmental feasibility. 30. There is no ~dy infonnation on the costs of developing the IUs.  58 Moreover, competition for water from urban areas in the arid zones wiH limit irrigation development from both ground and surface sources. 4.10 Much of irrigation in Mexico is gravity-based, with large schemes developed around storage reservoirs, diversion dams, and conveyance canals. In these schemes, water conveyance losses are about 40 percent, while losses in the minor canal networks amount to about 20 percent. At farm level, losses are normally between 30 percent and 50 percent. Overall conveyance efficiency is, therefore, between 24 percent and 34 percent. By comparison, efficiency ranges from 50 percent to 60 percent in California and Arizona, neighboring states that irrigate a given area with almost half of the water used in Mexico. This difference is due primarily to improved minor networks used in the US (lined or piped), better on-farm systems (leveled or pressurized), and better farmers skills. 4.11 A further technical constraint to efficient irrigation is that most Mexican IDs were designed and built to serve the old sectoral strategy of producing cereals and oilseeds. Consequently, until the late 1980s, irrigation systems were primarily designed to serve one grain crop at a time, through gravity applications of large volumes with long irrigation intervals, and minimal flexibility in delivery schedules. These rigid schemes were not compatible with high-value crops (fruits and vegetables), nor with the optimization of grain yields, as both require more frequent, smaller water applications. The recently approved On-Farm and Minor Irrigation Networks Improvement (OFMIN) Project will help improve flexibility in water delivery through completion of unfinished on-farm and minor distribution networks, and training of water users organizations (WUOs), farmers, and technical staff in better system operation. Table 4.1 Planted Areas with Irrigation and Public Spending on Irrigation, 1950-1992 Period Irrigation Irrigation Total Rate of Public Districts Units (ha) growth of spending (ha) (ha) total (annual (1992 N$ %) million) 1950-58 1,297 937 2,234 6.7 n.a. 1959-64 1,884 1,259 3,143 4.7 n.a. 1965-70 2,201 1,165 3,366 -2.3 n.a. 1971-76 2,649 1,379 4,027 7.2 3,801 1977-82 3,069 1,822 4,891 2.3 6,495 1983-8?' 3,460 1,675 5,135 1.7 2,303 b1 1989-92 3,511 1,894 5,405 -0.3 1,067 1950-82 2,136 1,278 3,414 3.8 5, 148c1 1983-92c1 3,488 1,767 5,255 0.9 3,752 NOTES: 0/ No data are available for 1988. except for spending. bI Public spending on irrigation not available for 1992. '" Average for 1971-1982. The breakdown between units and districts is not available after 1987; to calculate the averages for 1971 to 1992 the missing years for units and districts are assumed to be proportional to what they were in 1971 through 1986. SOURCE: World Bank, LA2NR Mexico Agricultural Database.  59 B. Cost Recovery 4.12 The importance of cost-recovery is that it is a sustainable complement to public funding and, if properly structured, gives incentives to conserve water. The general rule for pricing excludable services derived from public investments is to charge user fees that recover operating and capital costs. In practice, full cost recovery fees have not always been charged31 . Earlier studies (summarized in World Bank, Report 8967-ME, 1990 at 3-5) identified insufficient cost-recovery from producers in IDs as a contributing factor to poor maintenance of facilities. The average percentage of operation and maintenance costs paid by users in Mexico's irrigation districts was 48.2 percent in 1950-58, rose to 65.2 percent in 1965-70, and then fell to 50.5 percent in 1971-76 and to 26.0 percent in 1983-88. Since 1988, cost recovery has been much higher (see para 4.14). In contrast, cost-recovery from producers in IUs, which have been managed and operated by users, has been around 100 percent. Capital cost recovery has also been very poor. 4.13 In Mexico, there is now explicit acknowledgement of the need for full cost recovery. For example, the new Ley de Contribuciones de Mejoras mandates full cost recovery in all water related investment projects in the country. The only exception is for flood control projects. This new legislation applies only to new projects and to date no new projects have come on line under if:!. For already existing infrastructure, policy is to set user fees to cover operating costs only. 4.14 How well has the current administration's policy of increasing the viability and self-financing of water projects worked? By some counts, progress has been significant. From 1989 to 1992 collections from the derechos jederales del agua increased four fold to 1.977 billion pesos. CNA also reports that in those irrigation districts that have not yet been transferred to associations of producers recovery of operating costs increased from 18 percent in 1988 to 83 percent in 1992. However, there is a long road ahead to full recovery of operating costs. The magnitude of the problem in irrigation is revealed by the fact that 70 percent of all cost recovery at present comes from only three states (Sonora, Sinaloa and Tamaulipas). 4.15 Extremely poor service has made it difficult to force reluctant users to pay for services in some areas. In drinkable water systems, wastage as high as 40 percent due to leaks and line breakups have been reported. Cost recovery has also been made more difficult because of the lack of meters for individual users and even for entire municipalities. For irrigation systems, the difficulties are similar. 4.16 Cost recovery by CNA exists in different types of charges. The most important is the derecho de aguas, by far the largest revenue producer. Collections for derecho de aguas represent about two- thirds of all water collection charges (including VAT) in 1993. CNA also charges for the supply of wholesale water (agua en bloque) to water systems (13.2 percent of collections), for irrigation services (6.2 percent of collections) and for effluent discharges {less than 1.0 percent of collections).33 31. Fee for service has little tradition in the public sector besides infrastructure projects. For example, land titling and registration do not include user fees. 32. The policy objective is to recover the federal funds invested in projects after consideration of the socioeconomic characteristics of each area. The plan is to recover 90 percent of the investment funds over a period ranging from 25 to 40 years. 33. CNA also charges 10 percent of the commercial value of extracted river sand and several administrative fees, together representing 1.7 percent of revenues.  60 4.17 The effectiveness of cost recovery has advanced at different speeds from state to state and from urban to rural areas. The experience in the four states visited in the field illustrates this point well. For example, in the municipality of EI Llano in Aguascalientes there is a fixed water fee of N$27.5 per household because few housing units have meters, only approximately 60 percent of the households actually pay the fixed fee. By contrast, in the municipality of Aguascalientes water service is provided by an autonomous enterprise which has been quite successful with collections. In the state of Tabasco water charges cover on the average 30 percent to 50 percent of the operation costs only. (fabasco is behind many other states, because during the oil boom all services, including water, were provided free of charge). Turning around that tradition in more recent times has proved difficult, even though at present water is the only municipal service for which there is a (partial) charge in the state. One problem is the lack of adequate infrastructure, as many dwellings do not have working meters. In the state of San Luis Potosi drinkable water systems, according to the state authorities, recover only approximately 50 percent of the operating costs. At the other extreme of the scale is the state of Sonora where water charges cover operating costs. For example, in Ciudad Obregon there is no subsidy for operation costs and in Hermosillo the subsidy is very low. III. Recent Policy Changes in Irrigation 4.18 Policy changes in the irrigation subsector have followed the broader changes in economic policy. The Government's new irrigation policy is to encourage efficient water use through: (a) reforms in the legal basis of land tenure; (b) reforms in the legal basis of water rights; (c) a major increase in irrigation water fees to achieve self-sufficiency for Operations and Maintenance (O&M); (d) transfer of irrigation districts to water users organizations; (e) changes in economic policy to promote a shift away from basic grains production; and (f) more private sector participation in decisions and investment. 4.19 Incentives for private investment in irrigation were limited historically by legal, regulatory and administrative restrictions. The amendments to Article 27 of the Federal Constitution, and the new Agrarian and Water Laws of 1992 created a better environment for private investments in agriculture and specifically in irrigation. The new measures make land tenure and water rights more certain, and strengthen producers' organizations. In the context of the legal reforms, policy needs to promote new private investment in the sub-sector through expanding cost-sharing from beneficiaries and improvement in the regulatory and administrative frameworks. The development of tradable water rights and incentives to save water, which the new Water Law promotes, are expected to promote the more efficient allocation of water as rural and urban users become more familiar with the new market mechanism. 4.20 Management of IDs, including responsibility for O&M, is being transferred to the users, helped by the establishment of volumetric water a))ocation and delivery systems. Payment by beneficiaries of full water delivery costs, as well as the creation of a market for water, are a recognition of water's true scarcity value. These changes will generate incentives to invest in water saving technologies, and otherwise save water. 4.21 With regard to water use and quality, important changes have taken place, among them the decentralization of ID management into smaller hydraulic units (called "modules"), and the enabling of those holding water rights to transfer their titles. Private participation in financing, building, and operating water infrastructure is also being promoted. The Irrigation and Drainage Sector (lDSP) Project has made progress, as planned, in transferring management of modules to WUQs, in rehabilitating major canal networks, and in achieving full O&M cost recovery. Collection of  61 irrigation water fees improved substantially from 1989 to 1992 (self-sufficiency increased from about 18 percent (US$20 million/year) to 76 percent (US$85 million/year) of O&M costs. In all IDs transferred to WUOs, O&M self-sufficiency is now 100 percent; the policy goal is to transfer all 77 IDs to WUOs by 1995. 4.22 Irrigation reform requires complementary actions to ensure that farmers can adapt efficiently to a more market-oriented environment. Chief among those actions is completion or improvement of minor networks within each module and at the farm level. Remaining regulatory and administrative constraints must be lifted. In addition, technical support for IDs, and for the training of technical advisors and the WUOs, must be established to guide and support the subsector's adjustment. Technical support would increase farmers' ability to respond to market signals and to adopt appropriate technological changes. 4.23 Policies that encouraged the production of basic grains are being eliminated giving producers better incentives to shift into fruits, vegetables, forages, nuts, and other high-value crops. For example, official production targets were eliminated in 1992. With PROCAMPO, the domestic price of maize will fall by 70 percent and that of wheat by 20 percent. With NAFTA, prices of some exportables will rise, so that the joint effect (higher exportable prices, lower cereal prices) should induce a shift in the irrigated cropping pattern. 4.24 Electricity rates. A contributing factor in the rapid expansion of groundwater use has been low electricity rates. However, from 1990 to 1992, policy changes required farmers to pay higher electricity rates and therefore to modify their irrigation systems in order to improve water use efficiency, which helped rationalize groundwater use as many inefficient pumping systems became uneconomic. The Government's power sector target was to set rates at 100 percent of the long-run marginal cost (LRMC) of energy by 1997. Consistent with that target, energy rates rose by 440 percent in nominal terms from May 1990 to January 1993, reaching 54 percent of LRMC. In February 1993, the nominal monthly increase slowed down to 0.5 percent, lower than the inflation rate. On October 5, 1993, the 0.5 percent per month increase was stopped. At inflation of 9 percent and no more increases in the electrical rates, stopping the monthly increases implies that the fees would be reduced by 9 percent per year in relation to the real LRMC. IV. Water Markets 4.25 The realization that public agencies do not always have adequate resources, incentives or information to allocate water efficiently has encouraged the development of private mechanisms to manage water. Examples are WUOs, other autonomous entities, private firms and water markets. Private enterprise has been important in expanding groundwater use, while WUOs have growing responsibility to surface water systems in many countries. 4.26 The legal framework for private water transactions was often weak even where trades had gone on for years. In Mexico, unofficial water sales occurred but were officially discouraged or even banned. Water rights or concessions were uncertain and therefore inadequate to facilitate permanent water exchanges. Uncertainty about the legality of water exchanges imposed additional transactions costs and caused less efficient water allocation. 4.27 An efficient water allocation is defined as one that equalizes the marginal value products of different uses, net of transaction costs between uses. Water allocation by markets can produce efficiency gains - that is, a movement toward equalization of marginal products - in comparison to  62 administrative allocation. The efficiency gains derive from several features of markets. Market allocations are based on information available to farmers, and other private market agents, that is too expensive for public irrigation agencies to obtain. In addition, users will resist administrative water allocation since they generally receive no compensation for water lost by them as the result of a reallocation. Examples of the benefits from water trades are many. Chan and Griffin (1992) estimated sales of water rights in Texas to be worth from $3,000 to $16,000 per looonr. Vaux (1986) calculated that trades within California agriculture would produce net gains of $10 million annually in 1980 prices. Gains may be much larger between sectors - e.g., between farming and industry - than within agriculture alone. A. Costs of Establishing Water Markets 4.28 In analyzing water markets, it is necessary to consider transactions costs, institutions governing the market, problems associated with groundwater use, water quality, social aspects of water use, and the overall distribution of rents from water. 4.29 Transaction costs. Transaction costs vary with the type of water use. Those among farmers within an irrigation district are likely to be small because the distances to transfer water are short, irrigation facilities are usually compatible, and costs of knowledge about applying the water are already sunk to some extent. However, if there are many small farmers, then there may be increased costs of measuring water use and of enforcing contracts governing use. Such within-district trades were probably common in Mexico, even before the change in the law, so implicit contracts already exist. A second transaction occurs between rural and urban users. These costs of metering and contracting are significant. Transport and management costs are higher than those within districts. Transactions between rural and urban uses were in fact practically non-existent before the new law. The new law permits them, but available information about physical transaction costs does not allow us to say jf they will become common. 4.30 Establishing new institutions. Several institutional arrangements are required if markets are to operate effectively. First, transferable water rights must be established based on the volume of water owned or on the share of stream or canal flow. These rights must be recorded, secure, enforceable, and tradeable. Second, a governing institution, such as the CNA, must be able to resolve conflicts over water rights and to provide guidelines for dealing with third party effects - for example, downstream water users who are not directly involved in a water transaction, but whose water use may be affected by it - caused by return flows and other discharges into water sources. 4.31 It is difficult to estimate empirically the actual costs of water trades because they depend on infrastructure and management, inter alia. Good management and operating procedures can compensate for poor infrastructure and vice versa. Consequently the cost of water trades will vary from district to district but should be lowest in those units with flexible operating procedures, good water control devices, and strong WUOs. The presence of reservoirs to store water for irrigation will also facilitate trades as will information concerning water availability for the cropping season. If water users have concessions to a certain quantity or share of the water in a reservoir then these concessions can be traded among users in the system on different canals and at different locations. The traded water can then be delivered to the appropriate canals assuming that there is enough capacity in the canals to carry additional water. When reservoir storage or adequate canal capacity do not exist in the system then trades will be more difficult among canals in the system. However, even with run of the river irrigation, an effective WUO can make adjustments and broker trades for farmers on different canals along a river.  63 4.32 Externalities involving groundwater. A third set of considerations involves groundwater. Groundwater is a special case because: (a) it is always a common property because of connections among aquifers; (b) its quantity and condition are more costly to monitor than those of surface waters; and (c) its connection to surface waters produces complex interactions in ownership and management. An example of the last issue is that surface water rights are not secure if adjacent groundwater use can tap surface water. Incompatible surface and groundwater rights have caused serious management problems in parts of Arizona that are similar to northern Mexico. 4.33 Costs of measuring water quality. Putative market agents must also consider water quality. If this is not done, the quality of water supplied may be unusable. An example has occurred in the Rio Lerma basin of central Mexico where sewage has contaminated water used to irrigate vegetables. The costs of measuring quality are an important, permanent, transaction cost in water markets. 4.34 Social issues. If there are important societal uses that cannot compete in the water market, then the public sector might be justified in buying those rights or in reserving them from the initial allocation. An example is government action to preserve traditional stream fishing rights. Non- governmental organizations might similarly buy or reserve water rights or concessions for other environmental purposes. 4.35 Unequal land distribution. If strong de facto water rights already exist, then the primary task in establishing a water market is to formalize those de facto rights. Where de facto rights are weak or absent, then a common solution is to allocate water rights based on existing land rights in the area to be irrigated. This is efficient if the distribution of land is reasonably equitable and the area to be irrigated is easily defined. If this is not the case, such as in the Central Valley of California where large scale farmers captured a large share of the benefits from subsidized public irrigation, then an alternative allocation criteria or a land reform is needed. One alternative is the allocation of water rights to all families (landowners and landless alike) in the irrigated area, as was done with a small village irrigation project in northern India (Joshi and Seckler, 1982). In this case, a water market was used to match water rights with land rights after the initial allocation. An alternative is to allocate water rights to a public water management institution, such as a WUO or a river basin authority, and sell those rights to farmers, thereby allocating the rights and providing initial capital to the management institution. In Mexico, the practice is to distribute water concessions in proportion to land holdings, recognizing the de facto water rights that already exist. B. Experience in Developing Countries 4.36 There are two broad types of water markets. One is spot market sales of water and the other is permanent sale of the water right. In India, water markets have evolved where farmers owning tubewells sell water to farmers without wells. When groundwater is adequate and there are many well owners, the price of water is close to pumping costs (palanisami and Easter, 1991). When groundwater is limited the situation is different. Water levels then fall and well owners must either deepen their. wells or stop irrigating. Water prices typically exceed pumping costs in that situation. 4.37 In the large surface water systems of Pakistani farmers have rights to a certain time period for irrigation. The volume received depends on the flow but whatever the flow is during the farmers' allotted irrigation time he has the right to use it. Water trades can involve all or part of the individual irrigation allotment. Yet, even on an individual watercourse, the coordination required among farmers can make it difficult to transact trades. If there are other farmers on the watercourse in between the two farmers who want to trade, then these farmers must also agree to the change in irrigation time. The fact that such water trades are illegal in Pakistan makes it difficult or impossible  64 for government irrigation officials to help in the coordination of changes in delivery schedules. Still numerous trades occur indicating that both buyers and sellers receive significant benefits from these trades (Renfro and Sparling, 1986). 4.38 Active water markets developed in Chile after the passage of a new water law in 1981. The water trading has been most prevalent in areas of water scarcity and among farmers. The water market is facilitated by a separation of water rights from land rights. In most river basins, the water rights are in terms of a share of stream flows. However, in one basin where the streams flow has been stabilized by reservoirs, the rights have been translated into volumetric terms much like the Alicante system in Spain. These rights are only use rights since the actual water rights are retained by the government. Yet what is critical for trades is that, these use rights are recorded, enforced, secure, and tradeable among any users including those outside the basin. 4.39 The benefits of water markets in Chile appear substantial because land reform caused a mismatch between land and water rights, thus creating a natural exchange. In addition, the mismatch has allowed urban sectors to obtain water. However, several aspects of these water markets are raising concerns among government officials. First, is the problem of water quality; water owners only have a right to a quantity of water. Second is the large initial allocation of free "surplus" water to privatized power companies. Although the production of hydropower does not consume water, it does change the timing of flows with adverse impacts on downstream users. Power companies want to save water for the winter by holding it in their reservoirs during the summer when the farmers need the irrigation water. Third, owners of water rights do not have to consider their impact on downstream return flows when they sell their water rights. They can transfer all of their rights to users in other sectors or other basins and downstream users appear to leave very limited legal recourse. Fourth, low taxes on water holdings make it attractive for some owners to buy and hold large quantities of water without using them. 4.40 Estimates of Benefits to Mexico. A CNA study of the feasibility of water markets in several major irrigation districts - RCo Colorado, RCo Yaqui, Culiacan, Delicias, RCo Mayo, Rio San Juan, and Upper Rio Lerma - indicated that farmers commonly lease and sell water in a type of spot market. In Rfo Mayo, WUOs already trade water by volume because the district has measurement devices in the lateral canals. In the Rio San Juan District the scope for water trading is much more limited because allocation is based on an inflexible rotation. 4.41 The study found that irrigation infrastructure affects actual trades. In the Rio Lerma the poor condition of infrastructure and lack of measurement devices affect trades. The leasing of water was greater in those IDs with more diversity in cropping patterns, suggesting a significant difference in the opportunity cost of water among farms, a possibility seemingly confirmed by water lease prices ranging from $25 to $126 per 1000 m3 • The Delicias District had the widest range in prices among the districts studied of from $43 to $126 per 1000 m3 • This does not provide an estimate of the potential benefits because we do not know the quantity of water that would be leased at each price, but the different rates for water leasing show is that it is possible to seH water; a reasonable presumption is that the resulting benefits to farmers exceed transaction costs. The CNA study of water markets did not estimate the economic benefits of markets and did not compare alternative allocation criteria, nor did it estimate the long term benefits of greater private investment in infrastructure.  65 C. New Institutional and Legal Arrangements in Mexico 4.42 One of the outstanding questions in refonning Mexico's irrigated agriculture is the legal and administrative basis of water markets. Mexico promulgated a New Water Law in December 1992. Because of legal restrictions under the old Water Law. pennanent water rights were not traded. The new law makes it legal to lease or sell water. It establishes water concessions on a volumetric basis separate from rights to the land where the water is used. Are the new law and its supporting regulations adequate to pennit efficient allocation of water within agriculture and between sectors? 4.43 Water ownership. The actual water rights are held by CNA for the Mexican people. CNA grants use concessions made for 5 to 50 years; most are for 50 years. These concessions will be renewed generally if the use and discharge have not changed. CNA has to approve renewals and wiII consider the overall river basin supply and demand for water before making renewals that involve changes in use or discharge. 4.44 Eligibility for concessions. CNA can grant water concessions to a variety of entities including WUOs, individual users, private firms and municipalities. Initially the concessions will be made based on historical use; in the case of irrigation most concession will be made to WUOs, IDs, or IUs. The IDs, WUOS and IUs will then determine the allocation of water concessions to individual farmers and WUOs. The idea of water concessions is not new to water users in Mexico since the concept of separating the water right from a use right (concession) was introduced during the Spanish period. 4.45 Registering a water concession. CNA registers concessions and water sales. CNA is developing a simplified registration procedure; and has an incentive to facilitate the process since concession owners must pay a water use fee that will help finance CNA. The use fee is in two parts, one for the volume extracted and the other for the volume and quality of the discharge. Currently the water extraction fee for irrigation is zero. 4.46 Concessions are based on consumption. This means that water sales are in tenns of the quantity needed for plant growth, plus the quantities in transpiration and evaporation. Water lost to seepage cannot be sold. Basing the tradeable water concession on consumption use will reduce third party effects and protect those users who depend on return flows for some or all of their water supply. The water must be used for three years to maintain the concession. In practice, this means the concession owner must pay an extraction fee. The fee requirement is to prevent speculators from holding water free of charge and not using it. An important question about the three-year limit is whether or not instream uses such as fisheries and aquatic ecosystems qualify as recognized uses. If instream uses do not qualify, then this limit could pose problems for aquatic ecosystems in Mexico. 4.47 Water allocation rules. CNA has flexibility in allocating water because the concession is based on volume, not shares. Not only will CNA specify the quantity to be delivered to each irrigation district (ID), it also decides how surpluses (the volume available minus the sum of concessions) and deficits will be allocated. Since quantities can only be specified on some probability bases, rules or rights must be developed that recognize the highly variable nature of water supply. If the water concessions were based on shares then the users would have the right to determine how the surpluses are used and how shortages would be shared among users. In the case of Mexico these allocation decisions will be made by CNA. 4.48 . Water sale mechanisms. Since CNA grants concession to IDs or WUOs, concession allocations among individual users vary according to the internal regulations of the IDs and WUOs. If  66 concessions are reallocated to individuals, the latter can sell them to other individuals or organizations within the 10 or WUO. The individual farmer then receives the revenue from sales within the district. When water sales do not change the intake or discharge then the only requirement is that they be recorded. 4.49 Two-thirds of the users voting in the election must approve sale outside the district. External sales also require CNA approval. Sale revenues return to the district and not to the individual. These restrictions will limit sales between districts and will cause most sales to occur within the district. Consequently, the marginal value of water will tend to be equalized within an ID more than within districts. At a later date if IDs and urban areas combine there will be greater opportunity for trades between agriculture and urban users. 4.50 Under the 1992 law, much will depend on how CNA administers requests for sales among districts and urban areas. It will also depend on how concessions are allocated among various individual water users. If IDs or WUOs retain some of the water concessions this will facilitate trades between districts and urban areas, since the districts receive the revenues and thus have the primary incentive to sell water outside the district. The problem with IDs or WUOs retaining some or all of the water concessions is that it will eliminate pricing incentive for farmers to use this water efficiently. Only when farmers receive the revenues from selling the water concessions will they make decisions based on their economic value. D. Is the New Law Adequate to Create an Efndent Water Market? 4.51 The new water law is an important step in improving water management. However, the law's effectiveness depends on its implementation. A strong feature of the law is transferring water management to the WUOs, which take responsibility for maintaining the system and allocating the water locally. They are, therefore, of critical importance for the effectiveness of any water market. In many cases they will need to make infrastructure changes or change allocation rules so that the farmer who purchased the water can receive the additional supply. How easily a WUO can make such changes will affect the transactions cost of trades and the number of trades that are feasible to make. If the infrastructure gives the WUO only very limited control over the distribution, then it will be more costly to implement the changes necessary to accommodate sales. It also is not clear who would initially pay for the changes that may be necessary to allow water sales, but in the long run the ultimate buyer does. 4.52 The new law will probably limit trades between rural and urban areas to ones similar to those between Los Angeles and the Imperial Irrigation District of California. Los Angeles pays to improve the water use efficiency of the district and receives the saved water. Farmers lose no water as Los Angeles only receives the water saved. This arrangement improves the allocation of water between agriculture and the city, but not at the farm level as it provides no incentives for farmers to use the water more efficiently. The efficiency gains are owing to improvements in water delivery, such as canal lining and improved control structures, and not from improved on-farm water use. 4.53 Barriers to Achieving Benefits. The major barriers to water trades are the management capabilities of WUOs, the flexibility provided by infrastructure and rules used to allocate water among users, information concerning water supplies and water trading opportunities, and the type of water concession that users receive. Trades will occur when it is clear to water users that they will receive the benefits from the trades and that the traded water can be delivered. Thus the combination of management and infrastructure must be adequate to deliver the water while the farmers must have accurate information to make decisions on water use. Here the CNA can playa key role in providing  67 technical assistance to WUO and information to users concerning water supply, water technology, water use alternatives and water loses during the transport of traded water. V. Recommendations 4.54 Given funding and technical constraints, the most economic strategy for increasing irrigated production is to complete and improve existing irrigation systems, not to expand irrigated area or groundwater exploitation. This is essentially the Government's current approach. World Bank· supported projects already incorporate actions consistent with that strategy, including measures to help raise yields, improve water use efficiency, and promote enterprise diversification. This strategy therefore seeks gains from improved water management, increased cropping intensity, higher crop yields, and the conversion to higher-value cropping patterns. 4.55 Investment planning. CNA's investment planning continues to be hampered by the uncertainty each year about the size and timing of the budget. The budgeting procedures of the Ministry of Finance and Public Credit (SHCP) prevent rational investment planning. The CNA should have a five-year budget. 4.56 Cost recovery. The policy of full cost recovery only to new projects is wrong. Existing, and in many cases rapidly aging, infrastructure will have to be replaced and full cost recovery pricing is needed for that purpose. There is no good equity reason to charge higher prices to new users who are likely to live in previously under-provided areas, and to charge less to old users who have resided in areas that historically have been better provided. Full cost-recovery should be required for old and new projects alike. 4.57 Research Cor more efficient irrigation management. The supporting research system has to become more efficient. The major changes in relative prices (fertilizer in 1987-1990), maize (1994), water (after 1991), electricity (from May, 1990 until January, 1993) can only do so much to promote better efficiency in resource use. Existing institutions - the National Institute for Forestry, Agriculture, and Livestock Research (INIFAP), CNA, the private sector and producers' organizations - have failed to generate techniques to exploit Mexico's comparative advantage in labor-intensive methods. The recommendation is to privatize much of the public research stations in the north that now serve irrigated farming and to leave financing for those stations to producers' organizations. Private extension is also needed for farmers to improve water management and water trading. 4.58 Electricity tariff. The electricity tariff should be raised to the LRMC (see para. 4.24). 4.59 Groundwater action plan. The Government's policy for groundwater lacks full scientific and technical information. Studies needed to prepare the groundwater action plan should be completed expeditiously (see para. 4.7). 4.60 Water markets. Because of infrastructure costs, imperfect information, and third party effects, markets cannot solve all the water allocation problems in Mexico. However, if the necessary investments--notably in water metering infrastructure and information to producers about the law-are made, water markets can help better the efficiency of water use and allocation. CNA's role should be to provide overall water resources planning and information and to facilitate decentralized water management through markets and WUOs.  68 4.61 Within IDs the only requirement for trades should be that they be registered. For trades between IDs and between IDs and urban areas, CNA should devise a cheap process to approve trades. If third party effects are minor then such trades should receive quick approval by CNA. The review process should be transparent and involve all market participants. 4.62 Institutional changes are needed to achieve a better distribution of water among JD and urban areas. The current low water costs separate IDs and urban areas and do not encourage individual water users to trade. Mechanisms need to be developed that allow IDs and urban areas to combine into water districts when they use the same source of water. This would facilitate the exchange of water among rural and urban areas which will need to occur in a number of water short areas. Furthermore, mechanisms should be developed so that individual water users can directly benefit from the sale of water outside the district as long as there are no significant third party effects. 4.63 CNA should devise an open process for allocating water from surplus concessions and for distributing supplies during scarcities during the time it takes for a water market to develop. In the case of surpluses, sales by open bids, as suggested by the water law, should be used where possible. During droughts. though human and urban uses have priority over agriculture, allocation rules still need to be developed for agricultural water. These rules need to be created in consultation with users before drought. These rules do not have to be the same for all IDs since cropping requirements and farming condition are quite varied across irrigation systems. In systems with fruit trees and other permanent crops, provision might be made to save these crops and not irrigate low valued annual crops. In regions of annual crops a proportionate reduction for all farmers might work as long as farmers know about the reduction before the cropping season. The critical elements are full participation by water users and allocation rules that fit the cropping and other conditions in the region.  69 s. RESEARCH, INTELLECTUAL PROPERTY AND EXTENSION 5.1 An earlier Bank review of research and extension asked a few simple questions about how to obtain greater impact from those activities (World Bank, 1992). It also reviewed inteIlectual property problems as they might affect the generation and transfer of agricultural technology in Mexico. Here we update the answers to those questions, with appropriate reference to Chapter 9 on the special problems of tropical agriculture. I. Research 5.2 The basic question is what is the optimal public role in agricultural research? Related questions are: are current levels of funding adequate? Are there means to improve private financing or self-financing of public research? 5.3 The standard justification for an official role is a public goods one. This argument has applied, and continues to apply, in Mexico, but in an inconsistent manner. The inconsistency is that the richer, irrigated, export-oriented agriculture of the north uses privately-generated technology that the poorer, rainfed, subsistence-oriented agriculture of the south cannot use. Hence, the public goods argument for public research in the north is weaker than it is in the south. The disparity between the north and the south will widen given the likely effects of NAFfA on foreign direct investment in farming, which transfers technology with it. The public goods arguments for publicly funded or executed research in the center and south remain that: (a) the diverse agroclimate of the center and south prevents the direct transfer of techniques from other regions; (b) the same diverse agroclimate mandates local adaptive research of techniques borrowed from elsewhere; (c) domestic consumers benefit from agricultural research; and (d) there is no cost-effective mechanism to recover research costs on subsistence and environmental goods (e.g., basic grains, soil conservation) as there is for research costs on traded export crops (e.g., coffee and cocoa). The policy implications for public research (and extension) are to abandon the irrigated areas to privately-funded research 34 and to concentrate on the rainfed areas. The implications of such a shift are discussed in para. 5.7. 5.4 The main problem in the Mexican national agricultural research system is comparatively simple: lack of money. During the 1970s,'Mexico spent from 0.27 percent to 0.59 percent of its annual agricultural product on public research; this value is low in comparison to its competitors (World Bank, 1992), who often spend more than 1.0 percent. At the beginning of the reform (1982- 88), Mexico spent about 0.46 percent of agricultural GOP; from 1989-1992, Mexico has spent about 0.27 percent and the expenditure per scientific staff has fallen from 215,000 to 101,000 (Table 5.1). Moreover, private research spending in Mexico is less than in competing countries, so that does not compensate for the lack of public investment. 3s 34. The question of whether the research should be done by public or private enterprises is different from whether research should be financed by public or private agencies. 35. Many studies, in Mexico and throughout the world, have shown very high returns to agricultural research, so it is a profitable investment whatever its funding or management.  70 5.5 Lack of money means, first, that it is impossible to attract the best scientists; in fact, many staff are not even scientists. Second, staff of whatever qualifications have no operating monies. Lack of operating monies means, in practice, no support staff and inadequate quantities of materials, such as seeds, fertilizers and other chemicals, bags, paper, and most recently, equipment and software for data processing. Third, there is no money for investment in physical facilities, leading to a general decline in morale and efficiency. Fourth, there is no money to acquire new information from journals, paper and electronic databases. and other scientific sources such as Internet. Fifth, and most importantly, there is no money to train new staff at high academic levels. Academic training is even more vital today because of the accelerating transition from physical and chemical sources of output growth to biological ones requiring application of more sophisticated scientific capacity. Table 5.1 Agricultural Research Scientists and Expenditures, 1982 - 1992 Research Public research Spending/staff Scientists Spending (N$ spending/ago GDP (N$1992 Period (numbers) 1992 millions) (percent) thousands) 1983-88 1,718 328.6 0.46 215 1989-92 1,853 209.6 0.27 101 Source: INIFAP. 5.6 One justification for lower public research funding in Mexico is that it can borrow technologies developed from public or private enterprises in the industrialized countries. This argument mainly applies to the north of the country, where climate is similar enough to that in some parts of the industrialized countries so that agricultural technology can be borrowed. The argument is likely to be invalid in much of the tropical south, where the need for adaptive research is much greater. 36 5.7 While the priority ought to be to increase public funding to a level comparable to that of countries with similar incomes, the Government has shown no interest in this and is in fact is going in the opposite direction. Given this fact, what would happen if public research were to concentrate only on rainfed agriculture, on livestock production in states where rainfed agriculture dominates, and on forestry, how much should it spend?31 Table 5.2 shows that even reallocating the entire 1993 budget of INIF AP (about N$270 million) to the southern and central states would amount to 0.75 percent of the total value of annual and perennial crops alone. INIFAP's budget would amount to 0.87 percent of the total value of crops minus the value of coffee, cocoa, and sugar cane, three permanent crops 36. There are many examples of foreign crop production technology borrowed into the Mexican tropics giving much lower yields than at the site of origin because of the lack of good local adaptive research. Instances include pineapple (from Hawaii), mango (from Florida), coffee (from Brazil) and citrus (from the United States). Another positive factor might be the reform in the ejido legislation which, by giving better incentives for agricultural investment, might hasten technology transfer. 37. Forestry research, because of the time needed to get results and the severe environmental problems, will remain entirely public except on very large private forests.  71 for which most research costs should eventually be paid by producers organizations. To put it another way - even if there were no publicly funded research on livestock or forestry and all federal funds now spent on agricultural research were devoted to crops in the central and southern states - those funds would be less 0.9 percent of the value of crop output. 5.8 Compare the expenditure on research to PROCAMPO spending in the central and southern states on the area sown to maize and beans alone. 38 That area in 1992 was roughly 6.9 million hectares, giving at an average payment of N$350Ihectare, a total payment of N$2,426 million. That amount, which is less than the actual amounts that PROCAMPO will distribute in the central and southern states, is nine times INIFAP's total 1993 budget. Table 5.2 Value of Crops and Needed Federal Research Spending for Central and Southern States (N$ 1992 millions) Central states Southern states Total Average value of annual and perennial 23,725 11,226 34,951 crops, 1991-92 1 percent of average value 237.2 119.3 356.5 1 percent of average - value of coffee, 237.2 72.6 309.8 cocoa, and sugar cane Public research spending, national total, n.a. n.a. 209.6 average of years 1989-1992 Notes: States are as dermed in Table 1.2. They are regrouped here as Central: Aguascalientes, Zacatecas, San Luis Potosi, Colima, Michoacan, Jalisco, Guanajuato, Queretaro, Hidalgo, Tlaxcala, Puebla, Mexico, and MoreIos; Southern: Tabasco, Veracruz, Chiapas, Oaxaca, Guerrero, Campeche, Yucatan, Quintana Roo. 5.9 Increasing private funding. The earlier agricultural technology review (World Bank, 1992) concluded that a second priority was to improve private research funding. There are at least five sources of private funding. One is public/private foundations, called patronatos. The Agricultural Technology Project (ATP) supports a consultancy to develop patronatos where they do not exist, or exist but are weak. Initial indications are that this consultancy has been successful in identifying patronatos and in assisting some of them to become more active; an additional benefit has been the creation of a trust fund, the Mexican Foundation for Agricultural Research (FMIA) to increase public awareness about the benefits of research and to better funding. A second source is the privatization of parastatals and the transfer of their costs to producers organizations. This has been done, with poor results so far (see Chapter 9 about the tropics). A third source is service fees (laboratories, crop sales, and variety testing); while this is now available to the public sector it is of limited extent. A fourth source is seed production royalties or licenses (see para. 5.38); this was nil before the change in the seed law in 1991 and while it is now small, it is likely to become more important.» 38. PROCAMPO payments will in fact be made for nine crops, of which maize and beans are two. 39. A filth source is tax reform to allow greater deductibility of research expenditures; this is discussed in para. 5.56.  72 Priorities for public research 5.10 The analysis of priorities in research is analogous to the analysis of the allocative efficiency of firms. In research, the measure of allocative efficiency is often taken to be the allocation of research funds (for example, across commodities) in proportion to the value of output of each commodity. Ignoring the possibly diverse research opportunities across differing commodities and agricultural regions, a good fit between resource allocation and value of output is said to be a congruent one. A congruent allocation basically means that changing priorities would not produce a more efficient research system. The earlier review concluded that the fit between research fund allocations across commodities was congruent with the value of output of those commodities and therefore that adjusting priorities would not make for more efficient research. Because that conclusion was arrived at through a congruence analysis using domestic prices, it will change as a function of the change in value of output at world prices and at the new quantities that are produced at world prices. 40 Specifically, and based on the analysis in Chapter 3, free trade would change value and of output and proportionate research resource allocations, roughly as follows: (a) maize, -20 percent; (b) wheat, -37 percent; (c) sorghum, -20 percent; and (d) beans, -9 percent. 5.11 Despite the previous sector review's conclusions that there was a good fit between the value of output by state and the shares of different commodities in the research portfolio, INIFAP's research activities have been, and still are, spread across too many commodities. 41 The recommendation stands to reduce research activities in some commodities which can be studied by producers organizations (e.g., coffee, cocoa, some other perennial export crops) or in others of very minor economic importance. 5.12 Regional priorities (north/center/south). A second issue in the analysis of priorities is the congruence of allocations among regions: north/center/south. The earlier review concluded that INIFAP allocations by states shows that allocations were congruent with the south's needs. Despite that finding, the review concluded that public research needed to be expanded in the south because: (a) the north and center received more private research; (b) the south had received less public research in the past; and (c) the south's greater environmental complexity raises research and extension costs there. The earlier review noted that the necessary expansion to the south would require additional resources and could not be achieved solely by a reaHocation from the north. This is shown in Table 5.2 which sets a standard of 1.0 percent of the value of annual and perennial crop output in the central and southern states of the country. Even a complete elimination of research spending in the north and a reallocation to the south does not raise public spending in the center and south to 1.0 percent of agricultural GDP; the average public spending during the period 1989 through 1992 was only N$21O million, which is roughly 60 percent two-thirds of the value of annual and perennial crops in the central and southern states, and roughly 68 percent of the value of all crops less the values of coffee, cocoa, and sugarcane. 5.13 A recent policy change affecting research in the south is the removal of parastatals and the transfer of their assets to producers organizations. Affected parastatals are, notably, the Mexican Coffee Institute (lNMECAFE), the Institute for Sugar Cane Improvement (IMPA), the National Cocoa Development Council (CONADECA), and the National Fruit Company (CONAFRUT), among 40. The earlier review concluded that INIFAP should do little or no basic research, but that it should rather build more effective collaboration with institutions doing basic research. This conclusion stands. 41. This is also the conclusion of the Policy Notes of 1990 (Mexico· Policy Notes on Agriculture, Food, and Rural Development, Report No. 8144-ME. February I, 1990).  73 others. In terms of the public goods rationale for research, these are appropriate reforms because the benefits to research on exportables can be largely captured by the producers organizations when the crops is marketed, and the costs of research recovered through voluntary taxes on members. In practice, this transfer has failed (see Chapter 9) because of the fall in world prices of coffee and cocoa and of the high initial costs of starting effective research and extension in inexperienced agencies. 5.14 A historical shortcoming of research in the south was that it has excluded cereals, oilseeds, and environmental goods. In contrast to the goods supported by producers organizations - coffee, cocoa, sugar cane, fruits, dairt~ - which are wholly marketed, cereals, oil seeds, and environmental goods are only partially marketed and in some cases, not marketed at all. The work in the tropics outside INIF AP, including that of the Post-Graduate College, the CNA, and the private livestock sector is not in varietal development or adaptation for basic crops. Therefore, a research strategy to improve grains, oiJseeds, and environmental goods will not benefit from those complementary efforts, and will require public support. 5.15 Subsector priorities Orrlgated/rainred). The earlier review observed that the historical research focus had been on irrigated agriculture, and that good productivity gains had been achieved. Because the transferability of breeding research for yield under irrigated to rainfed conditions was high - as seen in the wheat, maize, and bean cultivars used in Mexico - research under irrigated conditions partly benefitted rainfed areas. Despite the partial transferability of some irrigated results, the earlier review concluded that three facts make it impossible to rely on research done under irrigation in the north as a substitute for local research in the center/south. The main one is that research under rainfed conditions is difficult because of the difficulty of finding robust drought resistance. Second, transferring crop management research from irrigated to rainfed areas is costly. Third, the private research effort is negligible for some rainfed problems. Private firms will not study drought or heat tolerance in rainfed/tropical systems because the risks are too high. They will not study cropping practices or do adaptive research until resistant genetic materials have been developed by public agencies. These three facts justifying public research in the center/south still stand, but the government's strategy neglects them. 5.16 Disciplinary priorities. The earlier review concluded that the emphasis of public research had been too strong on plant breeding and too weak on crop management. Supervision of the ATP has shown that this error has not been rectified. Moreover, because private import of genetic material, resulting from changes in the 1991 Seed Law and from the expected consequences of NAFfA is going to reduce the importance of public plant breeding even more. 43 Hence, the need for a shift in disciplinary priorities away from plant breeding will become stronger. 42. Dairying is also supported to an important extent by private milk processors and by private sector animal health research, done both in Mexico and abroad. 43. The International Agricultural Research Centers (IARCs) are a principal source of germplasm for Mexican crop improvement. Mexico also receives training, new methods, and results from other countries through the IARCs. The chief barrier to better use of the IARCs is lack of funds on the Mexican side to use the IARCs in germ plasm and research methodologies. This is again quite different from countries with weaker scientific and technical capacity, in which the IARCs are substitutes for the national programs. The policy response has to be to raise the Mexican contribution to its own research system, so as to make more efficient use of external scientific contribution. The earlier review noted two criticisms of the IARCs' role in Mexico. One was that they could do more in the highland tropical and rainfed areas, but the review, based on an analysis of the IARCs activities, found that this was not a valid criticism. The other was that the Centers should do more in germpJasm bank development, but this did not appear relevant for Mexico which has access to adequate resources to develop gene banks on its own.  74 Individual commodity issues 5.17 The earlier review identified problems with research for individual commodities. This section updates recommendations for the major commodities - maize, wheat, beans, tropical livestock, coffee, and fruits and vegetables - which are more than half of INIFAP's portfolio - in light of recent developments and possible events related to NAFTA. 5.18 Maize. Maize covers about 40 percent of the annual crop area in Mexico and is grown in a dozen different environments. Despite some advances, productivity growth in most of the crop's rainfed environments has been less than that of irrigated crops. Because of the diversity of maize environments, a successful research strategy is particularly complex. The principal defect of the research/extension system has been, in fact, the failure to deal with this complexity in transfer the results of historical maize improvement. The indicator of this failure is the very high gap between experimental and farm yields in rainfed areas. While local successes have occurred, they have not spread widely. 5.19 The earlier review attributed this defect to not analyzing limiting factors in maize production at the farm level and concluded that the explanation of adoption failures should be a priority in maize research. Another shortcoming was identified in the neglect of hybridization for rainfed and tropical areas; the review concluded that the public strategy should concentrate more on hybrids for such areas. incentives to treat the cost/risk problem. 5.20 Remedying the failure to transfer breeding results will require more emphasis on crop agronomy in areas of high genetic yield potential. This implies stating regional priorities based on analysis of limiting factors, and then defining appropriate research strategies to lessen the effects of those factors. One illustration is to orient maize research toward the tropical and sub-tropical areas where competition from other crops (e.g., hybrid sorghum and rainfed wheat) is less. 5.21 Changing the breeding strategy is also necessary, especially putting more emphasis on hybridization. This can be done as part of a strategy of using additional resources to produce breeding materials for private sector single-cross hybrids in irrigated environments, for single and double-cross hybrids for high potential rainfed environments, and for synthetics and open-pollinated varieties for environments. 5.22 With respect to relations to other branches of the research system, INIFAP and other public maize breeding efforts can continue to rely on the International Wheat and Maize Improvement Center (CIMMYT) as their principal source of new germplasm. In collaboration with Mexican universities, more basic research on special topics is needed, especially for the tropics and rainfed areas. Where there are commercial materials, then public improvement programs have to make more explicit comparisons with them to ensure that their own materials are competitive. 5.23 Private contributions to adaptive maize research and extension will initially be small. The risks of adapting germplasm are high, and private firms are typically unable to capture returns to agronomic and farming systems research. However, advances in public research will engender subsequent private initiatives linked to supply of seeds, fertilizers, and other agricultural chemicals. Thus, the public research sector has to nurture close relations with private input suppliers so as not to lose the benefits of future interactions in adaptive research.  75 5.24 Wheat. Wheat research and extension have been the principal successes of the Mexican program. However, growth in wheat yield potential has begun to slow, the irrigated crop faces competition for water, and the rainfed crop lags in potential and actual yield. With freer trade in wheat, and with rising water costs, wheat will become much less competitive and should receive fewer research resources. This will comprise the expanded effort that is now required to keep raising yield potential and to maintain the gains, especially in disease resistance, of the past. While Mexico can continue to expect a major contribution from CIMMYT to its wheat program, CIMMYT's financial crisis suggests that Mexico will have to pay for more of that research at a time when it is less in Mexico's interest to fund wheat research. Rainfed wheats have benefitted from transfer of irrigated materials, achieving good yield growth since the mid-1970s. This trend is unlikely to continue for much longer without more emphasis on earliness, drought tolerance, and agronomy. Rainfed wheat research, like irrigated, should receive fewer research resources at anticipated world prices, however. 5.25 Beans. Beans are grown allover the country and cover 10 percent to 15 percent of the annual crop area. Research has developed many varieties and corresponding agronomic recommendations for the main types. Despite a long-term effort, promoted with generous financial incentives to producers, bean research has largely failed to have a production impact. Introduced varieties have either not been adopted or, where they have been adopted, have not raised yields very much. Though beans are the chief example of how more extension and adaptive research are needed to adapt station technologies to farm conditions, the many previous failures ought to impose a more critical attitude in publicly-funded research to the real likelihoods of success. 5.26 The expected drop in bean output of 9 percent puts even greater pressure public agencies to consider the economics of bean research. An important technique in setting further bean research priorities will be to compare them to other opportunities - crops, livestock, and forestry - in the same environments. 5.27 Another issue is the lack of successful international and private research on beans, suggesting a defect in the basic research on the crop. While some of the failure to import and adapt research results is owing to the crop's relative lack of importance in much of the temperate world, there is still a world-wide difficulty in raising yields. Therefore, it may be important to allocate more resources to very basic research outside INIF AP, in collaboration with universities and scientists in Mexico and abroad. 5.28 Tropical livestock. Livestock have been about 30 percent of the total gross value of Mexican agricultural output since 1950, and livestock research is an important share of the public portfolio. The principal area with unresolved research problems and extension is the tropics, where the limiting factors to livestock production are feed availability, the adaptation of improved breeds, and disease. The previous review concluded that livestock research could be made more efficient by short-term shifts in priorities, principally on primary production, which is always limited in the tropics by variable moisture supply and low soil fertility. Further expansion of work on primary production is therefore needed. 5.29 Coffee. Coffee is Mexico's largest agricultural export. Though farmer recommendations have been available since the 1950s for cultivars and agronomic practices, the impact of research has been small. The diffusion of new genetic materials is inherently slow, because of the permanent nature of the crop, so research efforts always take longer to manifest themselves. Another limiting factor was the near-total absence of private contributions to research. The previous recommendation was to merge the coffee research program of INIFAP with that of the parastatal, the Mexican Coffee Institute  76 (INMECAFE), into one research and extension agency. With the liquidation of INMECAFE, this has been done de facto, though without additional resources for INIFAP. The government has transferred most of the resources of INMECAFE to the coffee growers organizations but, as Chapter 9 shows, the latter have made little or no real beginning in research and extension. The basic argument for public research, especially the risky very long-teno nature of coffee, still implies some public support to coffee research. 44 5.30 Fruits and vegetables. Temperate fruits and vegetables. nearly always grown with irrigation. will continue to depend on private. largely imported research. The role for public research will largely be one of cropping practices and water management; even then. the public role may be only to execute work financed by producers organizations. as the latter have sufficient resources and technical skills to contract research with the government or with other private organizations. With respect to tropical fruits and vegetables. which are usually rainfed and whose technical basis for rapid growth is weaker, the public role should be to promote producers organizations. but not to be directly involved in their management or, after an initial grant period. to be involved in their financing. II. Intellectual Property Issues 5.31 Intellectual property and fiscal barriers may hamper private and imported research. The past two governments' insistence on private participation in the economy and on a reduced state role has lowered those barriers. Are there remaining legal barriers that could be eliminated so as to elicit more private partiCipation in research and extension? What policies affecting intellectual property rights and research investments could encourage private activity? The main intellectual property issues are plant variety protection (PVP) in the context of seed policy, general patent law, and trade secrecy." 5.32 Mexican law has changed dramatically to favor the private development and use of agricultural technology. The possibility for private sector participation in the development and marketing of new plant varieties, including those derived from INIFAP, has been widely expanded. The scope of the intellectual property law has been increased and barriers to technology transfer have been reduced. A. Seed Policy 5.33 Much of the growth of Mexican crop production resulted from higher-yielding plant cultivars. Despite the importance of those cultivars. the earlier report argued that deficiencies in seed policy had slowed the development and spread of promising materials. The earlier report stated that the "public seed system could be altered so as to allow greater private participation, while improving the efficiency of remaining government activities." The earlier review identified three key reforms: market regulation, the role of the National Seed Production Company (pRONASE). and scientific policy affecting the seed industry. Will the reforms in the 1991 Seed Law result in cheaper seed-based technologies for producers? 44. The FAO/CP paper (see Chapter 9) on the Mexican tropics found that research, both public and private, was inadequate to deal with the problems of many important tropical crops, especially ones with growth potential. In addition to coffee, the crops mentioned included cocoa, citrus, mango, and banana (especially on environmental issues). 45. The "Intellectual Property Issues· section (paras. 5.31 to 5.80) is the work of Professor 10hn Barton, Stanford University Law School, who was a Bank consultant.  77 5.34 Market regulation. The basic choice was between a United States-type regulatory structure, in which the seed industry is governed only by producers' beliefs about seed quality; or a European structure, in which a public agency certifies seed. The 1991 Seed Law is a compromise between the two. A category of certified seed has been retained. A new category of "verified seed" has been created that allows any private firm to sell seed that it verifies the viability of the seed. The process of varietal certification has been made more transparent. Revisions in the variety trial inscription or import/export procedures must be published officially, must be made available to the entities entering materials for inscription, and must include comprehensive, timely, statistical analysis. With respect to quality control, the capacity of the National Seed Inspection Service (SNICS) has to be strengthened, as it is not now capable of fulfilling its mandate. SNICS would have to become an independent Federal agency with budgetary autonomy, with increased staff and more independence from seed producers. 5.35 The role or PRONASE. While PRONASE has not been privatized, it has been put on a competitive basis with private firms. The market sets seed prices, subsidies to PRONASE have been eliminated, and PRONASE's preference on INIFAP materials has been eliminated (see para. 5.36). 5.36 Scientific policy. The 1991 Law on the Production, Certification, and Commerce of Seeds (July 15, 1991) has eliminated PRONASE's preferential access to INIFAP materials. It states (Article 4) that the varieties created by SARH may be transferred to any person interested in acquiring seeds in the basic, certified, or verified categories for their commercial reproduction and their commercialization. On the basis of this freedom, INIF AP now sells basic seeds to PRONASE, private firms, and farmers associations using a form agreement under which the purchaser pays five to seven times the commercial value of the seed and receives, in addition, a royalty of 3 percent of the price to the farmers. The transaction is essentially a sale of the basic seeds; it includes no promise of exclusivity. 5.37 This reform will have several positive effects. First, by giving Mexican seed companies better access to INIF AP materials, it will allow those companies to compete with international firms in Mexico. Second, it may provide license fees from cultivar development to be re-invested in INIF AP (see para. 5.38). Third, responsibility for producing foundation seed from breeder's seed was taken from PRONASE and given to INIFAP. Lastly, restrictions on international centers releasing germplasm in Mexico have been eliminated, with the exception of plant health regulations. 5.38 INIFAP has already (January through August 1993) received some Mex N$8 million through sales of basic seed, which have been to 26 different institutions. The value of sales is significantly more than the production cost of the basic seed, but is much less than the total seed production cost, including research. INIFAP considers the stability of the firm which is expected to distribute the multiplied seeds. In situations of scarcity, INIFAP is more likely to sell to PRONASE. INIFAP has not decided how to handle the situation in which a firm, e.g., a foreign one, uses its seeds in a breeding program. Nor has it decided how to handle the upcoming question of the distribution of the varieties after Mexico adheres to the International Convention for the Protection of New Varieties of Plants (UPOV), although it apparently expects to acquire the PVP certificate itself and INIFAP staff breeders are raising the question of what rights they will have in such varieties. 5.39 Variety screening. As under the 1960 seed law, all seed distributed or imported for use (as distinguished from seed imported for experimentation) must be of varieties approved and entered in the National Register of Plant Varieties (RNVP). The requirements are that the seed's properties be described in an application; the seeds can be inspected to ensure that they correspond to the description. According to the law, the Consultative Committee on Plant Varieties (CCVP) evaluates  78 the seeds when there is any doubt as to whether they satisfy the screening requirements. The CCVP includes 10 members: 3 public officials, 3 representatives of the seed producers organizations, 3 representatives of seed users, and 1 representative of technical or scientific institutions. The CCVP maintains a register of recommended varieties. This register is significant to the farmer, because it is used by lending agencies as a condition to granting credit. In approving entry onto this list, the CCVP requires three years of testing (some of which can be in foreign locations) and it requires that the seeds be superior to previous varieties in yield, product quality, and disease resistance, with some possibility of weighing among the different categories. The decision is made on the basis of information supplied by the public or private breeding institution. 5.40 Plant Variety Protection. PVP is not recognized in Mexico. While its historical absence has not affected incentives for seed-based research and extension, the legal situation will soon change. Plant varieties are currently protectable under !be terms of the 1991 Law on the Development and Protection of Intellectual Property. Moreover. the intellectual property chapter of NAFTA requires that each party provide substantive protection equivalent to that of the UPOV Convention and shall make every effort to accede to the Convention (Article 1701). According to Annex 1701.3, Mexico has up to two years after signature of NAFTA to comply with this requirement, but must begin to accept applications from plant breeders in all plant genera (for later grant) as soon as NAFTA enters into force. NAFTA was signed on December 17, 1992; assuming that it enters into force, Mexico must thus adopt PVP legislation by December 1994. 5.41 Mexican officials are acting on the assumption that the plant protection provisions of the current law are inadequate and that a new law will therefore be needed. In adopting PVP legislation, Mexico will have a choice between the 1978 and the 1991 versions of the Convention. The latter version gives somewhat stronger rights to the plant breeder. This is an issue which clearly deserves immediate study, possibly by the Instituto Nacional de Propiedad Intelectual created by the 1991 Intellectual Property Law (see below). B. The Regular Patent System 5.42 Under the 1987 Patent Law, the foJIowing were not patentable: (a) plant species, animal species, their varieties, and essentially biological processes for obtaining them; (b) foods and drinks for human consumption and processes for obtaining or modifying them; (c) biotechnological processes to obtain pharmaceuticals, medicines in general, beverages or foods for animal consumption, fertilizers, pesticides, herbicides, fungicides, and products with biological activity; (d) genetic processes to obtain plant or animal species or their varieties; (e) chemical products; and (t) chemical pharmaceuticals, medicines in general, beverages or foods for animal consumption, fertilizers, pesticides, herbicides, fungicides, and products with biological activity. The last four restrictions were to expire in 1997. 5.43 This structure was completely replaced in 1991 with a new Law on the Development and Protection of Intellectual Property (Diario Oficial, 27 June 1991). Under this Jaw, the corresponding provisions (Article 20) make the following patentable: (a) plant varieties; (b) inventions relating to microorganisms, such as those which use them, those which apply them, or those which result in them. Included in this provision are all types of microorganisms such as bacteria, fungi, algae, viruses, microplasma, protozoans, and, in general, cells that do not reproduce sexually; and biotechnological processes for obtaining pharmaceuticals, medicines in general, beverages or foods for animal or human consumption, fertilizers, pesticides, herbicides, fungicides, or products with biological activity. Pharmaceuticals, medicines in general, beverages and foods for animal or human  79 consumption, fertilizers, pesticides, herbicides, fungicides, or products with biological activity are now patentable. 5.44 The following will not be patentable: essentially biological processes for the development or reproduction of plants, animals, or their varieties, including genetic processes or those relating to material capable of reproducing itself, in itself or in any indirect manner, when they consist simply in the selection or isolation of available biological material and allowing it to act under natural conditions; plant species and animal species or breeds; biological material as it is found in nature; genetic material; and inventions relating to living material making up the human body. Certain inventions of biotechnological importance remain unpatentable. Of these, the most significant for agriculture are genes, which are clearly meant to be excluded by the term "genetic material" (Iniciativa de Ley de Fomento y Protecci6n de la Propiedad Industrial.) 5.45 There are important restrictions on patents involving living material. Article 22 of the New Intellectual Property Law states that a patent holder shall have no right to prevent: (a) a third party who, in the case of a patent related to living matter, uses the patented product as an initial source of variation or propagation to obtain other products, unless such use is carried out in an iterated way; or (b) a third party who, in the case of a patent related to products which make up living matter, places in circulation or commercializes the patented products, for purposes other than multiplication or propagation, assuming that the products have been legitimately introduced into commerce by the holder of the patent or a person to whom a license has been given. 5.46 This exclusion is likely to be significant for the development of transgenic agricultural varieties. An entity that isolates an agriculturally important gene and inserts it in new varieties (e.g., the toxin from Bacillus thuringiensis into a crop to confer pest resistance) is likely to want regular patent protection on use of this gene to protect its research investment against a breeder who transfers the gene from such a transgenic variety into other varieties. Under current Mexican law, protection of such a biotechnology invention is limited both by the exclusion of genes from patentability and by the maintenance of a third person's right to breed with patented live material. 5.47 It is of course important that a gene patent not reach use of the gene as found in nature- and the sensitivity of Mexico to exploitation of its genetic resources may be one of the reasons for the exclusion of patentability of genes. In industrial nation patent systems, however, a reasonable restriction on gene patents is generally achieved by the law's requirement that the invention be novel. The typical claims in a gene patent cover the gene sequence as isolated (on the logic that the gene was never in nature in isolated form), various plasmids and gene constructs containing the gene (which were certainly never found in nature), and plants - including their descendants - which have been transformed by use of the gene construct. Thus the patent reaches use of the gene in genetic engineering and reaches breeding with the genetically-engineered plants, but does not reach breeding with plants that naturally contain the gene. This interpretation of the novelty requirement therefore creates incentives for investment in research on genetic engineering while not restricting traditional breeding with natural organisms. 5.48 The inability to patent genes is a limited omission in a major change in Mexican patent law. The new law clearly permits patenting of pesticides and the like; it is significantly stronger than its predecessor and creates a variety of new incentives to support agricultural research. The overall number of applications has been increasing by about 25 percent each year since the passage of the new law. Nevertheless, the law appears weak for the coming new generation of biotechnological agriculture in which Mexico may want to encourage the development or acquisition of proprietary genes for insertion into plants.  80 5.49 NAFfA and intellectual property. Under the intellectual property provisions of NAFrA, the biotechnology-oriented exemptions may have to be narrowed. NAFrA (Article 1709) requires the patentability of .. any inventions, whether products or processes, in all fields of technology, provided that such inventions are new, result from an inventive step and are capable of industrial application. " Exclusions are permitted for "plants and animals other than microorganisms," and for "essentially biological processes for the production of plants or animals, other than non-biological and microbiological processes for such production," but parties are required to permit the protection of plant varieties through "patents, an effective scheme of sui generis protection, or both ... 5.50 Although the issue is not absolutely clear, it appears that the exclusion of gene patents is not consistent with this NAFrA provision. Genes are not "plants" or "animals," and genetic engineering is arguably not an "essentially biological process." Because such patents are likely to be useful for Mexican agriculture, it is recommended that the exclusion of "genetic material," be reconsidered as part of the process of implementing NAFrA. Moreover, to make this change effective, Article 22 should be revised to permit a breeder to prohibit commercial sale of varieties that are bred from the patented varieties and include the genes inserted through biotechnology. 5.51 Before 1991, the patent law contained a compulsory licensing provision applicable if the patent was not used within Mexico within three years of issue. Although such provisions are hardly ever used, they have been the source of substantial international controversy. The new law significantly narrows the possibility of such use by prohibiting the use of a compulsory license when the product involved has been imported into the nation (Article 70). Moreover, NAFrA includes other significant conditions narrowing the possibility of grant of such compulsory licenses. Trade Secrecy Law 5.52 Particularly in the process area, inventors have often protected inventions through secrecy. For example, holders of inbred parental lines for use in producing hybrids have frequently restricted physical access to these lines as a way of protecting their hybrids. To make such rights effective - and they form an important part of much transferred technology - the owners of the trade secret or proprietary material protect their interests through contractual agreements and through suit against those who acquire the information improperly as by theft. (There is never any trade secret right against a person who independently develops the invention.) 5.53 Historically, Mexico has been strongly criticized by U.S. firms for providing inadequate legal remedies for the protection of trade secrets. Title 3 of the 1991 Law for the Development and Protection of Industrial Property provides a weIl-elaborated legal system for the protection of such trade secrets. Although NAFrA includes a number of specific provisions on trade secrets (Article 1711), the new Mexican system appears already to include adequate protection in this area in so far as agriculture is concerned. Research Role of Civil Associations 5.54 One mechanism with which INIFAP augments its resources is the patronato, technically "civil associations". Patronatos are voluntary associations of farmers in specific regions, e.g., Sonora, that support research. They have alsq provided political support for agricultural research and have brought farmers' perspectives into research planning. 5.55 The laws regarding tax-deductibility of corporate contributions to non-profit entities were revised effective January 1, 1993 (II International Bureau of Fiscal Documentation "Mexico" 6,  81 1992). As described by a specialist in the organization of patronatos, contributions to these entities were not tax-deductible to the contributors before 1993, but SHCP has granted recently this deductibility. This change should contribute to increased financial support for this mechanism. 5.56 A related question is the taxation of Research and Development Expenditures. There is no limit on deductibility of actual research and development expenditures for a corporation engaged in such research, nor is there any requirement that such expenditures be amortized over the period during which the benefits of the research are received. Deductions are also allowed for payments into funds for technological investigation. These payments, however, are deductible only up to a defined proportion of a firm's gross income. As one reform in the funding of public research is to seek greater private participation through producers organizations, this limit ought to be changed. Technology Transfer 5.57 Early concerns about the fairness of international technology transfer agreements led Mexico to enact legislation in 1972, substantially amended in 1982, under which SECOFI reviewed and registered all international technology transfer contracts. The law forbade registration of contracts containing particular objectionable clauses, such as those prOhibiting export of products of the technology. those bringing in technology already available within the nation, or those establishing too high a price. The application of this law evolved over time to permit greater freedom to bring technology into the nation. New Transfer of Technology Regulations, issued January 9, 1990, significantly loosened the interpretation of the law and permitted technology to come in for a variety of economically desirable purposes. Transitory Article 2 of the 1991 Law on the Development and Protection of Industrial Property completely repealed the Technology Transfer Law and its regulations. Parties are therefore now free to enter any form of technology transfer agreement that does not infringe other bodies of law. 5.58 Mexican and foreign firms are now free to import technology. The payments, whether lump sum or in the form of royalties, are deductible to the firm importing the technology. They are however. susceptible to a withholding tax (reflecting a tax on the foreign supplier of technology and recipient of the funds) of 15 percent in the case of a technology transfer agreement or, 35 percent in the case of a patent royalty. These withholding rates are in addition to VAT paid by the recipient of the technology. Although there is no similar withholding on dividends paid to a foreign shareholder, the effective tax rate on such earnings is 39 percent. These numbers may be reduced under the terms of specific tax treaties. Such a treaty is expected next year with the United States and in the near future with Spain and Germany; treaties are already signed with France and Sweden. Research with Transgenic Organisms 5.59 Transgenic organisms - defined as organisms "whose hereditary DNA has been augmented by the addition of DNA from a source other than parental germplasm" - may become more important in the near future as ways of increasing productivity or importing quality of high-value crops, especially horticulturals. Scientific concern in Mexico and abroad about transgenic organisms has led to organization of a biosafety review panel within SARH. This panel was organized on an ad hoc basis when the private sector proposed tests of transgenic tomatoes in the winter of 1988-89. Private firms were concerned that the panel did not yet have precise rules or procedures so that the process lacked transparency. 5.60 Article 5 of the new Seed Law provides a statutory basis for dealing with transgenic organisms. It requires that SARH grant explicit permission for research with "transgenic materials of  82 high risk." The May 26, 1993 regulations implementing the seed law spell out this review more precisely and outline the factors that should be taken into in determining whether a transgenic material is of high risk. These factors are: (a) the production of substances that may be toxic to organisms beneficial to man or to other living organisms capable of substantially changing the biological network; (b) the effects of genes which, by escape into the environment and transmission into related plants, whether for forest, domestic, or commercial use, create a risk of producing new pathogens, increasing the harmfulness of those already present, or increasing their ability to dominate ecological niches, with the possibility of causing significant change in the distribution and diversity of the flora and fauna in such niches; and (c) the presence of genes which produce substances categorized as environmental pollutants. The outline of the application is presented and SARH then has a 45-day decision period. At this point, there is a review panel, and more detailed regulations are being developed to define the necessary standards and procedures. Pesticides 5.61 The main issues relating to pesticides are: (a) barriers to their sale hampering productivity by imposing additional costs on producers; and (b) risks to workers using pesticides and to consumers of treated products. The economic restrictions that once existed on imports are gone. 46 5.62 A new legal structure has strengthened the pesticide enforcement mechanism. The New Environmental Law (January 28, 1988) gave the Ministry of Social Development (SEDESOL) (then the Ministry of Urban Development and the Environment, SEDUE) a significant role in regulating pesticides. It supplemented regulation by SARH, the Ministry of Health and SECOFI. Under the new law, the various Ministries organized an Intersectoral Commission for the Control of Processing and Use of Pesticides, Fertilizers and Toxic Substances (CICOPLAFEST), and negotiated a "single window" arrangement for approving pesticides. There is now a standard application form that indicates the information needed for approval. A catalogue of approved pesticides is now published. For a pesticide to be imported, it must be on the approved list and its importer must be on a Jist of people regarded as capable of working safely with the material. Mexico has begun working with the new program for harmonized review of pesticides within OECD. It hopes that this harmonized system will replace the existing regulatory system sometime in 1994. It remains to be seen whether adequate resources will be allocated in the field to monitor health risks to workers. III. Extension 5.63 Extension and input supply are the chief means to transfer new methods to farmers. The efficiency issues in extension are analogous to those in research. Does the extension service have adequate resources to fulfill its mandate? Does it allocate resources appropriately to regions and production systems? Is it correctly balanced with the parastatals and private firms providing technical assistance? 5.64 Mexican extension has five main branches: SARH, the agricultural trust funds, the parastatals, producers organizations, and private firms (fable 5.3) in addition to input suppliers. In 1990, SARH had about 60 percent of public extensionists. (More recent figures on numbers of extensionists by public agency are not available, but they are smaller than in 1990). SARH staff work chiefly on basic crops and livestock. The chief agricultural trust funds are FIRA (the Bank of Mexico's Agricultural Trust Fund), a (recently closed) Trust Fund for Rainfed and Irrigated Areas (FICART), and FIRCO 46. Pesticides and other agricultural chemicals are now patentable (see para. 5.43).  83 (the Trust Fund for Shared Risk). FIRA provides technical assistance through its own agents, those of the commercial banks to which it re-lends, and those in a special program for low-income producers. FICART formerly had loan appraisers/extension workers with the same training as those of FIRA or of SARH giving technical assistance in crops, cattle production, agroindustry, fisheries, fruit production, other animals, and agroindustries. FIRCO contracts technical assistance through private sector firms and individuals. Extension by parastatals has disappeared with the elimination of the main parastatals, such as INMECAFE, IMPA, TABAMEX, CONADECA, and CONAFRUT. whose extension activities have been transferred to producers organizations. 47 In sum, after growing to reasonable dimensions in the 1970s, the apparent public extension system - SARH, the agricultural credit banks, and parastatals - has shrunk by about one third, with the latter disappearing almost completely. Table 5.3 Extension Agencies Institution \"'lIent Group Activities Research Links Cost-sharing SARH all fanners mainly crops and through demonstration 10-20% in irrigated livestock plots and INIFAP districts; very low publications elsewhere FIRA low-income and crops, livestock, agro- through FIRA recently begun with other farmers industries, agriculture demonstration fields some progress FIReO ejidos and other crops, livestock, agro- through INIFAP producers pay all low-income fanners industry, forestry extension after five years Input mainly commercial input use mainly through complete suppliers demonstrations INIF AP publications Commercial mainly commercial finance extension ? complete Banks Producers all types finance extension and ? high organizations fonn groups to receive Source: Mission interviews. 5.65 Mexican extension has been thoroughly reformed since about 1985. The main reforms have been to: (a) unify irrigation districts (DR) and rainfed districts (DT) as rural development districts (DDR); (b) delegate more budgetary responsibility to the DDRs; (c) create a generalist extension worker position as the main farmer contact, and to provide specialized technical advice from the districts; (d) give work a sharper socioeconomic focus; (e) improve linkages to research; (t) improve staff training; (g) raise cost-sharing; and (h) move extension away from its concentration on basic grains and livestock and toward more specialized enterprises that can provide higher returns.48 The various proposed reforms are summarized in Table 5.4. 47. See Chapter 9 on tropical agriculture for discussion of the role of the producers' organizations. 48. Some of these reforms were implemented under a pilot Bank project (PROCATI) that began in 1987 and closed on June 30, 1993.  84 Table 5.4. Extension Service Reforms Refonns in Aspect Before 1984 1984/85 Later Refonns Current Situation Organization DR"' and DT'" RDD'" RDD RDD . Budget authority Districts for DR, state RDD RDD RDD forDT Extensionists Technical assistance, Technical Technical Little change from pre- respons ibility planning, data collection assistance assistance 1984 Specialized support Absent Given by Given by districts Available in some to extensionists districts all districts but not Research linkages Thought weak To be Special program of Several special programs reinforced on-farm validation included Bank supported trials PROCATI and PRODERJTH Private extension Limited to irrigated areas Goal of 50 % cost- and cost-sharing and cash crops; some sharing in northern through credit system districts Staff incentives and Agreed to be inadequate Agreed to be Special incentive operating budget inadequate payments given Staff training Agreed to be inadequate Agreed to be Field and inadequate specialized courses Priority activities Too much on basic Nothing Nothing grains and livestock AI DR = Irrigation Districts hI DT = Rainfed Districts r;/ RDD = Rural Development Districts Source: Mission interviews. 5.66 The problems of the extension service remain much like those of research. Resources after the budget cuts of the mid-1980s are much sma11er than before and the cuts probably effect most the poorest farmers who are most dependent on public extension. In 1984, there were thought to be about 21,400 technica1 assistance staff in SARH, plus other public agencies and parastata1s. This fell to roughly 14,300 by 1990, with SARH and "other agency" staff numbers both dropping by about one third; though no more recent overa1l estimates are available, the numbers are known to be sma1ler. Cuts in SARH extension staff have not been restored by adding staff to other agencies and some extension staff may even be moved to ASERCA. 5.67 Two ways of relieving these resource scarcities without spending more are to change staff placement and to improve cost-sharing. An earlier ana1ysis of the congruence between numbers of extension staff against cropped area by state and va1ue of crop production by state was greater than 0.95; the indices for extension workers ranked as "good" with area and with va1ue were a1so greater than 0.95. Therefore, numbers and qualities of extension workers were a1located proportionately to planted area and to the tota1 va1ue of crop production. With respect to irrigated and rainfed areas, in 1985 the va1ue of rain fed crop output was about 52 percent of the national tota1, and rainfed districts had 74 percent of staff. This misa1location is justified by the fact that farmers in irrigated districts had better access to private extension, but it does show that there is little to be gained by shifting  85 public extension staff to the rainfed and tropical districts unless total numbers of staff are restored to those prevailing in 1984/85. 5.68 Previous World Bank reports have urged shifting costs to the private sector. CosHharing in SARH and the remaining parastatals is now minimal, with the only well-established source being the irrigation district water charges. Cost-sharing is complete in the private input suppliers, well advanced in FIRA and FIRCO. It now recovers nearly 40 percent of all costs of technical assistance from low-income producers, and has a goal of 100 percent. 5.69 Administration. Irrigation districts and rainfed districts combined to become rural development districts, but the irrigated districts have now been put back under the supervision of CNA. Hence, the previous review's criticisms stand. 5.70 Farmer participation and methodologies. The critique of extension was that it failed to use farmers' knowledge, via fanning systems research, and that it insisted on commodity-oriented work when a diversified approach could give better results. 5.71 The use of methods Iike those of the training and visit (f&V) system was well-established at the district level. T&V is not used, as such, by the other agencies, and in their work fanner participation is less robust. 5.72 Extension/research linkages. One critique of the research/extension linkage was that crops, livestock, and forestry research were in three subdirectorates of SARH. This critique has less validity now. INIFAP is the central agency for crops, livestock, and forestry, with explicit administrative links to extension. There are also formal mechanisms between SARH and other agencies to coordinate research/extension linkages (these include the technical committees of PROCATI and of PRODERITH, INIFAP input into the agricultural credit banks, collaborative field demonstrations by INIFAP and SARH, and INIFAP plus SARH collaboration with PRONASE). Other examples of effective linkages abound including INIFAP produces many extension publications for fanners and for technical staff. Linkages are strong in technical assistance from the input suppliers because they have a natural interest in providing advice about the use of the products they sell. Hence, at the national policy level, there is nothing more to be done to strengthen these linkages. IV. Recommendations A. Research 5.73 We confirm and extend the conclusion of the earlier review that greater public and private funding are needed. The Government has apparently decided not to fund public research but has not found an alternative strategy. There needs to be a long-term strategic plan, with a clear definition of the public role, the nature and size of the related institutions, and a serious plan for funding those institutions. 5.74 Whatever the total amount of resources made available for publicly funded research, their allocation to different commodities will change with the shifts in the quantity and value of output resulting from the trade and price policy reforms. The public research system, especially if it is continually short of basic resources, has to concentrate on fewer commodities and fewer problems. The major shift in priorities has to be to the poorer areas--in the center and south of the country. nearly always rainfed. often tropical. A general plan to shift priorities would be to grant seed money  86 to a few (new or existing) private institutions that will take over the privatized research facilities in the north. This money should be granted competitively, to national or foreign institutions, for five years, after which time the privatized institutions would be independent of core federal support, but could still compete for grants. The institutions would then rely on state governments, producers organizations, other donations, and their own commercial activities (e.g., consulting and farming). The same should be done for the research activities of the former parastatals in coffee, sugarcane, and cocoa. 5.75 A special area of reform of the allocative efficiency of the research/extension system is in water use efficiency. The rising shadow price of water to farmers, which will most probably result from the operation of water markets (see Chapter 4), will drastically alter the economics of irrigated farming. The Government again has to decide if private supply response, supported by private and imported research alone, will be adequate. B. InteJlectual Property 5.76 The seed law reform has been effective in that farmers have more choice among seeds and seed suppliers and INIFAP has an additional income source. The initial success of the reform is leading to a new set of problems, however. As the private sector becomes more sophisticated, it will look to INIFAP for material that can be further transformed in breeding; this will require a new agreement between INIF AP and the private sector. When Mexico adopts PVP protection (see para. 5.77 b) INIF AP will have to adjust its seed distribution policies. 5.77 Further reform is needed in several areas: (a) ensuring that newer varieties form a larger share of the Register of Recommended Varieties; (b) developing appropriate laws and institutions for PVP protection in compliance with UPOV; and (c) INIFAP should continue its efforts to prepare a policy governing its own use of PVP certificates. This policy should cover: how INIFAP should obtain PVP protection on its own materials; the rights to individual INIF AP breeders who have developed varieties on which a PVP certificate is obtained; the transfer of PVP rights held by INIFAP (Le., market exclusivity) to another entity (e.g., a private firm) and how it should choose among potential bidders for such rights; should INIFAP apply for PVP protection outside Mexico?; and INIF AP should face the question of transferring varieties, with or without PVP exclusivity, to entities that may further improve them. It may be able to do this by incorporating in its contracts appropriate provisions for obtaining a royalty on the derived varieties, taking into account the relative weight of the INIF AP variety in the parentage of the varieties that are ultimately marketed. 5.78 In order to ensure its farmers the benefits of advanced biotechnology, Mexico should change its intellectual property laws to permit protection of genetic engineering with natural and modified genes, using appropriate interpretations of its novelty requirement. This also requires adjustment of the provisions governing the use of descendants of protected biological materials. It is likely but not certain that these changes will be required by NAFfA. Prompt promulgation is required of regulations for testing of transgenic organisms will facilitate the use of biotechnology applications employing such organisms. 5.79 In making the technical decisions needed to develop its PVP law and to amend other intellectual property laws, Mexico would benefit from stronger efforts to tap a variety of expert opinions. There have already been studies and seminars on these intellectual property issues in Mexico--additional efforts to expand the knowledgeable community are desirable.  87 5.80 With respect to patent law, the new law in 1997 will provide some expanded patent protection. There will remain either explicit lack of protection, or significant ambiguities, in patent law as it affects foods, drinks, and techniques to produce them. It is important to resolve those ambiguities, by moving up the date of the new law and by being more explicit about its consequences for consumable products. c. Reforms in Extension 5.81 There is strong criticism with the Government of the size and role of the public extension service. This criticism has expressed itself as declining resource allocations to public extension, in what amounts to a de facto privatization. Moreover, the shift of much of the extension staff in the DDRs to acting as ASERCA agents reduces even more the true effectiveness of the service. Even before the cuts of the 1980s, the extension service was generally not intensive enough to achieve the kinds of results in rainfed and tropical areas which have been found with better coverage and it is now obviously even less able to contribute to renewed growth. Hence, the recommendation is to; (a) restore the cuts in public extension; and (b) reallocate much of the service to the center and south where smaller, poorer, rainfed farmers are most numerous. 5.82 Even with very limited resources for public extension, the target area has to be the south and the center because of the weakness of privately-funded extension there. More innovation is needed with publicly-financed, but privately-executed extension, services. An example could be the producers organizations for export crops which could manage all research and extension for coffee, cocoa, other tree crops, and sugar cane. The policy of unassisted privatization is failing, particularly for commodities facing very low world prices (e.g., coffee) or those where the eventual financiers of research and extension services have heavy debt (notably sugar cane). While Government assistance should not be given for an indefinite period, perhaps no more than five years, it is warranted in view of the crisis of technology transfer services in exportable crops and the time needed to create anew, private, effective structure.  88 6. RURAL FlNANCE411 I. Background 6.1 From 1982 to 1988, the Mexican financial sector was extremely restricted. In September 1982, the commercial banks were nationalized and the Constitution was amended to reserve banking to the state. At that time, many banks became insolvent as a result of a large foreign exchange exposure, a deterioration in the quality of their portfolios in an inflationary environment and interest rate controls that had reduced their margins. The Government merged and closed 60 nationalized banks forming a new structure of 20 banks, and in 1986 formed a fund to help rehabilitate and modernize the banks. There were eight development banks and 21 government financial trust funds. The development banks rivaled the commercial banks in asset size. 6.2 Since 1988, the financial system has undergone a remarkable transformation. Mexico shifted from a highly controlled financial system to one based largely on market forces. At first, commercial banks were permitted to operate an uncontrolled intermediation system in parallel with the controlled system in October 1988. Then, limits on commercial bank Jending to the private sector were lifted in January 1989. Later, all interest rate controls, forced lending schemes, reserves requirements and liquidity coefficients on marginal deposits were eliminated, while a remarkabJe reform and modernization of the supervisory and regulatory framework for banks and brokers was implemented. This more liberal and stable financial framework set the stage for the privatization of commercial banks, which was successfully concluded in July 1992. 6.3 With the reforms, serious efforts to improve the development banks began. The Government agreed to examine the role and functioning of development banks and trust funds under the Financial Sector Adjustment Loan (FSAL, Loan 308S-ME, June 1989), and began merging and liquidating trust funds with overlapping functions. At the end of 1989, the Government of Mexico (GOM) implemented a Program for Structural Change and Financial Modernization of Development Banks which redefined the roles of development banks and trust funds. They were reduced in size and number as inefficient institutions were reorganized or closed, their assets were reduced by 46 percent in real terms over the 1987-90 period, and the total staff was cut by 25 percent to be 34,000 in 1990. 6.4 Rural Finance. An efficient rural financial system is necessary to direct investments toward more efficient uses and to mobilize savings. For many years, the GOM has undermined the development of such a system by directing funds to favored producer groups through targeted credit. While the volume of targeted credit has fallen recently, and some other reforms have started, the Government still has not fully liberalized rural financial markets. The overall degree of reform in those markets is small in comparison with other policy changes in the agriculture sector. Rural finance has been a major object of Bank lending for Mexican agriculture. "-l The Bank's ample lending experience has, accordingly, been used for much of what follows. 49. Unless otherwise indicated, aU figures are real 1992 new pesos or real 1992 dollars. Fiscal costs refer to budgeted transfers; quasi-fiscal costs refer to indirect transfers that have an effect on the Treasury. SO. Much of it is reviewed in the Project Perfonnance Audit Report (Report No. 8860) of Mexico's Fifth through Eighth Agricultural Credit Loans, of the Operations Evaluation Department, and the Ninth Agricultural Credit (Loan 2837- ME) Project Completion Report.  89 6.5 The GOM's intervention in rural credit has been through its key rural finance institutions, namely, the Banco Nacional de Cr&lito Rural (BANRURAL - National Rural Credit Bank) and the Fideicomisos Instituidos en Relaci6n con la Agricultura (FlRA - Trust Funds for Agriculture.) Until recently, the Fideicomiso para Cr&Jito en Areas de Riego y de Temporal (FICART - Trust Fund for Credit in Irrigated and Rainfed Areas) played an important role in channelling external funds to BANRURAL. The Aseguradora Nacional AgrCcola (ANAGSA - National Agricultural Insurance Company, which was closed in 1990) insured most of BANRURAL's loans. 6.6 The problems facing Mexico's rural financial system have been numerous. Directed credit at subsidized and preferential interest rates, subsidized credit guarantees and agricultural insurance premia, and debt restructuring and forgiveness programs have characterized rural credit in Mexico. The massive presence of government institutions have crowded out private financial intermediaries, discouraging the development of private sector institutions that could provide alternative, efficient services. 6.7 Government policy discouraged commercial bank lending to agriculture as well, particularly to small farmers. Borrowers have paid below market, and sometimes negative, interest rates on their loans. Compulsory insurance of BANRURAL's loans by ANAGSA induced borrowers to default on their loans and encouraged BANRURAL loose recovery practices. As a result, many loans went uncollected, while ANAGSA and BANRURAL became large recipients of government funds. 6.8 During the past four years, however, the GOM has made substantial reforms in rural credit policies, all conducive to reduce transfers to the sector and improve the efficiency of its rural finance institutions. It reduced interest rates subsidies, made transfers to development banks and trust funds more transparent, and closed or reorganized inefficient government institutions. As a result of these reforms, government transfers to rural finance institutions have declined from US$2-3 billion a year to less than US$800 million in 1992. However, on-lending interest rates remain controlled and credit guarantee and agricultural insurance policies subsidized, requiring still considerable government support'to rural finance institutions, while the rural population remains largely outside the formal financial system. 6.9 This chapter presents an overview of agricultural financing in Mexico and the estimated fiscal and quasi-fiscal cost to the GOM of supporting the rural financial system during 1983-92, evaluates to which extent recent government poliCies have been successful in reducing the fiscal burden of the rural credit system and improving the efficiency of public development finance institutions, and makes recommendations for further reforms to improve the overall efficiency of rural financial markets. II. Agricultural Financing 6.10 The economic crisis of the early 1980s and subsequent budget cuts led to a substantial decrease in agricultural credit as a proportion of total credit to the private sector. Although total credit to the private sector grew substantially during the decade, agriculture received smaller shares of financing. Agricultural credit represented 22 percent of all credit granted in 1983; by 1992, it represented just 8 percent. (See Annex 2, Table 1.) When compared to agricultural GOP, net credit 51 flOWS to agriculture, which had reached a high of 17 percent in 1988, have to 8 percent by end of 51. Agricultural credit disbursements is a better measure of annual fmancing to sector. However, this data is not available for commercial banks. Therefore, to be consistent. net credit flows are shown for all institutions. They represent the annual increase/(decrease) in portfolio balances. Changes in BANRURAV. portfolio were negative in 1990 and 1991 due  90 Figure 6.1 1991-the level it maintained in 1992. (See Annex 2, Table 2.) Nevertheless, agricultural credit REAL INCREASE IN OursTANDING CREDIT has shown positive but TO TIlE PRlVATE SECTOR decreasing real growth rates 201.M , - - - - - - - - - - - - - - - - - - - - - , from 1988 through 1992. (See Figure 6.2.) The share of commercial banks in total rural lSO'Ho finance has risen lately, though this may simply represent a IO(M recovery to the share held before the nationalization of the banks in 1982. 6.11 The GOM's key rural fmance institutions contributed -SO'Ho'-1-_+-~-4_-+_-~~-_+--r-~ to the largest share of 19E3 1984 1985 1986 1987 1988 1989 1990 1991 1992 agricultural credit during 1983- 92. (See Figure 6.3.) This is not evident in 1990 and 1991, when the share of commercial banks to total agriculture portfolio balances represented 56 percent and 53 percent. However, it should not be interpreted as entirely due to increased commercial bank financing to the sector. The transfer of a substantial part of BANRURAL's overdue portfolio to the Fideicomiso para la Reestructuraci6n de la Cartera Vencida (FIRCAVEN - Trust Fund to Restructure the Overdue Portfolio of BANRURAL) and the Programa Nacional de Solidaridad (pRONASOL - National Solidarity Program) altered the composition of outstanding Figure 6.2 agricultural credit. AGRICULTURALnNANCING O'OTSTANDINGA.GRICULTURAL PORTFOUO 6.12 During 1983-88, u~~--------------------~ BANRURAL's contribution to agricultural financing was significant, providing an average of 45 percent of all agricultural credit, which fell in 1989-92 to an average of 8 percent. Despite the subsequent massive reorganization and cleansing of its portfolio, reduced fiscal transfers and still low recoveries rates led to BANRURAL's reduced participation in agricultural financing. As a percentage of • BANRURAL ~ FIR.A ~ COMM. BANKS agricultural GDP, significant portfolio write-offs; consequently, figutel shown arc actual net flows (new loans granted leu recoveries, as opposed to annual changes in portfolio balances).  91 BANRURAL's net annual lending oscillated between 4 percent and 7 percent during 1983-88, but decreased to 1.0 percent in the following years. 6.13 FlRA's presence in agricultural financing has been also significant. It discounted an average of 35 percent of outstanding agricultural credit during 1983-92. As a percentage of agricultural GOP, FIRA's net discounts oscillated between 3 percent and 5 percent over the same period. Additional commercial bank credit to the sector averaged 35 percent. m. The Cost of Rural Finance (1983-92) 6.14 Figure 6.4 presents the flow of government funds to Mexico's rural finance institutions. Through the annual budget process, SHCP assigns fiscal funds to FIRA, FlCART, BANRURAL and ANAGSA/AGROASEMEX (the GOM's agricultural insurance companies). BANXICO refinances part of FlRA's discounts. A small part of FlRA's funds supports discounts to BANRURAL, and until 1993, most of FICART's funds partly financed BANRURAL's loans. Another share of FIRA's funds reimburses commercial banks for their transaction costs and the provision of technical assistance to borrowers; still another part defrays the salaries of FlRA staff who also provide technical assistance services to farmers. Part of Government transfers to ANAGSA and AGROASEMEX have returned to FIRA and BANRURAL in the form of indemnity payments. 6.15 The estimated value of Government transfers to agricultural finance institutions exceeded NS95 billion (USS23 billion) during 1983-92. During this period, the overall cost to the treasury of maintaining the FIRA and FICARTIBANRURAL credit systems alone (including ANAGSA and AGROASEMEX) has been estimated to exceed NS94.4 billion (US$22.6 billion).s2 Transfers to other development banks and trust funds during 1983-92 were only NS724 million (USS176 million). (See Annex 2, Tables 3 through 5.) Thus, the focus of this Chapter will be on the major institutions. 6.16 Government transfers have declined markedly over time. The cost of the agricultural credit and insurance system (measured as the real, weighted average cost to the GOM of each dollar lent to agriculture) dropped from 83 cents per dollar lent in 1983, to 17 cents in 1992. (See Figure 6.5 and Figure 6.6.) 6.17 Table 6.1 summarizes the total cost to the GOM of supporting the agricultural credit system. Subsidized and preferential agricultural lending interest rates were responsible for most transfers, except the costs of technical assistance, credit guarantees and agricultural insurance premia. The costs of subsidized and preferential interest rates can be interpreted as the savings to the Government if it had not controlled interest rates. It should be noted, however, that a large part of total transfers during 1990, which substantially increased fiscal transfers during 1989-92, should be attributed to President de la Madrid's decision to assume the external debt of these institutions (Convenio de Posicion Cona de Divisas, 1983.) During 1989-92, the GOM paid for most of BANRURAL's foreign debt. Excluding debt retirement from the GOM's cost of BANRURAL, total transfers would decrease by N$8 billion, and the share of transfers for increased lending and operating losses would decrease to NS3.8 billion. 52. World Bank estimates based on the audited financial statements and/or data provided directly by the institutions involved (see Annex 2).  FLOW OF FUNDS SIlCP r ..... I_B~NCO DE MEXICO 1 ~ ------.... ANAGSA ) "--- - - -- --- ---------.....\ FICART """"'- _ _ _ _ _ .....J. _ _ _ _ _ --"'" I i"Jj ...... [AGROASEMEX o ~ •. '!'l (I'Q o C H'I ri ~ C\ \0 ::I L.. N l. COMMERCIAL BANKS BAN RURAL p,. en ---- I .. [ BORROWERS .. 18% PBI!! 82% OPs BORROWERS AIlPBIs ....... r. Discounts/rediscounts Fiscal Transfers -~ Loans ..... Insurance payments .......  93 Table 6.1 Cost to Government of Rural Financial System (N$ 1992 mill ions) Item 1983-1988 1989-1 Technical Assistance 419 228 647 Credit Guarantees 152 580 733 A ricultural Insurance Premia Subsid 11549 5 185 16644 Costs of subsidized and preferential interest rates: - Operating costs of BANRURAL, FICART, and FIRA 14,093 9,747 23,840 40623 11 864 52487 66,836 27,604 94,440 A. The Cost of Government Rural Finance Institutions 6.18 Government transfers to rural finance institutions responded mostly to subsidized interest rates, especially at times of high inflation, on loans to agricultural producers, both low-income (PBIs) and other producers (OPs) ..53 The GOM managed interest rate structure has not reflected the costs and risks of lending. Instead, interest rates have differed according to productive activity and class of borrower. As a result, government finance institutions have required the GOM continuous support to cover their operating costs and losses, and to sustain and increase the real level of their lending operations. Transfers were also due to other partly or Flgure 6.4 wholly subsidized producer programs (for example, GOVER.NMEl'IT SUPPORT TO THE agricultural insurance, technical RURAL FINANCIAL SYSTEM assistance and credit 14~ I------------;--~---___, guarantees) . 12~ 6.19 BANRURAL was the .. IO~ largest rural lender until 1990. co II It was created in 1975 by ~ merging three public sector t B~ Z agricultural credit institutions, 'S a 6~ ~ 4~ with the objective of providing credit to low-income agriculture producers unable to provide col1ateral. BANRURAL comprises 12 regional banks and one national bank with o '~~~-_+-_+-~-~-~-~-+_-~ 19113 1984 l!l8ll 1PB6 IlIB7 1_ 1!189 IlI!lO 1!1!11 1m combined net worth of about US$1 billion at end-1992. 53. PBls are producers (mostly ejidatarios) with net annual family income not exceeding 1,000 times the minimum daily regional wage and comprise about 80 percent of the rural population (23 million people). OPs are all other producers.  94 6.20 BANRURAL had no incentive to properly analyze credit risk or monitor borrowers, because of the security of government transfers and indirect loan recoveries from ANAGSA. Government transfers were necessary to keep BANRURAL in operation, especially due to large operating losses resulting from low rates of loan recovery, high operating costs, and very low and sometimes highly negative interest rates. BANRURAL's loans were compulsorily insured against default with the wholly government-owned insurer, ANAGSA. BANRURAL recovered most of its loans from ANAGSA's indemnity payments. 6.21 BANRURAL's operating costs have been significantly high. During 1983-92, BANRURAL's operating costs amounted to N$22 billion (US$5.6 billion). (See Annex 2, Table 8), Before its reorganization in 1989, BANRURAL's operating costs averaged around 40 cents per peso lent. Despite having halved its staff, this ratio increased to 53 cents per peso lent mainly because reduced fiscal transfers caused BANRURAL's lending program to shrink. In contrast, commercial banks' operating costs averaged 8-10 cents per peso lent to agriculture. 6.22 Mobilization of rural savings has also been slow. BANRURAL's need for deposits to lend to the rural sector was discouraged by the security of receiving government transfers to meet its lending targets. 6.23 Transfers to BANRURAL in the form of capital contributions to equity, assumption of foreign debt, and subsidies to cover operating losses amounted to approximately N$41 billion (US$1O billion), or about 45 percent of the real dollar value of all government transfers to rural finance institutions during 1983-92. (See Annex 2, Tables 3 and 7.) 6.24 Measured as the cost of each real dollar lent to agriculture during the period, the costs to GOM of lending through BANRURALIFICART (including agricultural insurance) oscillated between 22 cents (1987) to US$3 (1990) per dollar lent by BANRURAL. As mentioned earlier, a large part of total transfers during 1990 should be attributed to the GOM's decision to pay offBANRURAL's external debt. The GOM's cost dropped to 35 cents per dollar lent by BANRURAL in 1992 as budget transfers were Significantly reduced. (See Annex 2, Table 6.) 6.25 FICART was created in 1976, with BANRURAL as trustee, to insulate the Mexican agriculture credit operations of multilateral lenders, in particular the Inter-American Development Bank (lDB), from BANRURAL's bad loan portfolio. In 1987, the World Bank also started to channel part of its agricultural credit to BANRURAL's PBIs through FICART. Disso]ved at the end of 1992, FICART's role was to lend to BANRURAL and provide technical assistance to farmers. 6.26 Government transfers to FlCART in the form of capital contributions to equity (mostly in grants from IDB loans), and some smaller amounts to support FICART's operating costs, amounted to approximately N$3.5 billion (US$S46 million), or about 4 percent of the real dollar value of all transfers to rural finance institutions during 1983-92. (See Annex 2, Tables 3 and 9.) 6.27 During 1983-92, FlCART received capital contributions of N$2.6 billion (US$60S billion) in external loans disbursements. Neither the financial cost to the GOM of external loans transferred to FlCART nor outstanding debt balances are available. As these loans were not FlCART's liabiJities, the financial cost to the GOM of these funds cannot be assessed. 6.28 Not all of the GOM's transfers to FlCART were to discount BANRURAL's loans. From 1983-88 a smaller sum, N$56 million (US$13 million), covered FICART's production technical  95 assistance to BANRURAL's borrowers. Another part covered FICART"s operating costs and losses, NS475 million (USS120 million). (See Annex 2, Table 10.) 6.29 As the overall cost of running FICART as a separate institution to discount only BANRURAL's loans turned increasingly unjustifiable, the GOM closed FICART and transferred its assets to BANRURAL. These assets were mainly BANRURAL's liabilities with FICART and this transfer represented only an accounting entry in BANRURAL's books. 6.30 ANAGSAIAGROASEMEX. The GOM support for agricultural insurance amounted to N$16 billion (US$3.8 billion), or about 17 percent of the real dollar value of all transfers to rural finance institutions during 1983-92. (See Annex 2, Tables 3 and 11.) 6.31 During 1983-89, the GOM total transfers to ANAGSA amounted to an estimated at NS12 billion (USS2.8 billion). Until the creation of AGROASEMEX in 1990, there was not an explicit Government subsidy on insurance premium. Farmers' premium amounted to NS6,4 billion (USS1.3 billion), while indemnities were about NS16.7 billion (US$4 billion). This implies that losses covered by the GOM were about N$10.3 billion (USS2.7 billion), or an average NS1,288 million (USS337 million) of premium subsidy per year. Moreover, with low premia income, ANAGSA could not cover its operating costs. The estimated cash shortfall between fiscal transfers plus farmers premia and administrative costs plus indemnities suggests that ANAGSA required an additional NS1.3 million (US$262 million) in the GOM support to help defray its operating costs, which totalled NS2,4 billion (USS558 million) during 1983-89. 6.32 ANAGSA's operations were subject to adverse selection and moral hazard. The company insured the riskier of the Mexican farmers. During its last three years of operations (1987-89), indemnity claims represented 75 percent of the area insured. Moreover, claims on fictitious losses were widespread. After providing an additional NS2 billion (USS 538 million), GOM closed ANAGSA in February 1990. 6.33 In Iune 1990, the GOM Figure 6.5 set up a new agricultural ,------------------------, insurance institution, GOM COST PER DOLLAR LENT AGROASEMEX, as a 117 flU .... BANRURAL $3.5 subsidiary of the Mexican public insurance company ASEMEX. $3.0 From 1990 through 1992. total fiscal transfers to AGROASEMEX amounted to NS400 million (USS1l2 million), implying that the GOM has provided an average annual premium subsidy of NSl60 $1.0 million (USS45 million) to farmers. Apparently, 10.5 AGROASEMEX has managed 10.0 to identify a commercially viable portfolio, but it continues to rely on Government transfers "0- IOTAlCOST to cover subsidized premia. The GOM had agreed to subsidize AGROASEMEX's premia, on a declining basis, until the 1991  96 spring/summer crop cycle. By then, the GOM expected that the company would be self-financing and able to compete with private insurance companies. By 1991, however, AGROASEMEX was not yet able to break even. In December 1991, the GOM reversed its previous decision and announced an increase in premium subsidies from 20 percent to 30 percent. These premium subsidies helped defray AGROASEMEX operating costs, which totalled NS176 million during 1990-92 (USS52 million). 6.34 In February 1992, the GOM informed the Bank that it was considering extending the 30 percent insurance subsidy to all private companies issuing agricultural insurance policies. The GOM's objective was to increase competition between private insurance companies and AGROASEMEX. AGROASEMEX would be in charge of channelling the subsidy, but contingent to private companies purchasing its reinsurance policies. The cost of this program nor the income AGROASEMEX has received through reinsurance fees is not available. 6.35 The value of government transfers to AGROASEMEX appear low compared to those received by ANAGSA. However, the cost efficiency of AGROASEMEX would be better assessed based on government cost per hectare insured. Transfers to ANAGSA between 1983-89, for example, averaged NS1,750 (USS408) per hectare insured. Comparative data for AGROASEMEX is not available. 6.36 BANRURAUFICART/ANAGSAIAGROASEMEX. Overall, during 1983-92, the system cost the GOM about NS61 billion (USS15 billion) or 65 percent of the real dollar value of all transfers to rural fmance institutions. Measured as the fiscal cost per dollar lent by BANRURAL, it cost the GOM 86 cents (of each dollar transferred to the three institutions) to lend one dollar through BANRURAL in 1983. In 1992, this cost had dropped to 37 cents per dollar lent. (See Figure 6.5, Figure 6.6, and Annex 2, Table 6.) 6.37 FIRA is a group of four trust funds under BANXICO. The Fondo de Garantfa y Fomento para la Agricultura, GanaderCa y Avicultura (FONDO - Credit and Guarantee Trust Fund for Crop, Livestock, and Poultry), established in 1955, refinances short-term commercial bank loans to agriculture, livestock and agroindustry. The Fondo Especial para Financiamientos Agropecuarios (FEFA - Special Trust Fund for Agricultural Financing), established in 1965, refinances short-term commercial bank loans. The Fondo Especial de Asistencia Tocnica y Garantfa para Cr~itos Agropecuarios (FEGA - Agriculture Technical Assistance and Loan Guarantee Trust Fund), established in 1973, reimburses commercial banks the cost of loan evaluation and technical assistance for PBIs, and provides partial credit guarantees for their loans. The Fondo de Garantfa y Fomento para las Actividades Pesqueras (FOPESCA - Fisheries Credit and Guarantee Trust Fund), established in 1989, refinances commercial bank loans for fisheries activities and guarantees loan recoveries. 6.38 The total estimated value of the GOM's financial support to FIRA during 1983-92 was NS33 billion (USS5.2 billion) or 34 percent of the real value of all transfers to the rural fmancial system. Measured as the cost to the GOM of each dollar discounted by FIRA, the cost to GOM dropped from 78 cents in 1983 to 9 cents in 1992. (See Annex 2, Table 3). 6.39 Government fiscal transfers to FIRA, to compensate for low-interest discounts and to operate its technical assistance and credit guarantee programs, amounted to approximately NSI0.3 billion (USS2.5 billion) that is, 10 percent of the GOM's total transfers to rural fmance institutions. (See Annex 2, Tables 3 and 12.) In addition to these regular budget transfers, FIRA borrows from BANXICO and other government agencies at rates well below the GOM's cost of funds. The actual cost to the GOM these funds is not available. However, these quasi-fiscal transfers could be estimated as the difference between FIRA's actual annual income expense and income expense  97 estimated as the difference between FIRA's actual annual income expense and income expense restated at the higher of the GOM's marginal cost of funds (the CETES rate) or the rate of inflation. The estimated value of these transfers to FIRA represented an additional cost of N$23 billion (US$5 billion) to the GOM during 1983-92. (See Annex 2, Tables 3 and 13.) This implies that the quasi- fiscal cost to the GOM of FIRA was in effect twice as high through concessionary lending than through direct budget transfers (US$2.5 billion). 6.40 The relative importance ofBANXICO's rediscount lines in funding FIRA's discount programs has increased markedly since 1989, as external funds from lOB and IBRD loans were depleted. During the past few years, it has been the main source of quasi-fiscal transfers to FIRA, as it refinances FIRA's discounts at a interest rate equal to the weighted average interest yield of FlRA's portfolio less 3.5 percentage points. 6.41 Discount Facility. To induce commercial bank lending to the agricultural sector, the GOM operates FIRA as a discount window for first-tier lenders to discount their working capital and investment credit to PBIs and OPs. FlRA discounts different proportions of these credits depending on producers' income levels and type of production, allowing the banks a 5 or 6-percentage point discount margin. 6.42 FIRA's discount rates and percentages were changed frequently in the past as the GOM sought to provide greater incentives to commercial bank lending for agriculture. At present, FIRA's discount structure provides more attractive terms for credit to PBls, and to OPs engaged in the production of basic products, than for producers in other areas.Sot For example, commercial banks may discount with FIRA 90 percent of their working capital or investment loans to PBls (with a 6-points margin) and 80 percent of their loans OPs of basic products (with a 5-point margin), on the condition that these loans are granted at FIRA's preferential interest rates. 6.43 Banks are required to charge FIRA's rate on the portion of loans financed with their own funds, although they are allowed to set their own rates on short-term loans to OPs if they cofinance a higher proportion of the loan than the required 20 percent. 6.44 Interest rates are still differentiated according to the term of the loan, the type of borrower and the type of activity financed. The large number of different categories adds to the complexity of lending operations and hence to the lenders' administrative costs, because lenders must ascertain the appropriate category for each loan, secure documentation, and carry out inspections that can demonstrate that each loan was placed in the proper category. Moreover, the differences in interest rates among categories represent attempts to subsidize certain types of activities and certain classes of borrowers rather than reflecting, as they would in competitive markets, differences among categories in the costs and risks of lending. Consequently, incentives are created for borrowers to seek loans from categories carrying the lowest interest rates rather than reporting accurately their own characteristics and the planned or actual uses of loan proceeds. At the same time, lenders have incentives to focus on those categories of loans where the spreads are highest relative to the costs and risks involved in making such loans. 54. Basic products for the purpose of FIRA's credit include: (a) Agriculture: maize, wheat, beans, rice, sorghum, soybean, cartamo, sesame, sunflower, and copra; (b) Livestock: milk, eggs, meat, chicken, and other poultry; (c) Agroindustries: processed products derived from the above agriculture and livestock products and/or inputs required for their production.  98 6.45 Between 1983 and 1992, the real value of FIRA rediscounts grew about 33 percent on working capital loans, and more than doubled in investment loans. Thus, while in 1983 FIRA discounted N$4.2 billion in short-term credit and N$2.1 billion in long-term credit, it discounted almost N$5.7 billion and N$4.7 billion, respectively, by the end of 1992, a real increase in total lending of about 64 percent. 6.46 Despite FIRA offering greater incentives for commercial bank lending to PBIs than OPs (see para. 6.70), credit to PBIs is declining relative to credit to OPS. From 1983 to 1988, FIRA's discounts to PBIs had increased from 37 percent of total discounts to 50 percent. However, on 1990 and 1991, discounts to PBIs declined sharply, to 27 percent and 18 percent of total annual lending, respectively. In 1992, FlRA changed its categories of producers from PBI and OP to Productores en Desarrollo I and II. Producers in category I are those with net annual family income below 3,000 times the minimum daily wage; producers in category II are all other producers. Even by broadening the range of lower income producers served, in 1992 producers in category I received only 37 percent of all FlRA discounts. 6.47 The cost to the GOM associated solely with FlRA discount operations during 1983-92 has been estimated at N$31 billion {US$7 billion)." Not al1 of the monies that FIRA received, however, covered the costs of cheap credit (low interest rates) and incentives for commercial bank lending to rural borrowers (intermediation margins). Transfers also helped defray the costs of FEGA's technical assistance services to farmers, and cover the deficits of FEGA's credit guarantee program. 6.48 Technical Assistance Program. Through FEGA, FIRA also sponsors two technical assistance programs. These programs have served to induce some commercial banks lending to PBIs, and to encourage small farmers to adopt advanced agricultural technologies. One of the programs reimburses, on a decreasing basis, individuals or groups of producers for the production technical assistance they contract. 6.49 The other program--recently discontinued--reimbursed commercial banks for part of the direct costs of making a loan, and partly for the costs of providing some productive technical assistance. This reimbursement was directly linked to the amount of the loan and represented about 3 percent of the banks' annual discounts. Although this program was labeled technical assistance, a significant share of this incentive to commercial banks represented actually a transaction cost reimbursement, and was not strictly related to production technical assistance. 6.50 FIRA later designed and introduced a new transaction cost reimbursement system (see para. 6.70), and a new reimbursement scheme for technical assistance under the Asistencia T~cnica­ Operativa Integral (ATOI-Integrated Technical-Operating Assistance) program for loan amounts larger than those covered under the Sistema de Estimulos a la Banca y a los Productores (SIEBAN - Incentive System for Banks and Producers), while it continues to partially reimburse, to farmers and commercial banks, for technical assistance to production under its Sistema de Asistencia T~cnica Integral (SATI - Integrated Technical Assistance System). 6.51 A breakdown of costs for the above programs is not readily available. Partial data shows that reimbursements to commercial banks for the costs of making loans are responsible for most technical 55. Estimated as the total value of fiscal and quasi-fiscal transfers to FTRA less the fiscal costs of technical assistance, guarantee deficits, and operating costs.  99 assistance expenses. During 1983-92, the total cost of FEGA's technical assistance programs amounted to NS234 million (USS149 million; see Annex 2, Table 14.) 6.52 Complementary Guarantees Program. FEGA extends guarantees to commercial banks on loans to PBIs to cover practically all risks (for example, climatic, price variability, borrower insolvency.) Since 1992, FEGA has extended these services to all Productores en Desarrollo 1. 6.53 During 1983-92, the GOM transferred N$733 million (USS200 million) to FEGA to offset its guarantee payments deficits. (See Annex 2. Table 14.) Until July 1988. FEGA provided credit guarantees free of charge. Since then. it charges 2 percent and 3 percent for its 60 percent and 80 percent guarantees (a premium of 1.2 percent to 2.4 percent of the annual outstanding credit balance.) In constant 1992 pesos, net guarantees paid (guarantees paid minus guarantees later recovered) by FEGA increased from N$38 million in 1986 (1.9 percent of the amount guaranteed) to N$130 million (15 percent) in 1991. Premium income as a percentage of net guarantee payments increased from 5 percent in 1988 to 40 percent in 1991. With premium income falling short to cover FEGA's losses. shortfalls have been covered with Government transfers. 6.54 Several reasons can explain the rapid increase in FEGA's losses. First. the decline in inflation produced a sharp increase in real interest rates (from highly negative to slightly negative or positive). albeit still subsidized, which led to cash flow and profitability problems. Second, the opening of the economy and the consequent reduction in the domestic price of important import competing crops like sorghum and soybeans contributed to farmers' cash flow problems. Third, the dissolution of the GOM's agricultural insurance company (ANAGSA) eliminated a source of finance for those producers who covered part of their repayments to commercial banks with the indemnities paid by ANAGSA. But above all, FEGA's main problem is that premia income is insufficient to cover payments of guarantees., Premia are equal for all farmers. independent of risk, and are issued against all risks, whether they are conventionally considered insurable or noninsurable. 6.55 Operating Costs. The bulk of FIRA's operating costs may be attributed to technical assistance services to farmers. FIRA's operating costs during 1983-92 totalled N$I.4 billion (US$360 million). FIRA does not report its operating costs by trust fund. Considering the large number of FIRA's technical staff in the field. it appears that most operating costs may be attributed to technical assistance services to farmers (provided by FlRA's field technicians), while the smallest share should be attributed to its discount operations. (See Annex 2, Table 14.) B. Estimated Loss in Value of the GOM Transfers 6.56 During the early and mid-1980s, FIRA's and BANRURAL's capital eroded seriously. The GOM chose not to force them to reduce their real level of operations. Instead, with the help of several external loans contributed to FIRAIFICART as capital, and of other foreign borrowing, rediscounts from BANXICO and budgetary transfers, the GOM maintained. and sometimes increased, the real level of FIRA and BANRURAL operations. The real value of the GOM transfers during this period also eroded sharply. An estimate of the loss in the real value of government fiscal transfers to BANRURAL and FlRA (quasi-fiscal costs are excluded) suggests that N$52 billion (US$12.6 billion) were passed on to borrowers in cheap credit and to defaulting borrowers as loan write-offs. (See Annex 2, Tables 15 and 16.) This amount equals total fiscal transfers to FlRA and BANRURAL during 1983-92.  100 c. Benefits Accrued to Borrowers of Government Funds 6.57 The following formula was used to estimate subsidies accrued to final borrowers as a result of negative and subsidized or preferential interest rates: S = P * (i + m)] - I where: S= Subsidy per period P= Average outstanding portfolio for the period i = Average cost of funds of commercial banks (CPP), except for periods with negative interest rates where the rate of inflation was used m= Financial intermediation margin (5.5 percentage points for FIRA (the estimated average intermediation margin FIRA has given to commercial banks; and 6 percentage points for BANRURAL. the intermediation margin FIRA gives BANRURAL on its loans to PBIs) (i+m) = Minimum market interest rate, assumes commercial banks and BANRURAL would have charged at least 5 percentage points for OP loans and 6 for PBI loans over the weighted average cost of funds of commercial banks (CPP) I = Interest income on outstanding portfolio 6.58 Under the above assumptions, the benefits of FIRA and BANRURAL borrowers totalled N$62 billion (US$14 billion) during the ten-year period 1983-92. (See Annex 2, Table 17). Most important, however, is that 81 percent of these benefits accrued in just five years, 1983-87, when on- lending interest rates to agriculture were highly negative. As inflation subsided, and the GOM raised and made adjustable on-lending interest rates to agriculture, borrowers' benefits decreased accordingly. 6.59 Most benefits accrued to BANRURAL's borrowers (62 percent) because the bulk of BANRURAL's clientele are PBIs who have benefitted from the most attractive financing terms. FIRA's borrowers did not benefitted as much because PBIs comprised only half of its clients until 1988; PBI participation dropped markedly later to reach a meager 18 percent in 1991 (see paras 6.97 to 6.105), 6.60 Total benefits accrued to borrowers were actually higher: benefits need to be adjusted upward by adding write-offs of overdue loans and debts settled through insurance and credit guarantee payments. The Bank does not have any information on write-offs of the overdue portfolio of BANRURAL transferred to the FIRCAVEN and PRONASOL in 1991, and to the Fideicomiso Liquidador de Instituciones y Organizaciones Auxiliares (FIDELlQ - Trust Fund for the Liquidation of Auxiliary Institutions and Organizations) in 1992 to make this adjustment. However, as recoveries rates for FIR CAVEN were reportedly insignificant. total benefits to borrowers of BANRURAL must have been substantially higher when write-offs of bad loans are included.  101 6.61 Debt forgiveness programs have made BANRURAL's problems worse and it continues having severe loan recovery problems. In December 1992, a report published by the Secretarfa de la Contralorfa General de la Federaci6n (SECOGEF - Secretariat of the Comptroller General of the Federation) indicated that during the first semester of 1992, the overdue portfolio of BANRURAL was about N$867 million (US$289 million), almost 3 times the amount of overdues in December 1991. 6.62 Canceled debt through insurance and credit guarantee payments have been also important. After adding the GOM subsidized premia and net guarantee payments, total benefits of BANRURAL and FIRA borrowers totalled N$67 billion (US$17.7 billion) during 1983-92. IV. Rural Finance during the Salinas Administration (1989-1992) 6.63 In 1989, the GOM embarked on the liberalization of financial markets and reprivatization of commercial banks, and announced its Program for Structural Change and Financial Modernization of Development Banks. The major commitments under the program were to: (a) redefine the roles of development banks and trust funds; (b) eliminate of financial subsidies; (c) charge development banks and trust funds the full opportunity cost of funds raised by the GOM for on-lending to these institutions; and (d) limit transfers to development banks and trust funds. The goal of development finance institutions would be to help fund viable projects to producers with productive potential. The newly created PRONASOL would serve producers with little or no productive potential. 6.64 In 1993, the GOM moved to allow FIRA and BANRURAL to lend for rural productive activities other than agriculture, and to negotiate the restructuring of the agricultural portfolio of the banking system which is facing severe arrear problems. At present, the GOM is focusing its attention on understanding nonbank finance in rural Mexico. A. Streamlining the Responsibilities of Development Banks and Trust Funds 6.65 The GOM closed inefficient and costly institutions, such as ANAGSA, the Banco Nacional Pesquero y Portuario (BANPESCA), and FICART. It reorganized and downsized BANRURAL, set up new incentives schemes for commercial bank financing of small farmers in FIRA, and new transfer and financing programs under PRONASOL. 6.66 . The GOM also redefined the clientele of each institution: BANRURAL would serve poor farmers who might have viable investment possibilities, but who lack access to commercial financing, perhaps because they have no collateral or lending history; FIRA would serve all others borrowers, including poor ones with collateral and a lending history; and PRONASOL would serve poor farmers with what amounts to income transfers, not credit. 6.67 BANRURAL. BANRURAL was reorganized and downsized. It closed 300 of its 500 branches, and reduced its staff from 22,000 in 1988 to about 10,000 in 1992. Moreover, around 40 percent of its portfolio was transferred to PRONASOL (US$3.2 billion, benefitting 426,703 borrowers) and FIRCAVEN (US$l billion, benefitting 503,111 borrowers) in 1991. PRONASOL received the overdue loans of small farmers with little or no productive potential; borrowers were expected to repay their loans with labor. FIRCAVEN received all other overdue loans, which the GOM expected to recover, to allow defaulting borrowers who would make at least partial payments to start borrowing from BANRURAL again. In November 1992, the GOM announced an additional transfer of overdue investment loans to FIDELIQ. Nor the amounts transferred or FIDELlQ's  102 recovery rates are available. In August 1993, the GOM informed the Bank that FIRCAVEN recoveries were insignificant, and that it had closed this trust fund. 6.68 BANRURAL has been unable to serve its new target population, that is, PBls with viable investments but no collateral. With reduced government transfers, the need to increase loan recoveries accentuated, and management moved to serve those larger producers able to provide collateral. 6.69 FlRA. With the downsizing of BANRURAL, FIRA became the most important development bank in Mexico in terms of its contribution to agricultural financing. 6.70 In July 1992, FIRA designed and introduced a new system to link repayments to banks per farmer served rather than by loan size, and a separate system to reimburse banks for their costs of providing production technical assistance to farmers. The new transaction cost reimbursement scheme (known as SIEBAN) offered commercial banks a fixed decreasing annual fee, for two years, per new low-income borrower brought into the system. Banks could opt for this subsidy instead of the 6 percentage points intermediation margin. The annual fee was set such that the incentive to opt for the SIEBAN was greater the smaller the size of the loan. FlRA expected that lending to PBls would increase and that banks would indeed opt for the fixed fee. However, the value of loans to PBls remained at around 18 percent of total FIRA lending in 1992, and banks showed virtuaJly no interest in the transaction cost subsidy. In 1993, the SIEBAN was revised to provide commercial banks both, the intermediation margin and the two-year transaction cost subsidy, and to cover all Productores en Desarrollo. 6.71 Like BANRURAL, government funds through FIRA programs have reached also increasingly smaller numbers of poor producers, as the relative importance of PBI discounts to total annual lending decreased from about 50 percent in 1988 to 18 percent in 1992. Despite FIRA's efforts, commercial banks lack adequate incentives to lend to PBls relative to OPs, and continue to be reluctant to incur high transaction costs and bear the risk of lending to small, disperse, borrowers with risky projects, at interest rates that do not reflect their costs or risks of lending. 6.72 FlRA argues that the problem lies with the definition of PBI. According to FlRA, with the deterioration of the real wage and the divergent rise in input and output prices, the definition of PBI leaves out a considerable number of small farmers who deserve to qualify for preferentiaJ credit. In fact, FIRA's 1992 annual report notes that the qualifying level of annual family income for low- income producers (PBIs), established in 1992 at less than 1,000 times the daily wage, represents only 385 times the minimum wage of 1972 (in 1991 prices). While this is true, income and census data show that 80 percent of the rural population still have annual incomes below 1,000 daily wages. Although FlRA claims that its customers are not the entire rural population, the fact is that with the current structure of discounts 80 percent of FIRA's funds help finance the top 20 percent of the rural population. In effect, FIRA's customers are not significantly poor and are benefiting from Government transfers, when a vast majority of the rural population is still living in poverty. 6.73 PRONASOL - Solidarity Funds for Production. In 1990, GOM created the Solidarity Funds for Production (Fondos de Apoyo a la Producci6n, also known as credito a la palabra) to assist peasants excluded from coverage by BANRURAL--originally about 269,000 producers with 1.2 million hectares-with the cultivation of basic products in areas without irrigation and/or with low productivity. The program was later expanded to cover all other low-income producers of basic products in rainfed areas and/or in areas with low productivity who do not have credit with BANRURAL or another banking institution.  103 6.74 The Production Funds consist of payments made directly to producers. The payment is on a per hectare basis, up to a maximum of 3 hectares. Producers must reimburse the monies received under the program into a community fund. The community later decides on the use of recuperated funds, generally for community works or to form Solidarity Businesses. Those who fulfill their obligations and repay 100 percent of the amount received from the Production Funds or demonstrate their inability to pay due to a bad crop year or natural disasters are entitled to receive another full payment the following year. Those who repay at least 50 percent, are entitled to receive another payment equal to the amount repaid. New producers receive a first annual payment of N$350 per hectare. In 1993, the Production Funds planned to channeled US$350 million to over 1 million producers. 6.75 In 1990, about 90 percent of the Funds assisted maize and beans producers in rainfed areas. No breakdown is available for later years. Assuming the proportion of crops supported was the same in 1992, the area supported with Production Funds represents about 22 percent of the rainfed area cultivated with maize and beans in 1992, and about 35 percent in 1993. 6.76 PRONASOL - Solidarity Businesses. In 1991, the GOM constituted the Fondo Nacional de Empresas de Solidaridad (FONAES) to provide risk capital for productive projects for the low-income population. FONAES participates with up to 35 percent of total investment costs. With this investment, FONAES acquires an amount of the firm's assets proportional to its financing participation at the time the empresa is created; that is, FONAES does not hold a proportional interest in capitalized earnings. 6.77 FONAES has partially financed a number of urban and rural small businesses (Empresas de Solidaridad) under three different financing schemes: (a) through risk capital investments; (b) through trust funds and credit lines provided by other Government institutions; and (c) a combination of (a) and (b) or risk capital and other Solidarity funds. From the 3,643 empresas formed in 1992 and the 7,024 formed in 1993, most (52 percent in 1992 and 62 percent in 1993) were financed through trust funds established with FONAES, state governments, and other state businesses participation. Through these trust funds development banks have: (a) opened credit lines ten times the trust fund capital (in NAFIN, Mexico's National Finance Company) or (b) provided credit, 25 percent of which is guaranteed by guarantee trust funds (BANRURAL). 6.78 FONAES has funded 817 rural business with US$283 million, benefitting 166,235 producers with an average investment of US$I,700 per producer or US$346,000 per business. Forty- four percent of these funds served to finance 369 livestock production businesses with 8,300 producers, an average investment of US$15,OOO per producer or US$338,OOO per business. The rest of FONAES funds were invested in a wide variety in agriculture, forestry, fisheries, agroprocessing, and marketing businesses. 6.79 FONAES, however, appears to cater the better-off of the small producers, as producers need to contribute or borrow from other sources 65 percent of their investment costs. 6.80 PRONASOL - Solidarity Banks. In December 1992, FONAES undertook the creation of Cajas Solidarias (solidarity banks, a form of rural savings and loans association) to promote savings and credit opportunities for low-income farmers who receive Government support through the Solidarity Funds for Production. This program consists of promoting the use of recovered funds to form community savings and loan associations (an informa1 arrangement) which then gather to set up a solidarity bank (a legal entity).  104 6.81 Producers contribute their payments to the Solidarity Funds for Production program as capital to the rural bank which allows them to borrow up to 90 percent of their individual accumulated contributions. Because the first three years the capital of the bank increases with new repayments for the Production Funds, producers guarantee themselves access to increasing amounts of credit at competitive market interest rates to meet their production and consumption needs. 6.82 The banks also provide a savings outlet for small savers which also receive competitive rates on their deposits, and may borrow using their savings as collateral. 6.83 Members decide on the interest rates to be charged and paid, and on the distribution of profits. The spread must ensure a self-financing venture. If there is a surplus and profits are distributed, the effect of dividend distribution is to lower the effective interest rate charged on loans and to increase the yield on deposits. It should be noted that the adjustment takes place only ex-post, if the members' general assembly agree to distribute any profits, in which case 45 percent are retained to increase the capital of the rural bank. 6.84 Sixty-three solidarity banks have been formed in the past 18 months, comprising about 991 informal local banks in 19 states, serving the 38,134 members of 33 banks in fun operation (the rest is still under the process of organization and training of staff). In general, the number of members per bank ranges between 33 to 1,359, with two banks over the 1,0OO-member limit. Minimum capital ranges from US$49,OOO to US$604,OOO, with several banks (14) below the minimum required capital of U$160,000. 6.85 Personal savings accounts range from an average of US$14 to US$I,403 per bank, while the national average balance on saving accounts is US $447 . Loans average between US$179 to US$I,761, with a national average ofUS$494. 6.86 The experience so far has shown that although few banks have failed, a good majority is working fine, attracting savings from their founding members and from additional members of the community that do not receive support from the Solidarity Funds for Production. B. Elimination of Financial Subsidies 6.87 The GOM also sought to reduce credit subsidies, that is, interest rates on loans through development banks and trust funds were no longer to be subsidized,~ except for PBIs. First, in August 1988, it increased the interest rate charged on working capital loans to PBIs to 95 percent of CPP and in March 1991, to the CETES rate + 2, the interest rate applicable to OPs engaged in the production of basic products. 6.88 Second, FIRA introduced a new mechanism for short-term credit to OPs, under which commercial banks could choose to finance a larger proportion of sub loans rediscounted with FIRA than the required 20 percent of cofinancing. In exchange, commercial banks are allowed to set the interest rate on that portion of the subloan financed with their additional contribution. In other words, while FlRA funds and the stipulated contribution of 20 percent of the banks' own funds continued to be lent at the Government preferential rates, additional bank financing beyond the required 20 percent could be lent at free interest rates. This measure constitutes an important step 56. In Mexico. subsidies are measured as the difference between subsidized interest rates and a reference rate, such as CPP or CETES, but not against market interest rates. Preferential interest rates are on or above the reference rates, but below market rates.  105 toward the future liberalization of interest rates since the combination of interest rates charged for this type of loans, known as the rasa mezcla, tends to approximate market rates. 6.89 Third, and most important, the GOM control1ed inflation, which dropped from 159 percent in 1987 to 12 percent in 1992. 6.90 As a result of the above, interest rates to agriculture have been positive since 1988, and interest rate subsidies have declined. An estimate of the benefits accrued to FIRA's and BANRURAL's borrowers shows a significant decrease in financial subsidies, which fell by an estimated 76 percent in real terms. 6.91 Despite the above improvements in decreasing financial subsidies, the current structure of government imposed interest rates for rediscounted agricultural loans is too complex, has resource allocation costs, and increases lenders' transaction costs. c. Other Reforms 6.92 On-lending at the CETES rate. In 1989, the GOM abandoned its policy of transferring external funds as capital contributions to trust funds and development banks. Instead, it now lends these funds at its marginal cost of funds, the rate on the 28-day CETES. To compensate its institutions for the resulting losses of their on-lending at controlled interest rates and spreads to final borrowers and other financial intermediaries, the GOM makes specific budgetary provisions to cover the interest rate differential related to these operations, and assumes the foreign exchange risk. 6.93 Limiting transfers to development banks and trust funds was also a major part of the GOM's program. Fiscal and quasi-fiscal transfers dropped from 1.6 percent of GOP in 1990 to 0.9 percent in 1989 to 0.2 percent in 1992. 6.94 Financial sector reforms. The GOM took significant measures to reform the financial sector supported by the FSAL among which the reprivatization of the commercial banks and the liberalization of deposit rates to promote domestic resource mobilization are worth mentioning for their expected positive impact on the rural credit system. 6.95 Non-agriculture rural finance. Lastly, in 1993, the GOM moved to allow FIRA and BANRURAL to diversify their lending activities and cushion the risk in their loan portfolios. FIRA may now discount and BANRURAL make loans for a wide range of rural activities not related to primary production. 6.96 How has the performance of the reforms compared to its goals, as described in paras. 6.63 - 6.95? The GOM has made major achievements in rationalizing the development banks and trust funds and in reducing transfers to them. In general, despite the very positive reforms of the past five years, controlled on-lending interest rates and subsidized credit guarantees and agricultural insurance policies continue to require substantial government transfers to the major rural finance institutions and hamper the development of efficient rural financial markets. The GOM appears committed to credit subsidies as an element of social policy, although the bulk of subsidies have not accrued to the very poor but to the better-off segment of the rural population. Poor farmers have remained poor and their access to credit and other formal financial services is still difficult.  106 D. Nonperforming Agricultural Loans 6.97 The GOM has recently negotiated with commercial banks a program to restructure the nonperforming agricultural loan portfolio of the banking system.~7 The problem of nonperforming agricultural loans, however, is part of a more general trend of increasing nonperforming loans in all sectors. BANXICO reports that the overdue loan portfolio of commercial banks increased 55 percent in real terms between December 1992 and December 1993, while overdue loans to agriculture increased by 26 percent during the same period. 6.98 In agriculture, arrears have been generally due to the high real interest rates that have prevailed since 1992, combined with low profitability of sector investments. According to the Economic Indicators published monthly by BANXICO, as of December 31, 1993, loans to agriculture represented about 51 percent and 13 percent of the canera vencida held by development and commercial banks, respectively. 6.99 In the past few years, the GOM dealt with the nonperforming loan portfolio of development banks and trust funds differently. As mentioned earlier, it transferred 40 percent of BANRURAL's 1991 loan portfolio to the FIRCAVEN and PRONASOL. Only those loans transferred to PRONASOL constituted an outright debt forgiveness, as the GOM expected to recover, at least partly, loans transferred to FIRCAVEN. However, recoveries were insignificant and FIRCAVEN was closed. In November 1992, the GOM announced an additional transfer of overdue investment loans, this time, to FIDELIQ. Nor the amounts transferred or FIDELIQ's recovery rates are available. The current value of BANRURAL's nonperforming portfolio is not available. 6.100 In the case of FIRA, loan payment dates were closely monitored. Overdue loans were immediately rescheduled or refinanced, either by extending due dates or making contingency loans to consolidate producers' debts. During the past few years, FIRA's refinancing and loan consolidation programs have shown an upward trend. These refinancing and rescheduling mechanisms have somewhat disguised or prevented higher losses while helping commercial banks to show a healthy portfolio of FIRA rediscounted loans. 6.101 The Bank does not have any official, current information from the GOM on the magnitude of the overdue loan portfolio nor on the current restructuring program. The Mexican press, however, reports that both FIRA's and BANRURAL's overdue loan portfolio is being restructured. The program covers all producers (PBIs and OPs). Investment and working capital loans for up to US$67,000 fully or partly delinquent on March 25, 1994 qualify. The distribution of overdue loans by class of producers or type of loan is not available. 6.102 Under the program, about US$1.8 million of the commercial banks nonperforming agricultural loans. Part of these loans had been previously discounted with FIRA and were in arrears. FIRA will discount an additional number of loans, originally financed with the commercial banks' own resources. In the case of BANRURAL, the press reported that the delinquent loans of 60,000 farmers, amounting to US$730 million (45 percent of BANRURAL's portfolio) are being restructured . 6.103 Restructuring terms are as follows: 57. In Mexico, loans are classified as nonperfonning immediately after one payment has been missed.  107 (a) 15·year amortization schedule, with an eight-year grace period for repayment of capital, and a five-year grace period for interest payments during which only the fixed portion of the interest rate will be amortized; (b) the interest rate will be the rate of inflation plus a fixed margin of 4 points for PBls and 4.5 for OPs; and (c) FIRA will discount these loans up to US$67,000 and will guarantee 50 percent of the outstanding portion of the loans. 6.104 This massive restructuring implies that neither FIRA's nor BANRURAL's clientele will shrink since previously delinquent borrowers will have a good credit standing and, therefore, will be eligible to receive credit by paying just 4-4.5% of their old debt for the next five years. 6.105 Issues. This debt restructuring program raises two main concerns. One is the implication that it could have on the capacity of FIRA and BANRURAL to maintain their real level of operations in the future without additional government transfers. The other, and more important, is the disincentive that this massive restructuring of loans creates for borrowers currently in good standing to continue repaying their loans and for continued commercial bank lending to agriculture. 6.106 The GOM can either choose to have FIRA and BANRURAL reduce their real level of operations or substantially increase government funding to these institutions. With large parts of FIRA (about 40 percent) and BANRURAL (about 45 percent) loan portfolio placed now at 15 years, and at lower than originally contracted interest rates, fiscal transfers to these institutions will have to increase or their lending program will have to shrink. 6.107 The very favorable terms of the restructuring program are likely to deter borrowers currently in good standing to continue repaying their loans at less favorable terms than their delinquent neighbors. If this happens, restructuring programs will need to be extended later to all borrowers regardless of their credit standing. 6.108 The Mexican press has estimated that commercial banks will lose over US$300 million during the first four years of the agreement. Prior to this agreement, the Mexican Banking Association had also agreed to waive delinquency charges for 1992 and 1993, but no information on the cost to the banks is available. Despite these losses, the banks have-agreed with the GOM to provide working capital loans for the next agricultural cycle to producers with viable projects. 6.109 The US$300 million in losses to be incurred by commercial banks in the next four years will also constitute a detriment for continued commercial bank lending to agriculture. The latter had increased substantially since 1987, especially with internal bank funding. In 1987, commercial banks held US$583 million (real 1992 US dollars) of discounted FlRA loans and another US$460 million of internally funded loans. By 1992, they held US$4.2 billion of discounted FlRA loans and US$5.2 billion of internally funded loans, a 623 percent and 1,000 percent increase respectively. 6.110 In summary, the GOM will have to choose how much to support agricultural financing in the near future. With reduced funds for lending due to the extended recovery period of the restructured portfolio and decreased interest income, the GOM can choose: (a) to support FIRA and BANRURAL lending program in real terms through increased fiscal transfers and expanded rediscounts from BANXICO; or (b) let FIRA and BANRURAL live up to the consequences of the restructuring program, that is, a substantially reduced lending program. Because a reduced lending program to  108 agriculture will have adverse political and economic effects, particularly during the NAFTA/PROC~MPO transition period, it is more likely that the GOM will increase its support to FlRA and BANRURAL operations at least partially, to avoid a significant drop in financing to the sector. V. Recommendations 6.111 The cost effectiveness of rural credit and the permanence of rural finance institutions seems dubious. The effective intermediation of financial resources in rural areas and the ability to reach small farmers requires additional changes in financial policies for the rural sector. It also requires to effect additional institutional changes. The following recommendations are based on the Bank's past lending experience for agricultural credit in Mexico. A. To improve the efficiency or rural financial markets and restore the incentives to commercial banks or lending to PBIs, the level and structure or interest rates should be simplified and PBI rates increased relative to OPs, rollowed by a rreeing or interest rates 6.112 Interest rate subsidization continues to reduce the resources available to PBIs and encourages rent-seeking behavior, accentuating the need to eliminate interest rate subsidies. Moreover, Government controlled interest rate policies for a rather small group of better-.off producers (compared to PBls who make up the bulk of the rural population) who receive the bulk of FlRA's and possibly BANRURAL's funds, leads to unfair competition. Commercial banks cannot compete with the on-lending interest rates set on FlRA and BANRURAL funds as these rates do not reflect the costs and risks of lending in agriculture. 6.113 Interest rates subsidies should be phased out. The ultimate target, however, will be to liberalize on-lending interest rates to agriculture, to allow FlRA, BANRURAL, and commercial banks to set freely their lending rates on all their loans. Bo' To the extent that subsidies are desired to help the poor, they should be transparent and well targeted 6.114 The rural credit reforms of the late 1980s had the explicit objective of reducing the cost to Government of the rural financial system, which was accomplished. Despite the segmentation of rural borrowers into three different agencies (namely, PRONASOL, BANRURAL, and FlRA), it is not be possible to conclude today that the benefits of Government remaining financial support to rural credit institutions accrues to poor producers with viable investments. FIRA's clients are largely OPs, and BANRURAL's new strict collateral requirements hinders access to formal credit to many PBIs. PRONASOL, however, appears to reach poor producers through its transfer payment program (Programa de Apoyo ala Producci6n), but serves richer groups through its Empresas en Solidaridad program. To support the rural poor, the GOM should: (a) continue supporting the two Solidaridad programs: the Production Funds and Empresas, although the latter needs to be restructured and targeted towards subsistence producers; and (b) expand financial services offered jointly with these programs (Cajas de Solidaridad).  109 C. Ensure the institutional permanence or F1RA and BANRURAL through financial selr· sufficiency 6.115 F1RA needs to become financially self-sufficient, except for a portion of its costs of reimbursing transaction costs to commercial banks and its costs for technical assistance. To this end, FIRA should: (a) discount subloans at a rate equal to its cost of funds plus a spread to cover its costs of financial intermediation. To this end, FIRA needs to introduce cost accounting procedures to establish the cost of its different services; (b) simplify its interest rate structure and, while interest rates remain controlled, reduce the interest rate subsidy and its rediscounting margins to encourage commercial bank financing; and (c) discontinue its complementary loan guarantee system as the program has not been successful to increase commercial bank lending to PBIs. 6.116 In addition, to be more effective and better serve the needs of rural borrowers, FIRA should: (a) reorient its emphasiS on credit and technical assistance, away from basic grains toward a broader range of activities; and (b) favor discounting investment loans over short-term loans. 6.111 To induce increased commercial bank lending for PBIs, FIRA should continue providing a one-time reimbursement to banks for the transaction costs of bringing new, PBIs into the system. However, FIRA needs to modify this new program, which now covers all Productores en Desarrollo I, to cover the reimbursement of transaction costs only for PBIs. 6.118 Technical assistance appears to be FIRA's most effective mechanism to support increased agricultural production through adoption of new technologies and improved agricultural practices. In view of the social and economic gains of this investment, and the efficiency of FIRA's technical assistance programs, it seems justified that GOM continues to pay at least part of the costs of providing these services. 6.119 Finally, Government transfers to FIRA should be made more transparent. FlRA should pay CETES rates on all its outstanding loans with BANXICO. While interest rates remain controlled and below market, Government should absorb the costs of FIRA's rediscount margins to commercial banks through regular budget transfers. 6.120 BANRURAL needs to be weaned from Government transfers. BANRURAL has moved closer to the desirable objective of financing only producers with commercial potential. However, in so doing, BANRURAL has turned to the larger producers with collateral, leaving the viable PBIs with scarce collateral without access to the formal credit system. Despite these changes, BANRURAL is not yet a financially viable institution. Its loan recovery rates are still low and its costs are high compared to the commercial banks, and continues closing unprofitable branches. Additional reforms are needed for BANRURAL to be useful to the rural sector. To this end, BANRURAL should:  110 (a) continue current efforts to analyze loans carefully, evaluate the likelihood of repayment objectively, and design and implement sound policies to mobilize savings; (b) continue cutting operational costs by closing money-losing branches, reducing its work force and improving management; (c) design and implement a system of incentives where employees are rewarded for loans recovered and resources mobilized, and borrowers are encouraged to repay; (d) increase its efforts to attract savings in rural areas, for example by paying competitive interest rates on deposits and issuing long-term savings instruments with small penalties for early withdrawals, and aggressively promoting small savings as the basis for financing future consumption and production needs; (e) test innovative ways to serve the PBI clientele. for example by accepting and promoting collateral substitutes such as guarantee bonds (avales). group joint liability, character references, etc.; and (f) stop financing larger borrowers who could gain access and be better served by private financial intermediaries. D. Abolish insurance premium subsidies to promote competition between private agricultural insurance and AGROASEMEX 6.121 Experience with agricultural insurance in Mexico has been disastrous. The existence of insurance did not stabilize rural income, notwithstanding its large coverage. Although there was not an explicit subsidy on insurance premium, the high probability of getting compensation with fraudulent claims was equivalent to a massive subsidy to producers. An important step was to close ANAGSA in 1990. AGROASEMEX is now more cautious in providing coverage and accepting claims, but the insurance subsidy, although an explicit 30 percent ad valorem rate, is still present. As there is not a good economic reason to subsidize agricultural insurance, the agricultural insurance system needs to develop on a commercially sound basis which can only be achieved by letting firms compete for customers' without subsidizing insurance premia.  111 7. FISCAL POLICIES FOR THE RURAL SECTOR 7.1 The GOM has spent considerable resources on rural development. Expenditures in the rural sector (excluding social services) have averaged 37 percent of agricultural GOP since 1982. This chapter reviews the composition of rural expenditures and evaluates how they have affected efficiency and poverty alleviation in the rural sector. It also assesses the impact of the GOM's taxation policy on incentives for private investment in the rural sector. The principal conclusion is that despite improvements in fiscal policies, particularly during the Salinas Administration, significant scope exists to strengthen the impact of fiscal policy on efficiency and poverty alleviation in the rural sector. The first section of this chapter examines the overall level and composition of public expenditures in the rural sector since 1982. The second section focusses on the procedures used to allocate resources in the sector and identifies several issues undermining the impact of expenditures on efficiency and poverty alleviation. The third section evaluates the effect of various taxes on incentives for private investments in the rural sector. The last section presents the principal conclusions and recommendations. I. Evolution or Rural Development Expenditures 7.2 This section reviews overall expenditure trends by federal, state and municipal governments in the rural sector. Despite efforts to decentralize expenditures to state and municipal governments, the large majority of public expenditures in Mexico are made by the federal government. In 1989, the combined budgets of the states was 13 percent of the total federal budget, while municipal expenditures were 3 percent of the total federal budget. ~8 Nonetheless, expenditures by municipalities and states are increasing while expenditures by the federal government are decreasing. Aggregate expenditures by state and municipal governments each rose by about 20 percent between 1986 and 1989, while federal expenditures decreased by 12 percent during this period. The federal government appears to be the principal source of public resources in rural areas, as data from Aguascalientes suggest that the share of rural development expenditures by state governments is relatively small (5 percent). The data also suggest that Solidarity resources, although significant and increasing~ are not the principal source of investment funds for the state and municipal governments. Similarly, Solidarity resources appear not to dominate federal investment flows to the states. A. Federal Expenditures 7.3 Federal programs in the areas of agriculture, fisheries, rural credit, land reform, rural roads, rural electrification, and other basic infrastructure are reviewed from 1982 to 1992.59 This involved collecting data from SARH, SRA and SEPESCA and their dependent organizations. In addition, information was collected from SEDESOL (Solidarity), SHCP (credit), BANXICO (credit), SECOFI (CONASUPO), CFE, SCT, FERTIMEX and INEGI (land titling) for programs with a significant rural focus. The analysis of the fiscal and quasi-fiscal transfers to rural credit and insurance schemes 58. Aggregate state and municipal data is from: INEGI. Ellngreso y Gasto Publico en Mexico (1992). 59. The dermition of rural development expenditures used in this paper differs from the GOM dermition which includes only expenditures by SARH, SRA, and DFls.  112 is presented in Chapter 6 and therefore is not treated in this chapter. The expenditure data is classified by program and by functional category (recurrent or capital). Unless otherwise indicated, the data reflects actual expenditures taken from the national accounts (Cuenta Publica) prepared by SHCP. All federal expenditures at the state and local levels as well as transfers to dependent institutions are included. All values are in constant 1992 prices. The aggregate level and composition of rural development expenditures are presented below, while annex 3, appendix 1 contains a detailed description of expenditures by institution and annex 3, appendix 2 presents the statistical data. 1. Level of Expenditures 7.4 Reflecting overall reductions in the federal budget, rural development expenditures declined significantly during the de la Madrid and Salinas Administrations. During the de la Madrid Administration, the total federal budget declined by 5 percent, while during the first four years of the Salinas Administration (1989-92) it dropped by 19 percent. Similarly, rural development expenditures receded by 38 percent during the de la Madrid Administration and decreased by 11 percent between 1989-92 (Figures 7.1 and 7.2). The brief peak in rural development expenditures in 1990 reflected the GOM's decision to assume the debt of the DFIs and increased marketing costs for CONASUPO. By 1992, rural development expenditures were N$21.3 billion (US$7.1 billion), compared to N$3S.2 billion (US$ll. 7 billion) in 1983. During the entire decade, the share of rural development expenditures averaged 8 percent of total federal budgetary outlays, roughly equal to the share of agriculture in Mexico's GDP. The share of rural development expenditures in agricultural GDP dropped steadily from 49 percent in 1983 to 24 percent in 1992. 7.s Despite the expenditure reductions of the 1980s, the relative level of agriculture spending in Mexico remains higher than in other developing and industrial countries. The share of agriculture in total expenditures in 1992 equaled the share of agriculture in GDP in the case of Mexico; whereas in the sample of industrial and developing countries, the share of agriculture in the total budget was significantly less than the share of agriculture in the GDP (fable 7.1). Similarly, the share of agriculture expenditures of agriculture GDP in Mexico was 39 percent compared to 21 percent and 12 percent for the industrial and developing countries. Table 7.1. Government Expenditures for Agriculture in Industrial and Developing Countries (percent) Mexico Mexico 15 Industrial 40 Develop. (1984-88) (1992) Countries Countries (1984-88)* (1984-88)* Ratio of agriculture's share in total expenditures over agriculture'S share of GDP 0.9 1.0 0.6 0.4 Agriculture's share in total expenditures 8 8 2 1 Share of ago expend. in ago GDP 39 23 23 12 SOURCE: Bonni van Blarcom, Odin Knudsen and John Nash, "The Reform of Public Expenditures for Agriculture.· World Bank Discussion Paper 216, at 5.  .. Federal and Rural Development Expenditures 500000 450000 II FEDERAL BUDGET • Rural Development ~ 400000 c. ('10 "'f e. ~ 350000 .- :l C. ::tl I: "'f III c 300000 ~ -"Tj a t:1 _. (JQ ('10 ...... 'E < I: ('10 "'f ...... W <'II 250000 _('10 - m Q ......:J "'C. m ..- 3 ('10 ~ 200000 .... :l l"'1 ;.< "0 ('10 150000 _. :l C. .... I: "'f ('10 100000 ..." 50000 o -;- 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 m:\mx\asm\dala\rd2.xls  Rural Development Expenditures by Institution 40000 CI'.l c _Other a 350001 IDOFls a ~ -CONASUPO ~ <:> -, mSOLIOARlOAO 30000 + I-SARH " c .., ::. t::l «'II III c: 0 25000 - < «'II <:> "0 3(JQ ~ _. ...... ...... 'E «'II .., ::2 = ..... «'II .I:"- N 20000 II ~ O'l t:"I:-.J O'l ~ N ..- "0 «'II ~ ::2 2 c. 15000 :::r .., = til Vl 0'" '< 10000 I H~~:~~~~~~jr~~i~~H b:;~MJ.*I:!f::~:1 h:::i:k::::::::::;::l ..... ::2 Vl ::. -- ..... 5000 I "." .. , 1@~~~\~~@tJ r~"'''''''"''';'FB' = ..... o· ::2 o 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 m:\mx\asm\data\rd2.xls  115 2. Composition of Expenditures 7.6 Programs. The tendency to favor market interventions at the expense of public goods was overturned during the first four years of the Salinas Administration. During the de la Madrid Administration, the share of the rural development budget allocated to market interventions (credit, marketing and fertilizer subsidies) increased from 31 percent to 66 percent, mainly due to rising transfers to support rural credit programs. Also during this period, the share of resources for public goods declined from 26 percent to 18 percent (Figure 7.3). The latter principally reflected decreases in the two largest public good subprograms - infrastructure and water development. Expenditures for administration and policy analysis also decreased during this period, mainly due to declining personnel costs. In contrast, the share of resources allocated to market interventions dropped almost in half, while the share allocated to public goods rose, during the Salinas Administration. The increased share of public goods reflects expanded expenditures for infrastructure and water development. By 1992, market interventions accounted for 27 percent and public goods had increased to 35 percent of the rural development budget. This compares favorably to Brazil, which spent 49 percent on market interventions in 1992 and 40 percent on public goods and land reform. ro Nonetheless, the share allocated to market interventions and transfers risks to increase in the 19908 with the continued expansion of the Solidarity credit programs (Solidarity Funds for Businesses and Production), and in particular with the PROCAMPO program (see Chapter 2). Also during the Salinas Administration, the proportion of resources allocated to administration and policy analysis more than doubled, due to increasing costs to prepare NAFI'A and the regulatory changes in the sector (Reform of Article 27 of the Constitution, revised Water and Forestry Laws). 7.7 Institutions. The budget of the institutions in the sector has reflected the policy priorities of each Administration. During the de la Madrid Administration, the budgets of the line secretariats were cut more severely than the budgets of CONASUPO and the DFls, which provided marketing and credit support respectively (Figure 7.2). More specifically, the budgets of SARB, SRA, SEPESCA, SCT (rural roads) and Solidarity/SPP fell by more than 50 percent, while the budgets of CONASUPO and the DFls dropped by only 41 percent and 22 percent respectively. During the first four years of the Salinas Administration budget allocations decreased for the institutions providing subsidies and increased for the institutions providing productive investments. In particular, the budget of CONASUPO declined by 25 percent and the budgets of the DFls and FERTIMEX dropped by 70 percent or more. These significant cuts were somewhat offset by a tripling of allocations to Solidarity, as well as smaller, but still significant, increases in the budgets of SARB, SRA, SEPESCA and IN!. Table 7.2 provides an overview of expenditures by institution and program. 7.8 Reflecting the decreased importance of market interventions during the first four years of President Salinas, the budgets of CONASUPO and DFIs decreased from an average of 61 percent of the total budget for 1983-1988, to 30 percent for 1992. Reflecting the increased importance on investments, particularly water infrastructure, the budget of SARH rose from an average of 25 percent during de la Madrid to 34 percent in 1992. The budget share of SPP/Solidarity also rose significantly during the Salinas Administration from less than 5 percent during 1984-1988 to 19 percent in 1992. 60. World Bank, "Brazil The Management of Agriculture Rural Development and Natural Resources: Volume II,· at 107.  Rural Development Expenditures by Program (in percent) 100% -, .... ::-: -:-, -, ::' :0,-: --:- _'''-';-:-0:' - :-:-,"'-::-:--' ,-,....,.::-:::,:-".:-:-:-:-:,-:,'-:-,-,-:-:-::-:-,."",:::,:-:::.:;.:,,-;-:-:--:--:--:-:-:::-:--- 80% " = ~ :. o t1> ~ !!.. o '0 3 t1> liO':" ::l "Tl o Public Goods ~ (~ ;..< ~ .... .... aOther g '0 t1> ;---l 0'\ Q.,(.;.j 110% OMa/ket 2 IlIterV(!lItioll5 rri .. Administr atioll ami .9." Policy Analysis '"C ~ o (JQ 20% i! 3 0% N Q) ('I) Q) q- Q) It) Q) <.0 Q) ...... Q) Q) Q) en Q) o en ... m N m m ..... m .- m .... m .- m ..... en m ... ... m en ... m m ..- M:\mx\asm\dta\rdprog2.xls  117 Table 7.2 Summary of Rural Development Expenditures by Institution Share Share Market Adm. and Public Land Other Total Investment Recurrent * Interventions Policy Goods \5 Tenure iSARH 1983-88 58% 42% 0% 26% 65% 0% 9% 100% 1989-92 52% 48% 0·* 25% 71% 0% 4% 100% SRA 1983-88 3% 93% 0% 22% 15% 63% 0% 100% 1989-92 11% 89% 0% 23% 3% 74% 0% 100% SEPESCA 1983-88 28% 72% 0% 66% 28% 0% 6% 100% 1989-92 30% 70% 0% 54% 42% 0% 4% 100% CONASUPO \1 1983-88 0% 100% 100% 0% 0% 0% 0% 100% 1989-92 0% 100% 100% 0% 0% 0% 0% 100% DFf \1 1983-88 0% 100% 100% 0% 0% 0% 0% 100% 1989-92 0% 100% 100% 0% 0% 0% 0% 100% Solidarity/SPP \2 1983-88 100% 0% 0% 0% 100% 0% 0% 100% 1989-92 100% 0% 21% 0% 79% 0% 0% 100% CFE\6 1983-88 100% 0% 0% 0% 100% 0% 0% 100% 1989-92 100% 0% 0% 0% 100% 0% 0% 100% SCT 1983-88 100% 0% 0% 0% 100% 0% 0% 100% 1989-92 97% 3% 0% 0% 100% 0% 0% 100% INI \3 1983-88 50% 50% n/a n/a n/a n/a n/a n/a 1989-92 50% 50% n/a nla nla nla nla n/a INEGI (land) 1983-88 0% 0% 0% 0% 0% 100% 0% 100% 1989-92 80% 20% 0% 0% 0% 100% 0% 100% FERTIMEX \1 1983-88 0% 100% 100% 0% 0% 0% 0% 100% 1989-92 0% 100% 100% 0% 0% 0% 0% 100% Total 1983-88 21% 79% 61% 10% 23% 1% 5% 100% 1989-92 31% 69% 52% 9% 29% 3% 7% 100% • Expenditure breakdown for recurrent and investment costs is only classified as part of services \1 All costs are assumed recurrent and for market interventions. \2 Recurrent costs appear in budgets of Sedesol and other implementing agencies for Solidarity programs. \3 It is assumed that 50 percent of INl's rural development budget (excluding social services) is recurrent. \4 Includes credit and input and output subsidies. \5 Includes expenditures for: research and tech., water development, infrastructure, education, extension, plant and animal health. ASERCA marketing supports for rice, sorghum, wheat and soja are technically market interventions, \6 All costs assumed to be capital investments for rural electricity and transportation expenditures.  118 7.9 Recurrent and Capital Costs. The overall level of capital expenditures has increased since the mid-1980s due to the increased priority placed on public goods and investment (see figure 7.4).61 The share of recurrent costs dropped from 75 percent of total rural development expenditures in 1985 to 55 percent in 1992, mainly due to decreasing expenditures by the DFIs, CONASUPO and FERTIMEX. Recurrent costs also declined for SEPESCA and SRA, reflecting decreasing transfers to dependent agencies (the majority of which have been privatized or liquidated) and lower operating costs (mainly due to personnel reductions). 7.10 Until 1992, SARH was the principal investor in the sector • however, since then, Solidarity has emerged as the largest source of rural investment funds. However, the importance of Solidarity in the capital expenditure budget is misleading, as its entire budget is classified as investment, despite the fact that 25 percent (1992) supports quasi-credit programs (Solidarity Production Funds and Solidarity Business Funds). The approach of putting all funding for Solidarity exclusively in Chapter XXVI as capital spending helps create the impression of a very efficient program, since there is no current expenditure associated with the program. However, it also makes it impossible to monitor the full costs of implementing Solidarity. Most of the current expenditure for the implementation of the Solidarity Program is included in the recurrent budget of SEDESOL although other federal secretariats involved in the program also contribute. In addition, state governments and municipalities also incur current expenditures because of their involvement in the implementation of the program. B. State and Municipal Expenditures 7.11 Decentralization of expenditure responsibilities from the federal government to state and municipal governments has been a key tenet of the de la Madrid and especially the Salinas Administrations. The aim has been to strengthen the normative role of the federal secretariats and to transfer responsibilities for investments and the provision of services to state and local governments. This section reviews expenditures by state and municipal governments in order to: (a) review the extent to which expenditure responsibilities have been transferred to state and local governments; and (b) obtain a more complete overview of rural development expenditures in Mexico. Unfortunately, the analysis which follows is very limited, particularly with respect to the second point, as only aggregated data was available for one state. Moreover, data on municipal and state expenditures differed significantly from once source to another. 62 Nonetheless. the analysis does suggest increasing decentralization of expenditures to state and municipal governments and implies that a relatively small share of state resources are for rural development. 7.12 Information on total expenditures and revenues generated at the state and municipal levels is drawn from four case studies (fabasco, Sonora, San Luis Potosi and Aguascalientes), INEGI, and the eight states in the Decentralization and Regional Development Project II (Oaxaca, Guerrero, Chiapas, Hidalgo, Zacatecas, Michoacan, Veracruz and Puebla).63 61. The SHep Cuenla Publica presents two different data series for recurrent and capital expenditures. The data presented here is based on the expenditure breakdown by program and category (e.g. recurrent or capital) and is only available since 1985. It differs from the tables presenting aggregated capital and recurrent expenditure data. 62. Data provided by the state finance secretariats and presented in the state annual statistical yearbooks differed from data presented in the INEGI publication: Ellngreso y GaslO Publico en Mexico (1992). 63. The INEGI data is from: INEGI. Ellngreso y GaslO Publico en Mexico (1992).  Rural Development Expenditures by Category (in %) 35,000 30,000 II ~ = 25,000 + I et tj (t -< III C - (t Q ." ~ 20,000 a "e - (t = lo!j '2'" _. n - o Total Capital C\I en en :: 15,000 z - r-l • Total Recurrent ." '""'." ~(t(t = Q,' '-' ..... .... A. c ~~ ., ., ~ 1.0 10,000 t_ 5,000 -i- - - - - - fl ~ n ... ~ J8 ~ Q o 1985 1986 1987 1988 1989 1990 1991 1992  120 7.13 State Expenditures. In contrast to the federal budget, which has shrunk since the 1980s, expenditures at the state level have increased significantly, facilitated by increased levels of revenue sharing from the federal government. 64 Table 7.3 indicates that average expenditures in Aguascalientes increased by 570 percent in 1991-92 compared to the mid-1980s. Federal revenue sharing to Aguascalientes increased by more than 750 percent in 1991-92 compared to the annual average during mid-1980s. Similarly. the combined budgets for the states in the Decentralization and Regional Development Project n (ORO m. excluding Chiapas and Guerrero. increased by about 90 percent between 1987 and 1992. facilitated by a 63 percent increase in federal revenue sharing. 1SS Table 7.3 Expenditures and Revenue Sharing for Aguascalientes 1982-1987 Average 1991-1992 Average % Change Total Expenditures N$81.1 million N$471.6 million 570% (US$26 million) (US$152 million) Total Revenues N$81.1 million N$472.9 million 570% (US$26 million) (US$153 million) of which: N$44.4 million N$337.8 million 660% Revenue Sharing (US$14 million) (US$109 million) SOURCE: Annual Statistics Bulletin of Aguascalientes. 7.14 Total investment in the states remains dominated by the federal government, but federal investment appears to have decreased. while state investment expenditures have risen. Total federal investment declined by 4 percent in Aguascalientes between 1988 and 1992, while total state investment rose by more than 450 percent. Similarly, federal investment in the DRD II states dropped by 22 percent between 1988 and 1992, while state investment increased by 53 percent. Nonetheless. federal investment still dominates. particularly in the poorer states: in Aguascalientes. federal investment was 20 percent greater than state investment. Likewise, in the DRD II states federal investment was 225 percent greater than state investment in 1992/>6 7.15 While the importance of Solidarity investments has grown both in the federal and state investment budgets, its overall share of federal investments to the states remains limited. In 1992. Solidarity accounted for 30 percent of federal investments in the DRD II states and 22 percent in Aguascalientes. Similarly, state matching contributions for Solidarity programs were about 30 percent of the state investment budgets for the DRD II states and only 13 percent of Aguascalientes' budget. 7.16 Expenditure data by program category was only available for Aguascalientes. In this state, expenditures for agriculture development and credit, land titling and Solidarity programs were less 64. For example, in Aguascalientes revenue sharing provided 41 % of total state resources in 1982 and 74% in 1992. 65. Information concerning the states in the Decentralization and Regional Development Project is from: Sergei Soares, "Intergovernmental Relations in Mexico: An Examination Focusing on Disadvantaged Rural States and Municipalities; Working Paper to assist in the Preparation of the Second Decentralization and Regional Development Project, 1993. 66. However, if federal investment data is taken from the Informe Presidencial, as opposed to the state finance secretariat, federal investment is approximately equal to state investments in the DRD II states.  121 than 5 percent of the total state investment program (expenditures for marketing, electricity and road construction were not taken into account).67 7.17 Municipal Expenditures. Overall municipal expenditures have increased since the late 1980s, reflecting increased resources due to higher revenue transfers, taxes and fees (Table 7.4).68 " Table 7.4 Growth in Municipal Expenditures Percent Change Percent Change 1986-92 1990-92 Aguascalientes 110% nla Sonora 64% 35% Tabasco 21% 3% 3 sample municipalities located in Puebla and Guerrero nla 114% SOURCE: 1990-92 data is from the Annual Statistical Bulletins for Tabasco, Sonora and Aguascalientes. Data for 1986-89 is from: INEG!, El lligreso y GaslO Publico ell Mexico (1992). The majority of expenditures (50-75%) are allocated to recurrent costs. While the share of investment costs in the municipal budgets varied by state (10-40%), the overall share allocated to investment by municipalities remained constant within each state. In Aguascalientes, total municipal expenditures were significantly less than state expenditures: in 1992, total expenditures by the 9 municipalities in Aguascalientes were only 35 percent of the state budget. Similarly, in 1989, total municipal expenditures in Mexico were 22 percent of total state expenditures. n. Evaluation of Rural Expenditure Policy 7.18 What has been the impact of government expenditures in the rural sector, and in particular, on rural efficiency and poverty? A detailed evaluation of the specific programs and their impact on efficiency and poverty is outside the scope of this study. Rather to address these questions, this section first examines the processes used to allocate resources in the rural sector and identifies four factors that undermine the efficiency of public expenditures. It then evaluates the extent to which certain federal programs are targeted towards the poor and reveals that the poverty focus of these programs could be strengthened. 67. Following the same proportion for the national Solidarity program, half of total state Solidarity expenditures were assumed to be in rural areas. 68. Municipalities in the states of Aguascalientes and Sonora relied on revenue sharing for 45-50% of their total budget in 1991-92, compared to approximately 90% for the poorer states of Tabasco, Hidalgo (3 sample municipalities), Puebla (2 sample municipalities) and Guerrero (one sample municipality). Sources: Contadurias Mayores de Hacienda Estatales; INEGI for sample municipalities in Hidalgo, Puebla and Guerrero. Annual Statistical Bulletins for Tabasco, Sonora and Aguascalientes.  122 A. Efficiency in Rural Expenditures 7.19 The efficiency of rural expenditures in Mexico could be improved. The following four sets of factors undermine the efficiency of rural expenditures. First, the absence of a well-defined rural development strategy has led to an expenditure program that does not necessarily reflect the development priorities of the sector, nor include the most efficient investments. Second, efficiency gains from decentralization are limited by an inadequate assignment of expenditures among the • different levels of government, a lack of clear distribution criteria for Solidarity resources: and insufficient revenue sources for local and state governments to carry-out their expanded responsibilities. Third, the absence of mUlti-year budgets undermines investment planning as well as the availability of resources to support operations and maintenance activities. Fourth, the budget process is very centralized and controlled by SHCP and as a result, the budget does not always reflect the expenditure priorities of the sector. This section reviews those four factors in more detail. 1. Expenditure Strategy 7.20 The allocation of federal funds in the rural sector does not necessarily reflect an overall development strategy nor the quality of individual investment projects, despite a detailed planning process (see Box 7.1) and recent improvements in project evaluation requirements. The principal causes of the problem are: (a) the multitude of rural development institutions in Mexico and their lack of coordination; (b) the large proportion of resources allocated via political commitments; (c) an inadequate set of project evaluation criteria coupled with weak institutional abilities to implement them; and (d) weak coordination of federal and state investment at the state level. As a result, resources often have been channelled towards poor quality projects/programs and there is a lack of coordination in rural development expenditures leading to overlapping programs, in some cases, and in others to underfunded programs. Some specific examples of some of the problems that may arise are: (a) Rural Roads. Rur3.I roads are constructed by SARH, CNA, the Secretariat for Communication and Transport (SCT), PEMEX, Solidarity, the state highway agency, the state rural development agency, and the state social development agency. Yet, the individual investments of these agencies do not reflect the priorities of a well-defined and agreed upon rural transportation network scheme, nor are they subjected to a standard set of economic and environmental criteria. Rather, they reflect generally a haphazard response to a variety of community and investment project requirements. Moreover, there is no clear strategy as to which agency is responsible for the maintenance of the roads. As a result, there is a large rural road network in Mexico that is poorly utilized and maintained. (b) Forestry Development. Within the forestry sector, the policies of SARH and SEDESOL are not sufficiently integrated. There is also a lack of clear and consistent set of environmental norms commonly agreed upon by SEDESOL's National Environmental Institute and SFOR (SARH's Forestry Subsecretariat). (c) Extension Services. Extension and technical assistance are provided by several institutions with different policies. SARH, FIRA, Solidarity (Solidarity Funds for Business and Production) and SRA provide technical assistance directly to producers, as well as the state rural development secretariats. Yet, many of the poorest  123 producers, who are the least able to purchase private technical assistance, have minimal or no access to government extension programs. 7.21 Political Commitments. Regardless of the project's economic and environmental rating and the allocation of resources arrived at during the budget process are never final. Priority is given in the allocation of resources across sectors to political commitments (compromisos) of the President. The allocation of investment resources within a particular sector can be similarly affected by the compromisos of the secretary and even general directors. Typically no new resources are sought from Congress, or from the SHCP, to satisfy these new priorities. Instead, existing projects experience funding cuts or are altogether postponed. The deleterious impact of this practice is in direct proportion to the share of total investment resources that eventually are diverted. 69 No official estimates exist of what share of investment funds are finally allocated in the capital budget via compromisos. The casual evidence suggests that they represent a substantial share of total resources. For example, in the case of CONACAL (Comisi6n Nacional de Caminos Alimentadores) a large share of the project portfolio appears to be absorbed by the compromisos. The importance of this exceptional route in the allocation of public funds is underlined by the existence of specialized offices in the President's staff and each secretariat for the monitoring of the compromisos. 7.22 Project Selection Requirements. Despite several improvements, the process of project selection still does not consistently follow a rigorous evaluation methodology. Since the mid-1980s, project evaluation techniques have improved considerably in Mexico, and funding approval and general ideas are no longer sufficient for a secretariat or agency to start disbursement on an investment project. SHCP now requires a project evaluation study before funds are allocated and expenditures authorized for a particular project. Nonetheless, 20-25 % of investment projects are not supported by a formal study. When an evaluation study is prepared, in many cases it is just completed to fulfill the bureaucratic requirement and does not provide an accurate estimation of the project's o~rall qUality. These problems are present even in those agencies where more rigorous application of evaluation techniques is expected, such as CNA. Moreover, the SHCP does not have the staff to carefully review all the feasibility studies submitted by the secretariats and decentralized agencies. In fact, only major projects get reviewed. Finally, there is no consistent attempt to evaluate and rank projects by standard criteria, such as benefit/cost ratio or internal rates of return. iO 7.23 Policy Coordination at the State Level. In many states, the coordination role of the COPLADE (see Box 7.2) is not funy utilized and federal and state institutions execute their agricultural programs in a somewhat isolated way, following the pertinent federal or state policies. COPLADEs seem to be more active in informing participating institutions of the different activities in 69. The reallocation of resources via 'compromiBoli may not always be to projects with lower social marginal rates of R return than those for existing projects. Even though some new projects are submitted to project evaluation, in practice it appears that they will be stopped for lack: of economic merit only under very exceptional circumstances. 70. The use of the environmental assessment process all an effective instrument to guide investments in the sector also appears limited. Any (publicly or privately funded) project that involves a change in land use requires an environmental assessment. The environmental assessment is the responsibility of the promoter (the responsible secretariat in the case of public investments) and is reviewed by the National Ecology Institute (lNE) of SEDESOL. The problem is not with the official methodology, but more with its application by potential investors and its review by lNE, which has limited resources to properly evaluate the studies.  124 . .. . . . . . ···········~~r!~Wt~9gJ»ro(:ess • • •• •• • ~~~.W~! . . . . . . . . . . · .· ·.•.·.• •.·. ·•· . ·. .·. . . . . . . . . . i}··. . •.•.• • •.·. . ·• •. . i • • . .............. !~~I~~£~~~~~~!'i\i'~;~j~k~wo}~~~~~f~~~.r~~1i6r~~Yll .Plan. ·wbicb ... Ian .. ....... hasaduratiOnofsix(earsccHncidiriwiththemaridateo{ thePtesident. . ...........................:. . . . . . . . . :. . . . . .:. . . . . :l. . . . . . . . . . . . . . . . . . . . . . . . . . . .:..J~....................................................... .The.. .P :. :..... . .. . .. is . . ...: :........ . •. openlti()nl¥~ .irl~ry¢t~~~¢P~'*4.pplic.y4999fQ¢~1proyidil)g . t}exi~ilityt(). the<pl.~iQg pr9C~i . . ··The.natioqaJpl2#~!~~.tb.~.~QP~4~¥~~§Pm¢pt2P.je¢iivesaItdestablishespii~J.'iHe$by~cdnomie· .sector. TheNatiollarPlanfQf:f9S9~1994estab1ish.edtheforinulati()n of 21 Medillm.TerrriNiltionat. ·:ri~~e~~::t~~.~I~~i';~lr~'~~II:?i~C~fd~;t~h~:~~T:0iO~~~~wl~t:riai.·• • • • •· . Pl~· is publish~!J'!lecUVmJ~p!~!t~ipt~p#f4lpr~entirig . rrioredet(lileQ.·.s~torill·. atld .• r~gi.()~t . . . analysis, objectiy~I1I1<lPQli¢y~g~~~gi,~~ljj~tJ.l>§ta.tl~e.ho\Yever,they •..at"~sti11 general· docuni~~~·.· . . .. devoid of budge~. Orsp~cific¥~lpm~iiij~l1i~9lJJ¢Ciives iJ1·the~ationafa1ldMedhiniTermPl~ are then translatedill.to¢phCl'etepr.p;~~Yi~Qn~Yearinstrumtmts· known.· as Annual .• Operating Programs (programasOperatiV()$A.JiiIiles;-PQi\).The POAs allocate funds for each agency and form the basis oftheanhuatbudge~~dtb.~firiancial agreements between the federal and state governments: theSoCiaIl)ev~I()Pm~tttAgree11i¢ritS(Convenios de Desarrollo Sodar, CDSs).whicb . . are Signed bythefederaI$ecre,tariatsia:ndtheiUovemor ,andprovidetbe legal framework. for· the sector than in proactive coordination.71 Moreover. the closely integrated system of decision making represented by COPLADE is best suited for those situations when there is political agreement at all three levels of government. It appears, based on analysis of the four states analyzed in this report (fabasco, Sonora, San Luis Potosi and Aguascalientes) to be less effective when the federal government, the state government. or the municipal governments have elected officials from different parties. Finally, the lack of a consistent and well-coordinated policy framework for rural .. development policy ma1ces it difficult for COPLADE to carry-out its role of policy coordination. 71. FAO, "Mexico - Review of Agricultural Development Experiences and Potential in the Tropics, • Volume I, Report No: 4/94 CP-MEX.37. 1994, p. 13.  125 . 7.24 The final distribution of state resources by agency and municipality also reflects the phenomenon of compromisos similar to what was observed at the federal level. At the state level. the COPLADE is expected to have a critical role in coordinating joint federal/state investment flows and in allocating these investments to the municipalities according to the state plan (the latter integrates an approved list of proposals and requests from the municipalities and state agencies and local interest groups). In the four sample states it was apparent that the COPLADEs did try to be responsive to the proposals of the municipalities and the plans of the state government departments, but the ultimate decision was clearly controlled by several key players in the COPLADE, and especially by the governor of the state. Furthermore, because of the intertwining of state and federal funds in the CDS and the matching grant nature of Solidarity programs, the compromisos of the federal authorities also have a significant impact on the priorities of fund allocation at the state and local levels. 2. Decentralization or Rural Development Finance 7.25 Decentralization is the devolution of power and responsibilities to lower levels of government in a federal system. An effective decentralization policy increases the responsiveness and accountability of government, leading to a more efficient utilization of resources. Decentralization of public finance is a principal objective of the current six-year National Development Plan, the revised Article 115 of the Constitution, Solidarity, and the new Water Law. The reform of Article 115 in 1983 gave municipalities responsibility for providing water supply and sewage, streets, public lighting, sanitation and slaughterhouses. The 1992 Water Law transfers responsibilities for investments, rehabilitation and maintenance of new irrigation facilities to producers (as well as operation and maintenance responsibilities on existing systems). The Solidarity matching-grant program is also a cornerstone of the Government's decentralization strategy. This section first highlights several general issues that undermine the efficiency of the GOM's decentralization program and then reviews the Solidarity program.  126 7.26 General Issues. This section highlights six issues that undermine possible efficiency gains from Mexico's decentralization policy for the rural sector. First, many rural development services are provided in a mode of shared responsibility and are not clearly assigned to a particular level of government. This leads to the inefficient provision and usually underprovision of these services; as each level of government attempts to free ride on other levels for this particular service in order to free funds for other activities for which they have exclusive responsibility.n The most notable example of this problem is the maintenance of rural roads. Several federal and state agencies finance the construction of new roads, however, the assignment of maintenance and rehabilitation responsibilities for rural roads is often unclear. This has led to insufficient expenditures on road maintenance and the deterioration of many rural roads. 7.27 Second, responsibilities for current services and capital investments are not always assigned to the same level of government. Most of the resources for capital investments at the state and local level are from the federal government, while the financial responsibility to maintain them belongs to the regional governments. Within each state, the COPLADE decides how the majority of the federal investment resources will be allocated (see Box 7.2 and para. 7.24). Although this approach offers some advantages, especially in states with few human resources at the local government level, it is too central ized. More importantly, it is likely to lead to inefficiencies because the capital investment decision and the operation and maintenance decisions of the facilities are taken by different agents with conflicting incentives. 7.28 Third, the assignment of expenditure responsibilities to local levels of government has not always been accompanied by adequate increases in revenue sources. For example, the decentralization of sewage and drinking water to municipalities was not accompanied by increases in municipal revenues. While the Government through the Solidarity Municipal Funds has sought to address this issue, the later only provide a small share of the total amount of resources needed. The lack of revenue sources to effectively implement services decentralized to the state and municipal level is particularly pressing in poorer areas. Poorer states (and municipalities) such as Chiapas, Oaxaca and San Luis Potosi do not have the necessary funds to invest in irrigation, sewage, road maintenance and to develop their management capabilities. 73 Many state governments are reluctant to assume additional responsibilities without an overhaul in the present revenue assignment and tax sharing system. 74 7.29 Fourth, the decentralization process in Mexico has been confused by the fact that some state officials perform a dual role also as representatives of the federal and state governments. For example, the director of the fisheries department in the state administration may be also the 72. The lack of a concrete assignment with shared responsibilities can also lead to geographical disparities and inequities across income groups. For example, the federal government may be more responsive to its shared responsibilities in the more visible urban areas than in rural regions. 73. Some of the wealthier states have been able to assume an increased leadership role in rural development policy planning and implementation. For example. in Aguasea1ientes. the governor created in 1993 CODEAGA (Comisi6n de Desarrollo Agropecuario del Estado de Aguascalientes) to lead the planning for the state sector and to coordinate policies from all state offices involved in rural development to delegate efforts by federal ministries. In Sonora there are two secretariats involved in rural policies (Secretaria de Fomento Ganadero and Secretaria de Fomento Agricola, which have been assuming responsibilities previously exercised by SARH. 74. Concerns about too much centralization of government responsibilities were more frequently raised in the relatively wealthy state of Sonora than in Aguascalientes, San Luis Potosi, and Tabasco.  127 representative of the federal government. This arrangement may not always work well, especially when the executives of the two levels of government belong to different political parties: to whom should the state secretaries respond in case of conflict? the federal government or the state government? 7.30 Fifth, there is very little decentralization in SARH from the central level to its state delegations. Rather, the internal structure of SARH is deconcentrated to the extent that the state delegations carry out programs and expenditures, and the ultimate decision making power remains with the central offices. As a general rule, state delegations have little discretion over how funds are spent and cannot reallocate funds across programs in their budget. But once expenditure ceilings by program are allocated to the state, there is no micro-management by the federal authorities of how these resources are finally put to use. Also, the overall share of SARH resources allocated to its state delegations has declined during the Salinas Administration from 29 percent in 1989 to only 11 percent in 1992." 7.31 Finally, the mandate to channel most investment funds through the COPLADE is stifling to local initiatives and may limit the efficiency gains associated with true decentralization. A principal objective of the COPLADE is to incorporate the municipalities in the allocation of federal and state investment programs. This objective has been only partially fulfilled. COPLADE appears to work as a centralized planning institution within the state and the role of municipal authorities is quite limited. Within the current institutional framework of the COPLADE, it is not clear that the GOM's decentralization policy and in particular, Article 115 of the Constitution, which seeks to strengthen the powers and responsibilities of municipalities, can be fully implemented. 7.32 Solidarity. The entire program of Solidarity is an important element in Salinas' decentralization policy. Innovative design techniques, involving the community at all steps of the investment process increase the efficiency of expenditure programs in Mexico. The community also participates in covering the cost of the programs through either labor, materials or, less frequently, cash. As a result of the high level of community involvement, the Solidarity program is more responsive to needs of the people and its projects have been completed at lower costs. Building costs appear to have decreased, and even the maintenance of the facilities constructed with Solidarity funds have had a much better record.'6 Furthermore, in the four sample states, the perception of the SoJidarity programs was indeed very positive and is viewed by local officials as the most favorable fiscal development for the implementation of Article 115 in the Constitution. The different components of the Solidarity program are commonly seen by municipal officials and the community interviewed as the only chance to launch infrastructure projects for which until now they had neither the ability nor the adequate assistance from the state government to develop. 7.33 However, a principal drawback of the program is that it retains an excessively high political partisan profile and lacks transparency. The most significant fact in the allocation of Solidarity funds is that there is no formula or pre-established eligibility rules that would ensure that resources flow to 75. The share of state delegation expenditures rises when the budgets of SARR's dependent agencies are excluded: In a SARH rFederalizaci6n y Reestructuraci6n de 1a Secretaria de Agricultura y Recursos Hidraulicos· (1992», the decentralized expenditure is stated at 66% while the central administration of SARR spends the other 34%. 76. Many Solidarity projects have been fmished under budget. There is an incentive for the community to see that resources are economized because any unspent surplus from the project stays in the community to use in similar areas.  128 those areas and projects that are most in need of the resources and will use them most efficiently.71 Instead, the allocation is decided on an ongoing basis and the final allocation is driven by proposals submitted by the states and the communities. Within the expenditure ceilings for Solidarity's programs at the state level (which are determined by the federal government) COPLADE decides the particular uses and the municipalities or communities that will receive the funds. 7B The delegation of the federal SEDESOL and the state social development agency work closely in this committee and the suspicion in the field is that the funding decisions are personal and depend very much on the political and institutional connections of individuals. 3. Multi-Year Capital Budgeting 7.34 Mexico lacks a long-term budget framework. While the national and medium-term plans are, at least formally. integrated into the annual budget process, these plans are primarily descriptive and do not provide a long-range budgeting tool. Given the long run nature of most investment projects and the interplay of capital investments with fiscal policy. investment planning should be framed in a period longer than the annual budget cycle to permit planning for recurrent (operations and maintenance) and capital budgets and to facilitate fiscal policy. During the 1980s, maintenance and rehabilitation expenditures declined even faster than new investments, and by 1987 were 13 percent of their 1982 level. 19 A mUlti-year capital budget framework would have highlighted more easily the scarcity of funds allocated to maintenance and rehabilitation. The lack of appropriate funding for maintenance of capital infrastructure also reflects the current budgetary system. which itemizes the different types of expenditures carried out by government in "chapters" does not separately identify maintenance expenditures. These are sometimes included in the chapter of "general services" and other times in the capital budget. The lack of a clear classification makes it difficult to track the actual funds allocated to maintenance in the budget. 4. The Budget Process 7.35 Each of the three stages of the budget process (preparation, execution, and control and audit) can have a significant impact on the overall efficiency of expenditure policy. This section reviews these three stages and identifies several issues which hamper the overall efficiency of expenditure policy. 7.36 Budget Preparation Stage. The budget formulation process is similar for all secretariats. Even though each level of government has autonomy in formulating its current budgets, Mexico has a quite centralized budgeting system. The preparation of budget proposals for the following year starts in August after the SHCP provides each secretariat with its spending ceiling. There are two issues in the budget preparation process which merit attention. First, SHCP may introduce extraordinary budgets once the regular budget approval cycle has been closed, which provides the finance secretariat with more budgetary leeway than is conventionally the case. The Decree approving the Expenditure Budget for 1993 (Article 14) authorizes SHCP to incur additional expenditures for different purposes. 77. The exception is the four states in the Regional Development and Decentralization I Project where specific criteria are used to transfer resources to the municipalities. 78. Actually, in the state of Tabasco there exists now a committee of Coordination of Solidarity, which previously was part of COPLADE, but now operates independently of it. 79. World Bank, "Mexico in Transition: Towards a New Role for the Public Sector," No. 877O-ME (1991).  129 including investments in the social and rural sectors. These extraordinary expenditures, however, can only be funded out of surpluses generated by current budget operations, by public enterprises, or by privatization activities. The only requirement is that these activities be reported at the end of the year in the public expenditure accounts from the Executive to Congress (the Cuenta Publica). In the majority of democratic countries, unless there is a national emergency, additional expenditures require previous approval of congress. 7.37 Second, there is an especially limited role for the municipalities in the capital budgeting process. The process is bottom-up in the preparation stage, but at the decision stage the process is very definitely top-down. For example, in the case of Solidarity, the neighborhood committees present proposals to the municipal government, which in most instances, passes them on to COPLADE and the state SEDESOL. The state office passes the proposals to the federal delegation of SEDESOL, which in turn consolidates and sends them as preliminary budget requests to the federal SEDESOL. After the SHCP establishes budgetary ceilings by secretariat, the process to allocate the funds starts from the top down through the same institutions until the funds arrive at the municipality for one of the neighborhood committees. As a result, the final budgetary allocation transferred to the municipality may not reflect the most efficient and/or desirable use as defined by the local community. 7.38 Budget Execution Stage. The conventional tradeoff between expenditure control and flexibility of operations in the process of budget execution is clearly solved in Mexico in favor of expenditure control. There is in fact very minimum flexibility in the budget execution process. Secretariats and agencies require prior approval from the SHCP for the transfer of funds across programs and even across expenditure categories within programs. As a result, resources assigned to a particular program or unit are often used for other purposes, without officially changing expenditure records. This makes it difficult to evaluate the true cost of programs in the sector. The exception is the Solidarity program, where there is significant flexibility to reassign resources within the overall Chapter XXVI category without prior approval of SHCP. At the municipal level, there is also little discretion as all investment projects for the municipalities have to be approved by the state congress. 7.39 Evaluation and Control Stage. The budget control stage seeks to monitor and evaluate performance. The main instrument is the "Sistema de Control y Evaluaci6n Gubernamental" carried out in coordination with the Controloria General de la Federaci6n and the SHCP. This control stage produces several documents such as the Quarterly Reviews of the Economic and Social Situation, the results of internal audits, Government Reports, Report on the Execution of the National Development Plan, and Cuenta Publica. This is more than adequate information on the monitoring of financial implementation. Where the system is weakest is in the critical evaluation of the programs and reporting that information into the next period budgeting process. B. Public Expenditures and Poverty Alleviation 7.40 Prior to the mid-1970's, rural development expenditures in Mexico were oriented towards the northern states, reflecting the prevailing belief that large-scale commercial agriculture was more productive than small farming. Large capital intensive expenditures in irrigation were accompanied by public infrastructure investments in roads, electricity and water supply systems. Similarly, research and extension expenditures by SARH were oriented towards large-scale, commercial agriculture. In the late 1970s and early 1980s, a decreased agricultural GDP growth rate, coupled with increasing demands from the small farm and ejido sectors, catalyzed a shift in GOM rural expenditure policy. The GOM initiated small scale irrigation and drainage investments in rainfed areas and financed  130 regional area development programs in the poorer southern states. In addition, it expanded producer subsidies on inputs (fertilizer, insurance, credit, seeds, etc) and outputs (grains and oilseeds) and initiated broad-based consumer subsidies. 7 Al This section addresses, in a limited manner, the extent to which expenditures of certain key rural development programs, adopted in the early and mid 1980s, have been directed towards low income rural producers. In particular, the section reviews expenditures by SARH, the DFIs, CONASUPO, as well as the revenue sharing and Solidarity programs to assess if they reach the rural poor. A comprehensive evaluation of the impact of each program on productive opportunities for low income rural families is beyond the scope of this study. While the analysis of these programs is not complete, it does suggest that many of the investment and subsidy programs have not effectively reached low income producers. 1. Revenue Sharing 7042 A principal instrument used by the GOM to support its decentralization policy and transfer resources to the states is the revenue sharing program. Federal revenue sharing provides 82 percent (1992) and 48 percent (1990) of the total revenues for states and municipalities, respectively (see Box 7.3). As the Government continues to decentralize rural development responsibilities from the federal government to state and municipal levels, poorer localities will become increasing dependent on revenue sharing to provide quality services and investments than their wealthier counterparts. However, the overall level of poverty in a state has little effect on the amount of revenue sharing it receives. 7043 Regression analysis was used to determine the influence of population, the state's level of poverty, and per capita GNP on the amount of revenue sharing received for all 31 states between 1990 and 1992.80 The results reveal that the state's level of poverty had a negative impact on the program's General Fund allocations it received during all three years (annex 3). The state's population was positively related to the amount of revenue sharing received, reflecting the fact that 45.17 percent of the General Fund is distributed according to population. The analysis revealed no significant relationship between GNP capita and the distribution of the General Fund by state. The regressivity of the General Fund reflects the fact that 45.17 percent of the fund is distributed according to state collections of several taxes, mainly the impuesto predial and water charges, which are positively correlated with income. Population was the only significant factor behind the distribution of the Municipal Development Fund in 1992 (annex 3). Per capita GNP/stae as well as the states's CONAPO marginality index did not affect the amount of Municipal Development Fund resources it received. 7044 The states are then responsible for transmitting 20 percent of the General Fund and the Municipal Development Fund to the municipalities. The progressivity of revenue sharing distribution to municipalities was evaluated in the three states of Tabasco, Sonora and Aguascalientes. In general, the distribution of revenue sharing funds to the municipalities is not correlated with the relative poverty level (as indicated by the CONAPO index) of the municipality. The only exception is for Tabasco where relatively poor municipalities appear to have received less revenue sharing funds in 1992. The lack of statistical significance for the margination index in Sonora and Aguascalientes suggest that the distribution of funds may be neutral with respect to the municipalities' poverty level (annex 3). 80. The CONAPE (Consejo Nacional de Poblaci6n) marginality index is calculated with data from the 1990 Census based on a series of infrastructure and income indicators.  131 '.:i .'wllii,;~'i~\i'~i'll~,~!lIrl~~:>.. ,. . . ,. . . ·······/·;:R~ye#qf:~arhlg ·1953, but the reVeIltiesll]bc~t@ffl~t~AA4Q1@icipalitiesunder this systel1l\\ferebase4<q~the.( ~~::~ ·°-in~~~s~~~~~ed~~~I~~18\Va.~• • f;t~.·i~u~:n~~0·f~:~~~t~n:~i:i;~ Nacional de .Coordindq~nlftsc41($N¢f'1~·W~i9hi$the. curr~nt legal foupdatiCln.fClrJ;evenu~ s~at4ig . .· ·. .in Mexieo .. ··{Jriderp.t~~~Cf;~~~~~~,m~£iJ>~ides .r~eivea cert~lf).p~r~~9ta~e(~~I~} f:~~ft::1~~~J~~~~t~i4fr~4~~i~~~~i4~r~~~et;tft.~~;efu~j1W~~~d~~it (11np~e~t.Q#Iy:t#'6t4~t~i(#@~9rl¥Ajg~.~9rf~efiitX~"~ev~nue)<iihda$pecl31Ta;(Q~\ Productionan,(f$erVi~~·9tseie.~~il~(2(j%9fFe4er3:11~es)j . ." .... '. '.. .....' . . . .. . .. ilie stu~:o.~~;m,lil:tSl'~~r~~'g~.E~~~i*~~~~~ advantages and (U~adY~f.igeSforthe~tllte§}1l.ieStateStPatbecome part ·oftheS,NGF areentidetl~ci a larger share ()tfe<teriUfurids.that'l~~Y\W()41dpeotherwise, but they are .require<tto distd.9Mt~~t Ieast 22; 1.per¢erit 6f~e.se.fullds3gt()ijgth~irl11uI1icipal ities~ .•. Although the biggesfdisadYarit~¢iS. that states los~<tl:1eirtax>bas¢;.aJlstat~baye~ignoothese agreements.' .' .. ..... ..... '. .. 'The .am()Mnt .• ot··f~er~· a11()~~fi~~;• ~~2~~ciones) •. ···that·· each· .•.state.• •• ret¢lY~ is····deterrtii~~ • •· • •· by •a formula furUteFa~9(fis~hl·~Qf4iIi@0~(Ley .deO:{ordinaci6n' Fis~Qj;orLCf)~.Ii.·. revenues ·collectedhfth~)hit~~g()¢,M¢~~~\~ity·• andforin .tb.e.··~ecaudq(;id,",<·Ee"eral1f(lfticipa1JJe.·.·. . ~:;iy.o~;~~~~·····~'·'~:• •~e~:~J~I~~~~taiih~t~u~ •.r~~~enb~al~~~~~~S:~~sl0~~• • • ·• municipalitieS.lnad~idon,theMW~~§ipMl?~yelopmentFund is made up of OA2percento(the< RecaudacMn FederalP{lniCip(lbleiindm~st~epassed on in its entirety to the municipalities.1busi/ in total, the statesreceive18~$perce#tof<~efederally collected revenue, 22;1' percent ()fwbjch they must distribute to their resPectiv~mtlriicipalities. .. . ... Sourc.e: ." ·.•··Setgei·.• •·.Soares ••..•.and • • •Andrea·.·.···.~ily¢r[I,~p)• •.••.. ·vict()ri~• ·• ·R~\i;igU~·.·• .• (~~;• • cQlIlbbr~tion•.• M "Intergoverilm~rijlR,elatiohsandfin~hilde:dco .• Focus on. Eight .pfoJ~ct .Siates... S~cqrid.···· .Decentral~ti<iaimd~egiollalp~V~12p~~~r~r()lt# .W~r~!ngPal'er· ·1.AllriI1994? <i< /.> ' .. No." 2. Solidarity 7.45 The National Solidarity Program is the Government's principal strategy to deal with rural and urban poverty. Again, a detailed examination of the Solidarity program is beyond the scope of the study, and the analysis focusses on the geographic distribution of Solidarity funds. While the overall distribution of Solidarity Funds by state reflects poverty criteria, the poverty focus of specific Solidarity programs, such as Production Funds could be improved. The poverty level did not appear  132 to be a significant determining factor for the distribution of Solidarity funds among municipalities within a given state. 7.46 The distribution of Solidarity Funds for all 31 states was analyzed using regression analysis: the state's allocation of Solidarity Funds was regressed on the states' per capita income, the marginality/poverty index and population (annex 3). The results suggest that poorer (as defined by the CONAPO marginality index_ and more populated states receive more Solidarity Funds than other states. The coefficient for GNP/capita was not significant. Within Solidarity, the distribution of Municipal Fund resources by state also appears to favor poorer states.'1 Again, both population and the poverty index are positively correlated to the amount of Municipal Funds received by each state. The distribution of Solidarity Production Funds does not seem to be heavily correlated with the state's poverty level. The distribution Production Funds favors more populated states and those with a large rural population (considered as towns with less than 5,000 habitants). However, there was no significant relationship between a state's poverty level and the distribution of Production Funds. The share of the population earning less than two minimum wages and per capita GNP also have no significant impact on the distribution of Production Funds by state. 82 7.47 Solidarity funds are distributed by the state COPLADE and SEDESOL to municipalities. The distribution of Solidarity Funds to municipalities was evaluated in Tabasco, Sonora and Aguascalientes. In the regressions, none of the explanatory variables, including the municipality margination index, were statistically significant. These results mean that the progressivity or equalizing effect of the Solidarity programs is diluted within the states.13 3. Distribution of Marketing (CONASUPO) and Credit (DFIs) Subsidies 7.48 A majority of CONASUPO's expenditures for marketing subsidies have been directed towards wealthy farmers, mainly because they have the largest surpluses to market. For example, 8 percent of the corn producers who cultivate 5 ha or more of corn, received 59 percent of the corn subsidies during the Spring-Summer 1991 cycle. Prior to 1993, CONASUPO purchased on average only 30 percent of the marketed corn crop. However, it has not focussed purchasing activities on the poorer states. In 1992, CONASUPO purchased brought 99 percent of the total output from the wealthier states of Sinaloa and Chihuahua compared to an average of 23 percent from the poorer states of Chiapas, Puebla and Guerrero. Under the PROCAMPO program, CONASUPO's role in the sector will be reduced. The phased elimination of the guarantee price system for maize and beans in exchange for direct income payments to producers under PROCAMPO will extend subsidies to farmers who previously did not have a market surplus. Nonetheless, the program will not have a 81. The Municipal Funds were created in 1990 to transfer additional resources to the poorest municipalities in the country and to strengthen their flllancial and administrative capacities. In 1992, they accounted for 14% of Solidarity'. rural (excluding social sectors) budget. 82. Solidarity Production Funds (Fondos de Solidaridad para la Producci6n or Credito a la Palabra) were created in 1990 to assist fanners not covered by BANRURAL, estimated at approximately 269,000 producers with 1.2 million hectares. The main purpose of the Funds is to support the cultivation of basic products (mainly com and beans) in areas without irrigation and/or low productivity. 83. The issue of equity or equalization with in the state is important to the extent that there are significant differences in the standard of living across municipalities in that state. The variation in the margination index within the state is a good indication that these differences do exist.  133 significant impact on rural poverty as farmers with greater hectarage will receive the bulk of the income payments given the program's current structure. 7.49 Similarly, only a small share of FIRA lending has been directed to the 80 percent of Mexico's agricultural producers considered to be low-income producers (PBIs) by the GOM.84 Between 1982 and 1992, only 18 percent of FIRA's borrowers were low income producers (PBIs), which comprise 80 percent of Mexico's farmers. During this time, FlRA accounted for 34 percent of the real value of all transfers to the rural financial system, however, In contrast, almost all of BANRURAL's loans (95 percent) were to low income producers. BANRURAL was the largest source of GOM rural credit from 1983 to 1989. 4.SARH 7.50 Given the available data, there are two general points to be made concerning the impact of SARH/CNA expenditures on rural poverty. First, irrigation expenditures, which accounted for 82 percent of CNA's budget in 1992 are focussed on irrigated areas, where the majority of the wealthier farmers reside. In 1985, SARH initiated a program to expand small-scale irrigation in rainfed areas, where the majority of low-income farmers are located. However, resources to the program have declined from the initial 15 percent in 1985 to less than 5 percent since 1988. In contrast, the development of irrigated areas consumed 56 percent of CNA's water program budget in 1992 (down from an average of 70 percent during the de la Madrid Administration). Moreover, 50 percent of the expenditures in irrigated areas are for rehabilitation, operations and maintenance, whereas these costs are covered by users in rain fed areas. 7.51 The second point concerning the impact of SARH expenditures on rural poverty is the distribution on state delegation expenditures. State delegations spent 14 percent of the total SARH budget in 1992. Regression results indicated that SARH delegations budgets are higher in poorer states (ranked by the CONAPE poverty index) (annex 3). Delegation expenditures also rise with population and with increase in per capital income, but the poverty ranking is the largest factor. This suggest that in states with weaker state rural development institutions, SARH partly compensates with a larger presence. as III. Fiscal System and Incentives for Rural Development Investments A. Introduction 7.52 This section reviews six types of taxes and evaluates their implications for private investment in the rural sector. The agricultural sector in Mexico has traditionally received favorable treatment in the tax laws, which has important implications for the allocation of resources in the economy and the equitable treatment of agriculture vis ~ vis other sectors. The fiscal reforms undertaken by the Salinas administration have changed the tax regime for agricultural activities, but the present tax structure still grants special status to the sector. 84. PSIs are producers with net annual family income not exceeding 1,000 times the annual minimum daily regional wage. 85. In the poorer state of San Luis Potosi, the Gabinete Agropecuario is supposed to be directed by the state's secretary of agriculture, but in fact. is led by the federal delegate of SARH. Moreover, some of the functions previously carried by SARH in the state such as extension services have not been taken over by the state government because of the lack of funds.  134 7.53 To some extent, the special treatment granted to agricultural activities in Mexico has been based purely on tax administration reasons. Because of the small scale and lack of development in many farming activities, it is too costly and almost impossible to bring these activities into the tax net on a conventional basis. From this perspective, some of the special treatment granted to farmers has not been different from the treatment granted to other activities such as street vending or small enterprises. Tax administration and operational considerations have been the overriding motivation for the simplified income tax regime for small businesses, which includes farming activities, adopted in the 1988-89 tax reform. 7.54 Nonetheless, a simplified tax administration system has not been the only or the most important objective in the tax treatment of the agriculture sector. Special treatment granted to agricultural activities by the present tax laws have been designed to give this sector of the economy a privileged treatment (in terms of output and specific inputs of production) compared to other sectors in the economy. For example, in the Value Added Tax (V AT) administrative considerations dictate the adoption of a single tax rate that includes agricultural activities. Instead, most agricultural activities and providers of intermediate products to the agricultural sector were granted a zero rate in the most recent reforms of the V AT. These measures have created serious administrative problems, increasing the complexity of the system and opening new avenues for tax evasion. 7.55 The non-neutral tax treatment of agricultural activities affects the overall efficiency of the allocation of resources in the economy. The potentially serious misallocation of resources takes place in different ways: too much labor and capital may be employed in this sector and too little in other sectors of the economy where their return would be higher. Special deductions or tax credits granted to capital investments in the agriculture section can also distort input ratios toward excessively capital intensive techniques of production, discouraging the use of abundant and underemployed labor. As is discussed below, the special tax credit for capital investments granted at present in the Asset Tax discourages the use of labor. 7.56 However, the special tax treatment granted to the agricultural sector as a whol e, and not just to individual factors of production, may sometimes be considered necessary in order to offset unfavorable treatment to the sector in other areas of economic policy. In particular, in many developing countries the agricultural sector has been penalized (implicitly taxed) by an external tariff that is overprotective of manufactured and semi-processed goods compared to unprocessed and agricultural goods. Often it has also been the case that an overvalued currency penalizes the export of agricultural products while it reduces the cost for imports of inputs in manufacturing activities aimed at domestic markets. In the past, international trade and foreign exchange regimes also gave unfavorable treatment of the agricultural sector. However, the recent reforms in Mexico's international trade, which eliminated most trade barriers and drastically reduced the average level of customs tariffs (specifically exempting agricultural inputs) has significantly reduced the penalization of agriculture. As a result, there presently is little justification for an across-the-board favorable tax treatment of agriculture. 7.57 If correction of distortions introduced by economic policies against the agricultural sector is not achieved, what then may justify the very generous treatment granted to the agricultural sector in the current tax laws? It could be argued that the fiscal privileges to the sector are justified on income redistribution grounds because a large share of the Mexican population still lives in poverty and a high percentage of this segment of the population is in the rural agricultural sector. However, there are other policies, such as the Solidarity program, that can be targeted to the rural poor and be more effective in combating poverty than a blanket favorable tax treatment which also benefits rich farmers and enterprises. Targeted policies are also less distortionary regarding resource allocation in the  135 economy. The irony is that some of the current tax policies encouraging capital intensive techniques of production in agriculture are not only distortionary but also anti-poor because they reduce employment possibilities in agriculture. 7.58 Special tax treatment, of course, can be justified in circumstances where significant externalities are present and taxes are used to reduce these externalities by facilitating their internalization in decision making. For example, the tax laws can be used to give incentives to reduce chemical pollution of water sources, to encourage soil conservation practices, or to promote the preservation of forests. As we see below, externalities have not been and are not currently a motivation for special regimes in Mexico's tax laws. B. The Income Tax 7.59 Before the tax reform of 1988-89, the agricultural sector was subject to a very preferential income tax treatment. The agriculture sector shared this privileged position with other activities, most notably the transportation sector. Under this preferential regime, known as the "Special Basis for Taxation," taxpayers paid a fixed amount of tax which was not related to their activity. In practice this special treatment represented almost the complete exemption of these businesses from income tax. 7 .60 The 1988-89 tax reform abolished the "special regime." To ease agricultural activities into the complex general income tax, the reform made them subject to a "simplified" income tax. This simplified tax, which also applies to the whole transport sector and to small businesses in the rest of the economy is actually a consumption tax based on business cash flow. The base of the simplified income tax is computed as the difference between cash inflows and cash outflows. Cash inflows include regular revenues and also proceeds from borrowed funds, interest charged, withdrawals from bank accounts and the sales of securities other than shares in the business itself. Outflows allowed include all operation expenses, the purchase of capital goods, the purchase of securities and deposits made in bank accounts. The simplified consumption tax offers several advantages to taxpayers such as a simplified accounting system and exemption from complex timing issues. In particular, all capital expenditures are immediately expended rather than being subject to depreciation rules. Although a lengthy discussion of the features of the cash flow consumption tax compared to a regular income tax is not appropriate here, it must be stated that there is a lot to commend the cash flow consumption tax (although no consensus has been reached on the superiority of either tax). It does deal with important efficiency losses associated with a traditional income tax. Nonetheless, the cash flow tax offers a much more favorable treatment of income from real capital investments than does the traditional income tax. This means that the entire agriculture sector is induced to use more capital intensive techniques of production than other sectors of the economy that are subject to the regular income tax. Given the importance of employment issues in the rural sector, the desirability of this is questionable. 7.61 Another important consideration is that the simplified income tax applies to all agricultural income derived by individuals or corporations regardless of their level of earning. The auto-transport is the only other sector in the economy that enjoys, with some restrictions, these same benefits. In all other sectors of the economy, only individuals with business revenues and interest income under 600 million pesos may opt to use the simplified income tax. 7.62 Furthermore, individuals and corporations involved in agricultural activities enjoy additional privileges (within the simplified tax regime) only applicable to them:  136 (a) Individuals may take a special deduction from taxable revenues calculated as the share of total revenues represented by 20 minimum wages on average in the past five years. Corporations can use a similar deduction for each shareholder up to a maximum of 10 shareholders or 200 minimum wages. (b) Individuals and corporations exclusively dedicated to agricultural activities may reduce their taxes due by 50 percent. The reduction falls to 25 percent if the business processes some of its products or if the firm conducts industrial or commercial activities from which it derives up to 50 percent of gross receipts. (c) Several categories of agriculture workers are exempt from the income tax: ejidos and comuneros, associations of ejidos and comuneros, other groups and associations for farming and cattle-raising, and all farmers with annual revenues under 20 minimum wages. (d) Other special treatment granted to the agricultural sector includes the exemption from taxable income of the proceeds from the sale of land titles by ejidatarios the first time the land is sold, and accelerated depreciation for those using equipment to provide services to farmers. For example, airplanes used for agricultural purposes can be depreciated at a maximum annual rate of 25 percent while for other uses this rate is 10 percent. 7.63 The Mexican income tax grants special treatment to agricultural businesses. Some of these measures are justifiable on purely administrative grounds. The simplified regime is a good way to bring into the tax net small farming activities which otherwise (under the general income tax) will pay virtually no taxes. Similarly, several of the other tax advantages to the sector, such as the exemption for ejidatarios and comuneros, may also be justified on equity grounds. However, the application of a general rule for minimum exempt income could accomplish the same on equity grounds and with savings in foregone revenues for government since wealthier ejidatarios would not be exempt. This alternative policy would also yield greater equity with respect to taxpayers in other sectors of the economy. Finally, other tax advantages, most importantly the reduction by 50 percent of all taxes for businesses exclusively involved in agricultural activities, is an outright fiscal privilege favoring the sector, regardless of size or even geographical area, vis a vis other sectors of the economy. c. The Gross Asset Tax 7.64 The tax reform of 1988-89 introduced a gross asset tax for all individual businesses and corporations of 2 percent. Agricultural businesses are also preferentially treated under the gross asset tax. In particular, in coordination with their treatment under the income tax, taxpayers that are exempt from income taxes, such as ejidatarios and comuneros, are also exempt from the asset tax. 7.65 Those taxpayers exclusively involved in agricultural activities (still within the broad definition of agricultural activities) pay the tax at a reduced rate of only 1.0 percent which is half the normal rate. Furthermore, these taxpayers compute their tax base by including land, valued at its assessed level for property tax purposes, and by excluding all other assets such as machinery and other capital equipment. In the general case, such assets are included in the tax base. Several other tax advantages are granted to the sector under the asset tax. Individual taxpayers exclusively involved in agriculture can deduct 15 times the minimum annual salary from the tax base. In addition, all capital investments receive a 100 percent credit against the taxes due.  137 7.66 The reduced tax rate and the exclusion of all assets aside from land tends to benefit richer farmers. The exclusion of capital equipment from the tax base and the 100 percent tax credit for all types of investments also introduces a bias against the use of labor in the production process. The impact on employment may not be as large, however, because of the small significance of the tax. Nevertheless, the asset tax should be reformed to provide a more equitable policy across all economic sectors. D. The Value Added Tax 7.67 The agriculture sector is very generously treated under the VAT. Most agricultural products are zero rated. This means that in addition to not paying the tax (Le., being exempt), producers are entitled to a credit for the tax paid on inputs of production. Zero rated products include all non- processed vegetable and animal products except india-rubber. Most basic food stuffs are also zero rated. In addition, the sale or rental of tractors, other agricultural equipment, fertilizers, fungicides, herbicides, are zero rated when used in agricultural production. Many services rendered to farmers are also zero rated including irrigation networks and the digging of wells, construction of electric grids for water pumping, fumigation, and the preparation of land. 7.68 Farmers and agriculture businesses that are zero rated need to maintain accounts and file tax returns. However, individual taxpayers who do not process their products can opt for an exempt status. In this case, farmers are entirely freed from filing requirements for the VAT, though, of course, they forego their rights to receive any credits. In order to qualify for exempt status farmers' annual revenues cannot exceed 77 times the minimum wage (of the taxpayer's geographical area) and their assets cannot exceed 15 times that minimum wage. 7.69 Finally, the sale and transfer of land is exempt from VAT, although these transactions can be taxed by the state governments. Actually most states do tax the transfer of land ownership, under the .. impuesto sobre traslaci6n de dominio de bienes inmuebles which in most states is a municipal tax. It 7.70 The extensive use of zero rated status granted to the agriculture sector in Mexico's value added tax law is unprecedented in international practice. The extensive use of zero ratings is largely undesirable on administrative, efficiency and equity grounds. The zero rating of so many activities complicates tax administration considerably and promotes opportunities for tax evasion and avoidance. The fiscal privilege granted to the sector leads quite likely to the misallocation of resources in the economy. On equity grounds, there are no good reasons to give this favorable treatment to all enterprises regardless of size in the agriculture sector. 7.71 A desirable agenda for VAT reform in Mexico would move the agriculture sector within the general regime, and substitute the zero rate with the general rate. The zero rate treatment should apply only to agricultural exports, as it does now to any other type of export under the general VAT regime. For administrative reasons, sman businesses in agriculture, like any other small businesses under the general regime, should be exempt from VAT. Equity considerations can justify the exemption, but not the zero rating, of those basic food stuffs which may have a bigger impact on the budgets of poor households. This is a problematic area in value added taxation everywhere. Many countries, however, have opted for taxing all goods at a uniform rate and addressing income redistribution concerns via targeted subsidies and other means. The Mexican Government should also consider both options.  138 E. Property Taxation 7.72 There is no federal tax on land or any other type of real estate property other than the asset tax. However, there are several municipal and state taxes falling on the ownership of property. The most important of these taxes is the property tax or Impuesto Predial. This is a municipal tax but in many states is mainly administered at the state level with revenues transferred afterwards to the municipalities. In most cases, the Impuesto Predial taxes both structures and land, including all types of land use (agriculture or forestry). The tax is levied on assessed values recorded in an official cadastre, in most states kept and updated by the state tax administration. The role of municipalities varies from simply hand..cJelivering tax bills to the entire administration of the tax including the cadastre. For example, in Baja California, the Impuesto Predial is entirely administered by the municipalities. The authority to change both rates and assessed values resides in the state legislatures, although some state constitutions and legislatures give their municipalities some discretion in this area. 7.73 The tax rates for the Impuesto Predial differ across states and within a state; the millage rate is different for urban and rural land. In some states, for example Tabasco, the millage rate on rural land is significantly higher than the rate on urban properties. The assessed values of agricultural land (predios rurales) often differs according to type of land use. For example, in the state of Sonora agricultural land is assessed at decreasing values depending on whether the land is irrigated by simple gravity, a pumping system, or rain; whether it is forest land or summer pasture land. States periodically update, usually by state law, the assessed values in the cadastre, but these values generally lag considerably behind actual market values. In some states assessed values are just a very small fraction of market values. Only the Federal District and the states of Mexico and Nuevo Leon have updated their cadastre to what can be considered market values. 7.74 Besides the Impuesto Predial there are several other minor taxes affecting land ownership. There is a tax on the purchase and transfer of real estate and there are also several fees charged for the right to the reallotment of land parcels. Typically these fees are lower for rural lots. 7.75 The tax burden falling on property and in particular on land remains quite low in Mexico. This probably means that taxation so far has had little impact on land use patterns or on the choice of technologies for agricultural production. However, not far in the future, local governments in Mexico will have to rely much more heavily on property taxation to finance services. This is a source of revenue that needs to be even more encouraged by federal policies. Even though the property tax tends to be sometimes a particularly unpopular tax and not simple to administer, it lacks many of the undesirable efficiency effects of other taxes. F. Social Security and Payroll Taxes 7.76 The coverage of Mexico's social security system could, according to the laws, be extended to many agricultural activities but, de facto, very few workers in the sector are protected. This makes labor input in agriculture cheaper relative to labor used elsewhere in the economy. especially considering its abundant supply to the sector. In addition to social security taxes, there is also a payroll tax exercised by some state governments, the Impuesto sobre nominas. However, this tax does not affect agricultural activities unless they include factory processing of agricultural outputs. 7.77 The Law of Social Security establishes a compulsory system and a voluntary system. The voluntary system allows workers that are not compulsorily covered to join voluntarily on their own after complying with some requirements. The compulsory system includes health, disability and old  139 age insurance. The Law explicitly establishes the possibility of coverage for: ejidatarios, comuneros and farmers organized in Solidarity groups; farmers incorporated or organized in credit groups; and ejidatarios and comuneros organized for the exploitation of forestry and industrial goods as well as commercial purposes. In addition, the Law also extends potential coverage for small owners with more than 20 hectares of land. However, the effective inclusion into the system of these workers is regulated by decree and very few of these categories of workers have been provided with coverage at present time. 7.78 The Law of Social Security (Article 16) establishes the possibility of extending coverage by the compulsory system to all salaried workers in the agricultural sector as well as to ejidatarios, comuneros and small owners otherwise not covered by executive order. They would be included in the system by geographical area, and coverage would be according to the type of services delivered and contributions. The extended coverage would take place by government decree. So far it appears that this avenue provided in the Social Security Law has not been used. 7.79 The overall lack of coverage of workers in the agriculture sector, places workers at a disadvantage relative to workers in other sectors of the economy and particularly to those in the government sector who are fully covered by the social security system. From a financial viewpoint, the slow incorporation of rural workers into the social security system has helped to prevent huge deficits for the system. The general principle to be followed in the extension of benefits is to tie those benefits very closely to required contributions. However, a strict match of benefits and contributions will not always be possible and it may not be totally desirable if some redistribution or cross-subsidization is intended. . 7.80 From an employment viewpoint, the lack of payroll contributions by employers in the agriculture sector lowers the relative price of labor services, which may offset the pro-capital bias in the current tax laws. The employment effects of the extension of social security into the agriculture sector and its impact on the actuarial soundness of the system need to be carefully evaluated. IV. Recommendations and Conclusions A. Trends in Rural Development Finance 7.81 Reflecting overall reductions in the federal budget, rural development expenditures declined significantly during the de la Madrid and Salinas Administrations. They receded by 38 percent during the de la Madrid Administration and decreased by 11 percent between 1989-92. By 1992, rural development expenditures were N$20.2 bi1lion (US$6.6 billion), compared to N$35.2 billion (USS11.7 billion) in 1983. Despite these expenditure reductions, the relative level of agriculture spending in Mexico remains higher than in other developing and industrial countries. The share of agriculture expenditures of agriculture GOP in Mexico was 39 percent compared to 21 percent and 12 percent for the industrial and developing countries. There have also been significant changes in the composition of the rural development budget. During the first four years of the Salinas Administration, the share of investments rose from 25 percent in 1989 to 45 percent in 1992, reflecting declining credit, fertilizer and marketing subsidies and increasing investment expenditures by Solidarity and to a lesser extent by SARR, SRA, SCT and SEPESCA. 7.82 The PROCAMPO program, initiated in 1993, risks to substantially increase the rural development budget and shift its composition towards subsidies and recurrent costs. In 1994, the GOM estimates that PROCAMPO would cost an additional N$6.1 billion (US$2 billion) over the  140 1992 costs of ASERCA and CONASUPO.86 This would increase the rural development budget by 28 percent and the recurrent cost budget by more than 50 percent, assuming other costs remain at their 1992 level. The cost of PROCAMPO is excessive and must be reduced. Even if macroeconomic conditions permit a transfer of resources from other sectors to support PROCAMPO, the GOM should not do this, as there are already more than enough resources allocated to the sector. Nor should the GOM reduce expenditures for public goods (which were N$7.1 billion in 1992) as this would diminish the ability of the sector to adapt to the changing economic conditions brought on by PROCAMPO and NAFfA. Chapter 2 presents a strategy to significantly reduce the costs of PROCAMPO. 7.83 On a more technical level, there is a serious problem of information. It is often unclear where money is allocated in the rural sector. Conflicting data appears within the Cuenta Publica and between the Cuenta Publica and the information provided directly by the secretariats, particularly with respect to Solidarity and program expenditure data for SARH, SRA and SEPESCA. The Government should ensure consistency between the expenditure data that appears in the Cuenta Publica and the data maintained by the functional secretariats. 7.84 Despite efforts to decentralize expenditures to state and municipal governments, the large majority of public expenditures in Mexico are made by the federal government. In 1989, the combined budgets of the states was 13 percent of the total federal budget, while municipal expenditures were 3 percent of the total federal budget. The decentralization of rural development activities and expenditures to state and local governments should be continued. Equally as important, the Government should ensure that the state and local governments have adequate revenue sources to carry-out their expanded responsibilities. B. Improving the Efficiency of Rural Development Expenditures 7.85 The efficiency of rural expenditures in Mexico could be improved significantly. Four sets of factors undermine the efficiency of rural expenditures. First, the absence of a well-defined rural development strategy has led to an expenditure program that does not necessarily reflect the development priorities of the sector, nor include the most efficient investments. Second, efficiency gains from decentralization are undermined by an inadequate assignment of expenditures among the different levels of government, a .lack of clear distribution criteria for Solidarity resources, and by insufficient revenue sources for local and state governments to carry-out their expanded responsibilities. Third, the absence of mUlti-year budgets undermines investment planning as well as the availability of resources to support operations and maintenance activities. Fourth, the budget process is very centralized and controlled by SHCP and as a result the budget does not always reflect the expenditure priorities of the sector. 7.86 Expenditure Strategy. The allocation of federal funds in the rural sector does not necessarily reflect an overall development strategy nor the quality of individual investment projects, despite a detailed planning process and recent improvements in project evaluation requirements. As a result, resources often have been channelled towards poor quality projects/programs and there is a lack of coordination in rural development expenditures leading to overlapping programs, in some cases and in others to underfunded programs. To avoid overlap and contradictions among the different programs in the sector, there should be increased coordination on a program by program basis for rural development expenditures. A consistent review of functions and responsibilities of the different 86. SARH. PROCAMPO: Vamos al Grano Para Progresar (1994).  141 agencies involved in rural development policy might allow further useful consolidation of activities in particular agencies. An institute for rural development policy would also help to develop a cohesive sectorial strategy. which could guide more effectively expdenditure decisions. 7.87 A standard methodology for ranking investment projects in the rural sector according to their economic merit should be developed and SHCP and INE should have expanded resources to evaluate the economic and environmental ranking of'proposed investment projects. The use of a standard methodology. which clearly highlights the economic and environmental contributions (or costs) of an investment project might make it more difficult for high-level officials to alter the allocation of resources with compromisos. 7.88 At the state level, the coordination role of the COPLADE to allocated joint federal state investment funds is not fully utilized and federal and state institutions execute their agricultural programs in a somewhat isolated way. The closely integrated system of decision making represented by COPLADE is best suited for those situations when there is political agreement at all three levels of government. As Mexico gains in political plurality the close-knit system of the COPLADE may have to be substituted by more indirect and decentralized institutions. For example. most federal countries pursue national objectives by enticing state andlor local governments into transfer programs earmarked for certain activities. or matching grant programs in which eligibility and funding level respond to well established objective and transparent criteria. The federal government has begun to introduce investment selection criteria for funds transferred to the states under Decentralization and Regional Development I and II. This experience needs to be evaluated and applied in all Mexican states. 7.89 Decentralization. Several issues undermine possible efficiency gains from Mexico's decentralization policy for the rural sector. Many rural development services are provided in a mode of shared responsibility and are not clearly assigned to a particular level of government. Many services are not clearly assigned to the central. state or local level of government and as a result are under provided. The assignment of expenditure responsibilities for rural development both for capital and recurrent costs should be evaluated. In the long run, both state local governments need to take full charge of the planning, financing and execution of the capital projects in the sectors or services for which they are responsible. Not all states and municipalities will have the same ability to take over these responsibilities at the same time. of course. A gradual and differentiated approach will be required. 7.90 Similarly. responsibilities for current services and capital investments are not always assigned to the same level of government. Over time. the assignment of expenditure responsibilities among the different levels of government for capital investments and recurrent expenditures should be the same. The integration of the maintenance and the capital investment decisions will avoid the inefficient use of resources implied by the lack of proper maintenance. 7.91 The assignment of expenditure responsibilities to local levels of government has not always been accompanied by adequate increases in revenue sources. Shifting capital expenditure responsibilities to state and local governments will also require a corresponding adjustment in the assignment of tax instruments, revenue sharing. or transfers to make possible the autonomous financing of these expenditures. Moreover. the federal and state governments should have certain budgetary obligations such as, for example, transferring funds in block grants or matching grants to promote investment in lower level jurisdictions when benefits spill over to other communities or when there is a desirable distributional impact.  142 7.92 The decentralization process in Mexico has been confused by the fact that some state officials perform a dual role also as representatives of the federal and state governments. Government officials should not have joint responsibilities for the federal and state governments as this could lead to conflict of interest. 7.93 Finally, the mandate to channel most investment funds and capital infrastructure needs through the COPLADE is stifling to local initiatives and may limit the efficiency gains associated with true decentralization. The role of COPLADE in the planning and distribution of resources should be eliminated and it should serve more as a discussion forum. Over time, local governments should be able to allocate their resources as needed. The institution of the COPLADE promotes central planning and works against decentralization to the municipalities. As the municipalities become stronger and better organized, the benefit of resource allocation decisions being made by local communities will offset the risk of overlap and inconsistencies that could occur with decentralization of expenditure decisions to local communities. 7.94 The entire program of Solidarity is an important element in Salinas' decentralization policy. Innovative design techniques, involving the community at all steps of the investment process increase the efficiency of expenditure programs in Mexico. However, a principal drawback of the program is that it retains an excessively high political partisan profile and lacks transparency. The most significant fact in the allocation of Solidarity funds is that there is no formula or pre-established eligibility rules that would ensure that resources flow to those areas and projects that are most in need of the resources and will use them most efficiently. Solidarity resources should be distributed to the states and within the states to municipalities using explicit formulas based on poverty criteria. Under the Second Decentralization and Regional Development project, the eight participating states are developing formulas for the distribution of funds to municipalities to improve targeting, transparency and reliability of funding. 7.95 Multi-Year Budgeting. Mexico lacks a long-term budget framework. Investment planning should be framed in a period longer than the annual budget cycle to permit planning for recurrent (operations and maintenance) and capital budgets and to facilitate fiscal policy. A multi-year capital budget framework would highlight more easily scarcity of funds allocated to maintenance and rehabilitation. The lack of appropriate funding for maintenance of capital infrastructure also reflects the current budgetary system, which itemizes the different types of expenditures carried out by government in "chapters" does not separately identify maintenance expenditures. These are sometimes included in the chapter of "general services" and other times in the capital budget. 7.96 The GOM should develop a rolling multi-year expenditure and revenue planning framework (three to five years). It would not be necessary for the multi-year budgets to be as detailed as the annual budget estimates, nor should they be regarded as strict blueprints or as a commitment of future resources. Rather, they should used as planning tools: to explicitly indicate maintenance and operation expenditures associated with capital expenditures; to analyze the expenditure and revenue levels implied by current policies to determine whether or not they support the planned fiscal targets; and to evaluate the sensitivity of future fiscal plans to changes in macroeconomic parameters. The information currently contained in the annual POAs could be incorporated into the multi-year budget. 7.97 In addition, the use of a commitment accounting system and functional tranches could also improve the long-term efficiency of capital expenditures in Mexico. Many budget systems around the world distinguish between program authorizations, whose validity extends beyond the annual framework, and payment appropriations opened for the year and necessary for the execution of the planned expenditures. Commitment accounting guarantees funds for the authorized programs and  143 therefore avoids the occurrence of unfinished projects. Under functional tranches each tranche enters into operation independently of the other tranches in the investment program and each tranche is funded by a specific authorization in the budget. 7.98 The most serious difficulty for the adoption of commitment accounting and functional tranches is that the Mexican Constitution explicitly prohibits the carryover and commitment of funds beyond one single fiscal year. The use of payment appropriations in the annual budgets for authorized multi- year programs still may allow the adoption of commitment accounting and compliance with the present Constitution. 7.99 The current budget classification system should be reformed so that maintenance appears under a separate new "chapter" in the recurrent cost budget. 7.100 The Budget Process. Mexico has a centralized budgeting system under SHCP. There are several issues in the budget process which merit attention. First, SHCP may introduce extraordinary budgets once the regular budget approval cycle has been closed, which provides the finance secretariat with more budgetary leeway than is conventionally the case. The authority of SHCP to introduce extraordinary budgets once the regular budget cycle has been closed should be eliminated to preserve the transparency of the of and to avoid compartmentalizing the budget process, both of which could lead to a sub-optimum allocation of resources. Discretionary decisions should be made by SHCP only for small amounts. 7.101 Second, there is an especially limited role for the municipalities in the capital budgeting process. As a result, the final budgetary allocation transferred to the municipality may not reflect its most efficient use as defined by the local community. Municipalities, not SHCP or the COPLADEs, should have decision making power for the capital investments for the services for which they provide. While they may need a loan or a transfer from the state and federal governments to carry- out the capital investment, they should have the freedom to allocate the resources towards priority investments as determined by the municipality. Smaller and poorer municipalities will need technical assistance to carry-out these expanded responsibilities. 7.102 Third, there is minimum flexibility in the budget execution process. Secretariats and agencies require prior approval from SHCP for the transfer of funds across programs and even within programs. As a result, resources assigned to a particular program or unit are often used for other purposes, without officially changing expenditure records. This makes it difficult to accurately monitor rural development expenditures. Federal agencies should have some flexibility to allocate resources within their overall budget to avoid informal re-allocations of resources. Changes in the original budget should be limited and should be reported to SHCP so that the national accounts reflect the true costs of the agencies' programs and to form the basis for next year's budget. C. Public Expenditures and Poverty Alleviation 7.103 Revenue Sharing. The overall level of poverty in a state has little effect on the amount of revenue sharing it receives. The revenue sharing system should be re-evaluated not only to increase the overall amount of resources transferred to the states (para. 7.91), but also to ensure that poorer states and municipalities, with the weakest implementation capacities are favored in the distribution of resources. Similarly, the distribution of revenue sharing funds to the municipalities does not seem to be correlated with the relative poverty level of the municipality. All states should have clearly  144 defined formulas to transfer revenue sharing to the municipalities that reflect, inter alia, the poverty level of the municipality. 7.104 Solidarity. The distribution of Solidarity funds is correlated with the state's overall level of poverty. Similarly, the distribution of Solidarity Municipal Funds by state also reflects the overall level of poverty. However, the state's poverty level is not a significant force in the distribution of all Solidarity programs, for instance, it bears no effect on the distribution of Solidarity Production Funds. Likewise, the distribution of Solidarity resources to the municipalities does not appear to be related to the municipalities overall level of poverty. The distribution of Solidarity resources by state should adhere to a clear set of criteria that reflects, inter alia, the state's overall level of poverty. Similarly, each state should develop a transparent set of criteria to distribution Solidarity resources to the municipalities that takes poverty level into account. 7.105 DFis. The role of FIRA as a development bank should be re-considered given the high proportion of its lending directed towards the wealthier producers in the sector. 7.106 SARlI. The role of SARH investments in poverty alleviation appears to have been largely overlooked. Most notably, SARH investments in irrigated areas accounted for 22 percent of SARH's total budget compared to 2 percent for rainfed areas in 1992. Moreover, investments in irrigated areas cover construction, rehabilitation, operations and maintenance, while they only cover construction in rainfed areas, suggesting that the other costs are left to the producers, who in general, have fewer resources than producers in irrigated areas. Finally, SARH should increase financial support to low income farmers to assist them purchase technical assistance. SARH should reinforce its practice of supporting larger delegations in poorer states that have weaker capacities to carry-out rural development activities. D. Impact of Taxation on Incentives for Private Investments in Rural Development 7.107 The overall tax burden on the agricultural sector in Mexico is low, reflecting preferential treatment for the sector in the income, payroll land and asset taxes and an even more preferential treatment in the VAT. Moreover, the tax structure favors capital intensive production techniques in agriculture, although the pro-capital input bias is somewhat offset by the lack of social security coverage for agricultural laborers. The overall burden on the agriculture sector has also decreased relative to other sectors because the implicit taxation of most agricultural activities via the protection of industry through import tariffs has been removed. Nonetheless, the appreciation of the exchange rate in the recent past however has undermined the competitiveness of the agriculture sector (as well as other tradeable sectors). 7.108 The current tax structure also favors larger farmers over smaller ones. The favorable treatment of large agricultural producers under the income tax may make it more difficult for small farmers to buy land and to compete. The extensive use of zero rating under the V AT for agricultural activities and many inputs used in agriculture also tends to benefit almost exclusively larger producers. Large farmers are the principal beneficiaries of the sector's generous exclusions in the VAT as small farmers would be protected almost as much by the general small business exemption also included in the VAT. 7.109 The removal of the preferential tax treatment from the agriculture sector as a whole could have some important employment effects. The goal of transferring resources, including labor, out of the agricultural sector and into more efficient uses in the rest of the economy assumes that these inputs can be absorbed somewhere else in the economy. Because of the known difficulties for further  145 absorption of labor in the industry and service sectors, the removal of tax privileges and advantages should proceed cautiously. 1.110 To summarize, the recommendations made here for changes in the tax treatment of agriculture are to: (a) incorporate large agricultural businesses into the general income tax regime; this would be best accomplished by applying the general eligibility rule of the simplified regime (annual proceeds of under N$600 million) to every sector of the economy including agriculture; (b) correct the anti-labor bias in the Gross Assets Tax by including all types of assets in the base and by eliminating or reducing the 100 percent tax credit for capital equipment; (c) incorporate the agricultural sector in the general VAT regime by eliminating the widespread use of zero rating except for exports, as in the general case. Basic foodstuffs could be exempt from the VAT. Small farmers would remain exempt through the appl ication of the general small business rule; and (d) increase effort to assist state governments update land cadastre so that assessed land values are closer to market values under the property tax. 1.111 These recommendations call for the uniform tax treatment of the agriculture sector with respect to other sectors. They do not, nonetheless, preclude the use of tax policy for some targeted purposes in the sector. For example, at present there is no incentive or special treatment for environmental activities nor are forest or park lands treated in any special way. The only exception to this is at the state and local level where the Impuesto Predial and several fees are lower in some states for land dedicated to forests and parks.  146 8. THE LAND MARKET 8.1 Land tenure has al ways been one of the most difficult aspects of Mexican farm policy. By 1910, the degree of land concentration in Mexico was greater than any other Latin American country and peasant movements were a principal force behind the Mexican revolution.17 As a result, the Constitution of 1917 guaranteed all Mexicans land rights through expropriations of large landholding and the peculiar institutions of Mexico, the ejido and communal lands system. These forms of tenure provided their members with usufruct rights over land and water resources, while prohibiting the sale or rental of ejido land, and imposing other restrictions on economic activities in the ejidos. 8.2 Although the commitment to land reform was enshrined in the Constitution, its implementation was gradual and sporadic reflecting the political will of the various presidential administrations and ambiguity over the definition of maximum land holdings. The principal focus of land reformers was, initially, to provide farmers with access to land. As the area available for redistribution became smaller, the focus shifted more to encourage private sector investment in the ejido sector and to regularize the legal status of al1 farmers by ensuring adequate title to land (Heath 1990, P. 6). A basic issue - that of developing a land market in the context of contributing to growth - tended to be overlooked. The recent constitutional and legal changes and the nascent effort to title land have changed the focus, at least partly, to concentrate on the development of a land market, but obviously cannot eliminate the popular concern with the distributive effects of a policy guaranteeing access to land. 8.3 This chapter reviews some of the emerging land market issues as they are likely to effect growth. First, we review succinctly the main features of the old system. Second, we summarize some of the old literature about the ejido system and its impact on productivity. Third, we discuss what the tenure reforms have been and review their legal adequacy. Fourth, we review progress with specific steps to title land in order to create an efficient land market". Obviously, there are many other issues affecting growth from the ejido and small holder sectors, notably finance, technical change, input markets, and product markets. but these are covered in other chapters. 87. Eckstein el ai, Land Reform in Latin America: Bolivia, Chile, Peru and Venezuela, World Bank Staff Working Paper No. 275, Washington, D.C. 1978:17 in John Heath, Enhancing the Contribution oflhe Land Reform Sector 10 Mexican Agriculture Development, World Bank Report 83-IO-ME (March 30, 1990) at I [hereinafter Heath]. 88. For the review of the old system and of the productivity effects of the ejido system, see Heath. Legal issues are from an internal Bank memorandum prepared by Rudolf van Puymbroeck of the Legal Department. Titling issues are adapted from documents prepared by Louise Cord for a land titling project. Much of section I is taken directly from Heath's work; all paragraphs quoted in their entirety from Heath's report show the paragraph in bold face italics, and parts of paragraphs are shown in normal quotation marks. Heath's report assumed that the "ejido may be considered a more or less flXed element of Mexico's cultural and political scene" (p. 1) and that is obviously not longer true, but most of his work is still pertinent. The Bank has never been involved in Mexico's agrarian reform efforts except to provide informal policy guidance and support studies, including that of Heath. After the passage of the reforms in early 1992 and at the request of the GOM, the Bank organized a conference in Washington in April 1992 to present its experience with land titling projects in other countries and to discuss the impact of land reform on agricultural productivity.  147 I. The Old System A. The Legal Basis 8.4 The legal foundation of the Mexican land reform is derived from Article 27 of the 1917 Constitution, which affirms that all land and subsoil resources are ultimately the property of the nation. While the principle of private property ,?wnership is conceded, the state reserves unto itself the right to modify the existing tenure pattern in order to ensure a more equitable distribution of wealth; to this end the state is authorized to expropriate private property. 8.5 The land was not given directly to farmers, but to ejidos and communal lands communities. These form the basis of the land reform sector (sector social). There are two types of ejidos. Individual ejidos are those in which member households work parcels of land on an individual basis and they have much in common with villages whose inhabitants are small private landowners. Collective ejidos work land on a centrally managed basis. These ejidos have been imposed on producers from above, are unpopular with farmers and probably account for less than 5 percent of arable land Bo • A third type of tenure in the land reform. sector are communal lands communities, which are comprised of indigenous groupings. 8.6 The ejidos manage about half of Mexico's crop land, and roughly equal shares of both rainfed and irrigated crop land as well. By 1991, there were 29,951 ejidos comprised of 3.5 million ejidatarios with 102.9 million hectares.!IO This represented 52 percent of Mexico's total land area and 25 percent of the population (INEGI, 1991). The largest concentrations of ejido lands are found in the states of Coahuila, Chihuahua, Durango and Oaxaca. Much of the ejido farmland is of low quality and in many instances contributes little to ejido income. The majority (74 percent) of the ejido land consists of pasture and forest. Agricultural land is only 21 percent, the majority of which is rainfed. B. Operational Restrictions on Land Use 8.7 The limits on size of holding were determined by production type. One hectare with irrigation were the equivalent, in the terms of the land reform law, of: two hectares of rainfed land, four hectares of pasture, eight hectares of brush. Private farmers could, with certain exceptions, hold 100 hectares of irrigated land or its equivalent in another type. Other limits were set for the maximum land area that could support 500 cattle and for individual crops (e.g., cotton, vanilla, cocoa, sugar cane). Nonetheless, various legal and political strategies were deployed to evade the holding limits, making it difficult or impossible to distinguish between the land owned per farmers and the land actually operated. For example, clandestine land rentals (see para. 8.8) were known to be common and might have tended to raise the size of operated holding and thereby to exploit economies of scale. (Illegal land sales also occurred and would also have tended to circumvent the land holding limits). Expropriated landowners were entitled to compensation, though actual payments were often very slow. 8.8 The laws also restricted land management. Ejidatarios could not hire labor unless they were fully occupied working with their employees. The sale and rental of land was prohibited. 89. Because the collective ejidos are and have been so uncommon, the aggregate productivity gains of changing collective to private tenure. as have been found in China for example. are likely to be much smaller in Mexico. 90. Hereinafter, all references to ejidos also refer to communal lands communities.  148 Sharecropping was also not allowed in most instances. Land left idle for more than two years was supposed to revert to the ejido community. Because land was inalienable. it could not be used as collateral and this was believed to limit the supply of credit to ejidos. Moreover. ejidatarios could not receive credit individually. only through the ejido leadership. thus making financial decisions subject to political influence within the ejido. The continuous process of land subdivision from fathers to sons led to a decrease in average farm size and thus may have promoted migration as many farms because too small to provide a decent income. The literature on the implementation of these restrictions was necessarily contradictory because compliance could almost never be determined accurately or broadly. Rentals and informal sales were reported to be widespread. despite being illegal. What was clear was the decisive importance of local political influence which determined the interpretation of the initial agrarian laws, their later amendments and the application of the entire body of agrarian law. c. Productivity Effects 8.9 The above legal and management restrictions might tend to reduce productivity in the ejidos by interfering with investment and allocative efficiency. An earlier World Bank paper (Heath, 1990) formulated five hypotheses relating to the performance of ejidos and private farms and reviewed the evidence then available about them. (a) Tenure insecurity in ejidos gives them less incentive to invest. (b) Restrictions on land transactions, notably renting and sharecropping, prevent the exit of inefficient enterprises and the consolidation of small enterprises into larger ones. (c) Communal land in the ejidos is less productive because it is poorly managed. (d) Ejidatarios have poorer access to credit and therefore are less productive. (e) Ejidatarios are more productive than private farmers because the latter are subject to complex and ambiguous restrictions that dampen incentives to invest and grow. 8.10 Many empirical problems confound the analysis of the productivity of ejidos. The status of an ejido depends on the quality of land received, market access, the possibility of irrigation, the availability of credit (which has always been politically directed in Mexico), and access to extension and marketing services. If those factors are not, or cannot, be controlled in the analysis then it is impossible to separate the tenure effect from any of the others mentioned. 8.11 The earlier review of several studies based on census data collected before the 1970s, "provide no firm basis for reaching conclusions about the relative efficiency of ejidos and private farms" (Heath, 1990 at 30). Some analyses showed ejidos to be more efficient, and others less. This is an important finding because it contrasts thoroughly with what has been found in comparisons of collective land tenure to private farm holdings in the former Soviet Union and China, and thereby weakens the case for an immediate jump in productivity after reform of the ejido system. Case studies of ejido efficiency are no clearer. Studies of irrigated production found that ejidos were more productive than private farmers of equivalent farm size. In rainfed areas, ejidos were less productive, perhaps because ejidatarios with irrigation had access to credit but those without irrigation did not. The previous review concluded that "recent [after 1970] evidence suggests that, comparing same-size enterprises in each tenure sector, ejido farms are likely to be at least as productive on average as private farms" (at 32).  149 8.12 An important critique of the above conclusion - that land reforms improving land rights and land markets in the ejido sector are unlikely to produce important short to medium-term productivity gains-is as follows. While productivity studies tend to support the conclusion, those studies compare constrained production in the ejidos to constrained production among the pequefl.os proprietarios (smallholders). A better analysis would be to compare both the ejidos and the pequeflos proprietarios to a less constrained situation, notably one in which both types of producers face more efficient input markets and receive better public services, especially research and extension. While it is empirically very difficult to analyze this issue, for lack of appropriate information, the contention is probably a good one - that rapid productivity gains and ensuing income growth in the ejido and smallholder sectors will require a package of reform in land markets, input markets, irrigation services, research and extension. II. The Recent Reronns A. The Legal Rerorms 8.13 In early 1992, Article 27 of the Constitution and the Agrarian Reform Laws were modified to give land ownership directly to residents of agrarian communities (ejidos and communal lands) and to allow them develop new forms of economic association. These changes were made to permit the development of rural land markets and to provide incentives for farmers to adapt to the Government's macro economic policies for the sector. More specifically, these changes are expected to increase the sector's productivity through the lifting of restrictions on capital flows into ejidos and communal lands, through the creation of a land market to permit the consolidation of minifundia into commercial farms and by raising the security of tenure in agrarian communities. 8.14 The new Article 27 contains the following modifications: (a) The land distribution clauses of the old Article 27 were repealed. (b) Land above the legal ceilings could be sold by the current owner within two years. (c) Size limits on holdings were maintained, although they can be released after investments in land improvement (e.g., irrigation); in other words, investments resulting in holdings greater than the limits are not at risk of expropriation. Holdings of individual ejidatarios were formerly limited to less than 20 hectares of rainfed land; this limit has now been changed to forbid an individual ejidatario from holding more than 5 percent of the ejido area. (d) Ejidatarios can enter into associations of all legal types, including joint ventures and rentaJs, without prior approvaJ of anyone, specificaJly including the ejidal authorities. (e) Communal land areas of ejidos (mainly pasture and forest) remain under ejidal tenure (e.g., they cannot be sold). (f) Ejidos can rent their land to anyone or sell it to members of the same ejido. A majority vote is required by the ejido members before an ejidatario can sell land to a non-ejido member. However, individuals will not be forced to cede their rights as ejidatarios if they do not want to.  150 8.15 From this review, and from other analysis of the new Agrarian Law, it seems clear that a stronger legal basis has been established in the ejidos for economically significant activities. The latter include land rentals and sales, with the effect of consolidating small units into large and perhaps raising productivity; land pledging, with the effect of using the land as collateral and thereby obtaining credit; joint ventures, with the effect of increasing the participation of outside capital; faster resolution of land disputes, thereby reducing their costs; nil risk of expropriation, thereby making investments more secure. Whether this new legal basis does in fact improve productivity depends on the interpretation of the historical productivity evidence which, as we have seen, is unclear. B. The Land Titling Program 8.16 To implement the changes in Article 27 and the Agrarian Reform Laws, the Government seeks to grant individual titles for up to 4.6 million agricultural parcels and 4.3 million house plots currently used by residents of ejidos and communal lands. The program, is called the Program for Certification of Ejido Rights (pROCEDE; Programa de Certificaci6n de Derechos Ejidales y Titulaci6n de Solares Urbanos). The Minister of Agrarian Reform is the head of PROCEDE. INEGI is formally responsible for surveying and preparing the maps and titles, which will be stored and updated by the RAN (National Agrarian Registry). The PA (Agrarian Attorney General), created by the new agrarian legislation, has, among its responsibilities, assisting ejidos organize themselves to carry out the land titJing program and arbitrating intra and inter ejido land disputes. The decision to receive a title, the titling process itself and the decision to open-up land sales to outsiders are all made by majority vote in the ejido assembly. In the event that the PA is unsuccessful, land disputes will be solved by the newly created Agrarian Tribunal (fA) which is also charged with resolving pending cases of land reform from the old regime. SARH has a general coordinating and monitoring role. 8.17 On January 6,1993 (the anniversary of the land reform program of 1915) the President launched the PROCEDE. The objective for 1993 was to title 30 million ha (approximately 33 percent of all ejido land), reaching 9,000 ejidos covering 1.4 million parcels and 1.25 million house plots. By August 1993, 164 ejidos had successfully completed the program. The programmed budget for the land titling program for 1993 was US$3OO,000 (then roughly N$1 million), with INEGI receiving 86 percent and RAN and PA the remainder. C. Costs and Benefits of Land Titling 8.18 In Mexico, as in other countries, benefits from land titling will not be experienced equally in all communities and among all ejidatarios within a community. Many commercial banks will remain hesitant to lend to individual ejido farmers with meager plots of rainfed land, given the high transaction costs and substantial risks associated with small loan sizes, and the already high levels of outstanding ejido debt. Furthermore, the demand for poor quality rainfed land located in isolated areas is expected to remain low, particularly once the NAFfA accords are implemented, which could constrain the increase in land values normally associated with titling programs. Finally, in many ejidos the existing system already is characterized by security of tenure and low levels of land conflict. 8.19 Moreover, the provision of a title is only one aspect of the required program; equally important is the establishment and maintenance of a registration system for all types of title (ejido boundaries, communal, individual), the development of a training program for ejidatarios concerning the use and management of a title and registration system and, in a wider context effective systems to deliver technical assistance, credit and other services (e.g., irrigation) as appropriate.  151 III. Conclusions and Recommendations 8.20 The GOM needs to view the titling program within the long-term objective of developing efficient land markets, based on legal contracts and with information easily available to buyers and sellers at a reasonable cost. However, the goal appears to be to provide as many titles as rapidly as possible, and less attention has been given to the creation of sustainable (registration and surveying) practices and institutions necessary to support an active, rural land market that is responsive to the varying needs of ejidos. In particular, (a) insufficient attention has been given to the institutional strengthening of RAN and the state public registries and to the creation of simple and accessible registration procedures, while too much attention and resources have been provided to INEGI for the initial surveying and titling processes; and (b) INEGI has adopted a surveying technology that will provide a precise picture of landholding at the time of measurement; however, insufficient attention has been given to the adoption of simple and less costly surveying techniques that will be affordable and available to residents of ejidos to document future land transactions. 8.21 There is no comprehensive framework for the program with clearly stated objectives, strategies and coordination mechanisms for the various institutions involved. While the Ministry of Agrarian Reform was selected as the lead agency for the program by President Salinas, its overall level of support for the reforms is relatively weak they essentially contradict SRA's mandate. The lack of a comprehensive program, an effective lead organization and coordination mechanism has encouraged institutional conflict and inappropriate resource allocations among the different institutions. 8.22 Cost Effectiveness. It is not sure that INEGI has selected the most cost-effective surveying technology. Less precise and costly technologies may be more suited for certain areas where the benefits from titling will be limited and they also may be more accessible to private surveyors to document future changes in land holdings. In addition, program costs may be overestimated as linkages with related programs have not been fully developed (e.g., land use mapping national cadastre and land information system). 8.23 Specificity of Program De'lign. The high cost of the surveying technology, the complexity of the registration procedures and the focus of the communications program are most appropriate for the wealthier more productive ejidos. The program does not reflect the needs of "special" ejidos, such as those in marginal rainfed. forest, coastal and urban areas. While the Government policy is to respond to all ejido demands for land titles, it focussed the pilot program and the current promotion campaign on the more productive ejidos. 8.24 Institutional capacity. While INEGI has experience in mapping and large scale survey/census projects, it appears to lack experienced personnel in the large scale development, management and analysis of spatial databases. Similarly, the capability of the PA promoters to facilitate boundary demarcation and to mediate and arbitrate a wide variety of land tenure conflicts- especially among indigenous groups or ejidos with squatters-is not clear.  152 9. THE TROPICS 9.1 The tropical states of southern Mexico are much poorer than the arid and temperate states of the north. Despite their current lag, in some areas they have good potential for growth of output, employment and poverty alleviation. Policies to raise productivity and incomes in the tropics must, however, consider several distinctive features of the zone: (a) a diverse, ethnically heterogeneous population, with deep social divisions; (b) contrasts of rich and very poor producers; (c) an ecologically fragile environment; (d) many very small producers who produce the majority of output; and (e) little potential for irrigation in comparison to the north of Mexico. Trade liberalization will severely damage maize (and probably bean) production in the tropics thus cutting producers' incomes in the short run. In the long run, the tropics have potential for export and domestic markets in crops, livestock and forestry but modernization of tropical agriculture faces tough technical problems and risks aggravating social and environmental problems. 9.2 To review the current situation and to outline options for the future, an FAO/CP mission analyzed agricultural development experience and potential in four states - Chiapas, Oaxaca, Tabasco and Veracruz - in the humid tropics. 91 Specifically, FAO/CP sought to: (a) evaluate the experience in and potential for agricultural development; (b) analyze the existing cropping and livestock systems and their likely changes; (c) assess possible impact of the North American Free Trade Agreement (NAFfA); and (d) analyze agricultural development programs. The review focused on food crops, perennial cash crops and livestock. I. Physical Characteristics 9.3 The study area covers approximately 266,000 km 2 which represents 14 percent of the national land mass (Table 9.1). There is a great diversity of agricultural production related to the variability in altitudes, soil types, topographies, and microclimates. Annual temperature is in the 22 to 26°C range and rainfall is mostly in the 1,000 to 2,000 mm range; altitude varies from 0 to 5,610 m.a.s.l. The study area represents some 15 percent of the country's total surface and has 42 percent of the national runoff. Oaxaca and Chiapas have about 850 km of Pacific shoreline and Veracruz and Tabasco some 930 km of Atlantic coast, representing in total some 16 percent of the Mexican coastline. The four states, although occupying basically a humid tropical area, have a wide variation of climates and soil, in addition to abundant water, which has allowed them to develop a diversified agriculture. 9.4 Agriculture's share of GDP in the four states varies from 3.8 percent in the oil-rich state of Tabasco to 12.3 percent in Veracruz, 15.5 percent in Chiapas and 23.5 percent in Oaxaca, though its share of employment is much higher. Much of the country's export potential lies in the tropics and in tropical crops. Coffee is Mexico's single most valuable export crop and approximately 80 percent of the coffee production comes from the study area. Coffee exports in the 1988-1991 period varied from US$333 to US$513 million. Maize is the most important crop in the region, covering over 1.5 million ha. 91. FAD/CP, "Mexico-Review of Agricultural Development Experiences and Potential in the Tropics," Vols. I and II, Report No. 4/94 CP-MEX.37 (1994). The full report is available from LA2NR in English and Spanish. Forestry issues are treated in other sector work underway in LA2NR.  153 9.5 The four states are diverse in other ways (fable 9.2). Veracruz has acceptable road infrastructure, installed processing plants and proximity to markets in Mexico City giving it good potential to intensify and diversify an efficient production of crops (orange, chili pepper, pineapple, etc.) for market and processing. A good network of dairy plants and abattoirs, indicate potential for the intensification of livestock production in Veracruz and Tabasco. In Oaxaca and Chiapas, the availability of labor, indicates a potential for the production of labor-intensive horticultural crops. All four states have adequate areas where the production of food crops (maize, beans, cassava, etc.) could be developed. All four states produce tropical crops, like banana, rubber, cashew and mango, which could be developed; in fact, the agroecological requirements of some of these crops can only be found in the study area. Table 9.1 Main Characteristics or Four Tropical States EJChiapas Area (km~ 73,887 Total Population (Ooos) 3,210 Rural Population (ooos) 1,913 Indigenous Population (000s) 1,178 Density (hab/kml) 43 Farm Employment (% or total) 58.3 Oaxaca 95,364 3,020 1,827 673 32 52.9 Tabasco 24,661 1,502 755 86 61 35.6 Veracruz 72,815 6,228 2,728 1,084 86 39.5 Total 266,727 13,960 7,224 3,021 52 46.3 Mexico 1,967,183 81,250 23,319 4,617 41 22.6 SOURCE: FAO/CP, "Review of Agricultural Development Experiences and Potential in the Tropics," Vol. I. Table 1 (1994). II. Principal Conclusions about Comparative Advantage 9.6 The FAO/CP mission prepared economic and financial enterprise budgets for the main crop and livestock activities. The conclusions about comparative advantage follow from those budgets, as summarized in Table 9.3.92 9.7 Maize. In Oaxaca, due to their relatively low yields, mechanized maize areas are likely to decline with the fall of maize prices. Producers in Chiapas are likely to find it more difficult to compete on the national market because of the high relative transport costs to Mexico City. Productivity improvements in no till areas, which have been expanding lately, could promote a substantial and beneficial expansion of this agricultural practice. Producers who currently have the highest yields would suffer the greatest losses in revenues from maize; those who have the means and market information will gradually diversify their production. Subsistence maize farmers have limited 92. Details of the budgets are in FAO/CP, "Review of Agricultural Development Experiences and Potential in the Tropics .•  154 possibility for diversification, but some would still present a potential for improving production for their own consumption and of marketable surplus; subsistence maize production would probably maintain its importance, therefore. Thus, in summary, the area planted to maize by commercial growers is expected to decrease and no significant change is expected in remote areas. 9.8 Commercial maize production in areas of moderate potential, typical of Oaxaca, would become unprofitable. There are no technologies to make it so and no obvious alternative field crops. Intensification of subsistence maize production would result in modest increases in yields, there would be little incentive for such intensification. Intercropping of maize and beans might save labor and provide greater returns per area cultivated, but has little growth potential. 9.9 Coffee. Although coffee plots have been neglected and most have been established in marginal zones, therefore presenting low yields, a decrease in area is not expected because an unattended crop still provides small farmers with an important supplementary income. Furthermore, duty free trade a1lowable within the NAFTA area for coffee products may increase demand for Mexican coffee in Canada and the USA. There is also scope for improving coffee quality, which would a1so have a significant and positive impact on returns and on the diversification of markets. Areas managed for the production of special coffees (organically grown, etc.) for demanding, mostly European, markets are likely to expand slowly to cover a maximum of 5 percent to 10 percent of the total coffee area. Coffee therefore appears to have better prospects than thought. Although the recent drastic price decreases have devastated the income of sma11 farmers, even at present low prices, returns to labor are reasonable even in low potential areas and as prices increase they would become attractive. The reasons for recent neglect of smallholder plantations are therefore unclear. Table 9.2 Land Use and Tenure in Four Tropical States Ejido Area Land Use (OOOs ha) Area E.jidatarios Crops Pasture Forests Total State Chiapas 4,064 16,000 1,082 1,972 3,091 6,145 Oaxaca 7,664 18,000 802 2,550 3,106 6,458 Tabasco 1,115 21,000 181 1,655 123 1,959 Veracruz 2,944 12~ 1,084 4,645 1,562 7,291 Total 15,786 67,000 3,14n 10,822 7,882 21,853 Source: FAO/CP, "Review of Agricultural Development Experiences and Potential in the Tropics,· Vol. I, Tables 2 and 3. (1994). 9.10 Cocoa. The rich soils pI anted to cocoa are likely to be slowly occupied by large farmers and companies. These have rented - and now have the possibility of buying - ejidatario land which is then put into production of tropical fruit crops (banana, papaya, passion fruit, etc.) for export and the internal market. This seems to be an undesirable trend due to deforestation (elimination of cocoa shade trees) and intensive pesticide use. Although the scope for improvement of cocoa production is limited, the environmental importance of the crop may warrant its rehabilitation. Cocoa production, even from established trees, appears to be unprofitable and in the future, with improved yields, would  155 struggle to break even. Bananas are already encroaching on cocoa areas and, given the profitable nature of banana production, is likely to continue until market limitations are encountered. 9.11 Sugarcane. Cane is a major crop in Mexico, which is one of the largest producers in the Western Hemisphere. In the past, sugar cane production alone has been nearly one percent of GOP, though its value has declined in recent years. As the relative value of sugar output has fallen, the sugar trade balance has become negative and Mexico has become an importer since the mid-1980s after being an exporter or at least self-sufficient for many years before then. 9.12 The area in sugarcane is likely to present few changes. On the debit side, mills with outdated and inefficient machinery as well as those located in areas presenting very low yields, such as Santo Domingo, may be deactivated, forcing growers to abandon the crop. For these, obvious alternatives include tropical fruit and vegetables (including fresh maize) and, for those with 15-20 ha, pasture for dual purpose cattle. In certain areas of southern Veracruz, tropical fruit crops (papaya, banana, etc.) for the internal market are slowly expanding and are already being rotated with sugarcane. Table 9.3 Current and Expected Profitability or Crops in Four Tropical States Current Profitability Expected Profitability al Crop State Technique N$/ha N$/daybJ N$/ha N$/day . Maize Chiapas & No tillage 1,652 53 1,423 41 Veracruz i Maize Chiapas Mechanized 1,819 70 1,225 47 Maize Oaxaca Mechanized 1,490 28 980 19 Sugarcane Tabasco Rainfed 724 20 1,413 15 Sugarcane Veracruz Irrigated 1,198 56 1,794 62 Cocoa Tabasco Rainfed (866) 7 (lOS) 15 Coffee Chiapas & Low (22) 15 783 23 Oaxaca potential Coffee Chiapas & High (101) 14 2,829 31 Oaxaca potential NOTES: oJ With yields from currcnt technology and expected 1994 producer prices. bI With yields from improved technology and expected 1995 producer prices as affected by free trade. SOURCE: FAO/CP, "Review of Agricultural Development Experiences and Potential in the Tropics.· Vol. I. Table 12 (1994). 9.13 About half the sugar cane produced in Mexico is grown in the four study states, of which Veracruz is by far the most important. The technical characteristics of sugarcane production add to the local importance of the crop because they create highly concentrated employment in small areas. Heavy government interference in the sugar economy has further added to the importance of the crop in certain areas within the four states. 9.14 The current Government has made drastic reforms in its sugar policy in an effort to reduce fiscal transfers to the sector and to make it more competitive. Since 1989, it has privatized 64 state  156 mills, liquidated the state company (AZUCAR-National Sugar Company) that formerly monopolized sugar distribution, eliminated restrictions on consumer pricing, liberalized producer price formation and introduced reforms in the labor relations of cane production and processing. The reforms have lessened, but not stopped, the clamor of cane workers and millers for public support. Very recently, Mexico gained some access to the US sugar market under NAFrA and this is expected to lead to an increase in producer prices and to a revival of activity in some of the less competitive mills. The Government, through SARH and ASERCA, has sought to create a spot market for sugar in hopes of bettering pricing information. 9.15 Oranges would appear to have good prospects, but continue to be subject to high price risk. The area planted to citrus is not expected to vary significantly due to a slowly expanding domestic market, a strong international competition for fresh concentrate and some agronomic limitations. As with other commercial crops in the region, citrus are expected to be gradually concentrated into agroecologically suited areas. 9.16 Livestock have good prospects, speciaUy due to the positive results obtained by the application of simple management practices. A major expansion of the livestock frontier and an indiscriminate diversification of crop areas into cattle production do not seem to be realistic under the present circumstances, and an increase in cattle numbers not recommendable. Gradual increments of milk and beef production would only be obtained through the intensification of the present practices, which would lead to a reduction of production costs. The area is likely to see an intensification of livestock production rather than an expansion of the livestock frontier. This intensification, already an on-going trend, is likely to be the result of improved technologies adopted by middle level ranchers. Under favorable soil and climate conditions, organized segments of the social sector would also be able to intensify production, reduce costs and become more competitive. 9.17 Other crops. Several annual (possibly vegetable crops and ornamental plants) and perennial crops (tropical fruit crops, rubber. african palm, ornamental plants, etc.) would gradually occupy, through replacement of crops. larger areas in the study area. This diversification, because it is risky, would have to be based on more detailed information on the agronomic, economic and environmental aspects of production than is currently available. III. Limiting Factors 9.18 Policies to raise productivity and income must take into account the different types of producers and the key constraints (institutional, ecological, technical, economic, infrastructure, social and legal) which limit the development of small farmers, who are most numerous in the study area. Institutional constraints 9.19 SARH's restructuring and consequent withdrawal has left a void that has not been filled by public or private institutions. 9.20 There is weak coordination among institutions, both federal and state, concerned with agriculture. Regional complementarity is commonly lost because of the lack of coordination. This is of special significance since parts of the states share agroecological zones and markets. Mechanisms for coordination. such as COPLADE. have not been effective.  157 9.21 State and municipal institutions for environmental protection are need for the intensification of tropical agriculture. Where there is intensification of export crops, such as banana in Tabasco, monitoring of pesticide application and residues is critically lacking. 9.22 The impact of the many and varied agricultural development programs on productivity, farm- level returns and income is unknown. Rational proposals for program adjustments by the federal and state governments therefore cannot be made in response to the rapid sectoral changes. 9.23 Agricultural data are inconsistent and incomplete due to lack of appropriate institutional resources and varying methodology; in certain cases, as represented by livestock population figures generated by SARH and INEGI, the difference reaches 50 percent. This severely limits detailed analytical work. Ecological constraints 9.24 Large areas are eroded. This process can be partially reversed by the adoption of proper farming practices. A more difficult option is to match land use to land suitability. 9.25 Profitable options to diversify agriculture are limited. This is of major significance to areas that might have to abandon maize in Oaxaca and Chiapas, or cocoa in Tabasco. Technical constraints 9.26 Much of Mexico's export potential lies in the poorly studied and varied tropical climates and tropical crops. The generation of agricultural technologies which could be used has been very limited, except for livestock. Annual and perennial tropical crops and agroforestry research programs have not received the necessary priority, except for maize and beans. Research programs have failed to consider major problems (e.g. erosion) and zones in the four states and the lack of alternatives will be felt even more strongly after liberalization of the maize price. 9.27 The dismantling of government-supported institutions (crop specific institutes and councils) which had branches active in agricultural research is further limiting technology generation. Major problems due to the lack of technologies, however, would only be felt in the medium and long terms. 9.28 Technical assistance is presently not available to poor producers who have commercial potential, but who cannot yet afford to pay. This is of special significance since the demands for change placed on this target group are rapidly and changing due to trade liberalization. Infrastructure constraints 9.29 Priority has been given to new irrigation/drainage rather than to the completion and/or rehabilitation of the existing one; this has resulted in a severe limitation of resources for intensifying agricultural production in those areas already served by infrastructure, with negative effects on agricultural growth and farm income. 9.30 Relatively little emphasis has been placed on promoting a sound divestment of the operation and maintenance of irrigation and drainage infrastructure to farmers - a key policy adopted by the GOM.  158 9.31 The poor conditions and/or lack of rural roads, particularly in the states of Chiapas and Oaxaca, does not allow the proper integration of many communities into the Mexican society. limits their access to basic social services (mainly education and health), restricts access to markets and affects the introduction and acceptance of improved farming practices which could increase production. 9.32 The poor conditions and/or lack of rural roads also limits the distribution of fertilizers and other inputs, and makes them more expensive. This constraint, which could only be removed in the medium term, indicates that, in the short term, increasing the use of fertilizer and other inputs will be unrealistic in important highland areas mainly of Oaxaca and Chiapas. Economic constraints 9.33 In the past few years, bank credit to small producers has decreased drastically while credit to larger farmers has increased significantly. with more emphasis on long term lending. Although FIRA and BANRURAL rates appear reasonable and adequate funds appear to have been available, small farmers have difficulty in meeting eligibility criteria and costs associated with documentation for loan applications. 9.34 The continuation of PRONASOL programs supporting specific crops Oike maize and coffee) - and not families - reduces incentives for the more rational land use which reforms and PROCAMPO have been designed to foster and NAFTA will make unavoidable. 9.35 Shared risk credit schemes are presently limited to demonstration plots: this severely restricts attempts at diversifying agricultural production in the zone, especially of export crops, already proven to be risky due to lack of research work and proven recommendations. Social constraints 9.36 Limited financial, technical and human resources. in addition to limited coordination. have decreased the effectiveness of recent, and laudable. poverty alleviation policies and activities since the latter are commonly fragmented, costly and overlapping. Furthermore, the unfulfilled expectations of the target popUlation could lead to undesirable social unrest. 9.37 The inadequate aVailability of social services (health, education and complementary nutrition) to the rural population, coupled with the aforementioned infrastructure problems and limited private sector investment in agroindustrial and urban-industrial sectors, which constrain off-farm job opportunities for unspecialized rural dwellers in the study area, are key factors which have promoted migration to urban centers and other agricultural areas. 9.38 The high correlation between rural poverty. marginality, ethnic distribution and the occupation of marginal lands by small farmers, who have a high dependency on the GOM for survival. 9.39 The loss of the most dynamic elements. who emigrate to seek better job opportunities, resulting in an unbalanced social composition of their original indian and ejido communities. 9.40 The limited coordination of programs, indicated above, has promoted the proliferation of producers' organizations, formed as requested by governmental institutions. Producers and/or  159 communities traditional or indigenous organizational structures and needs are therefore commonly neglected. 9.41 Formal and informal agricultural education has not been adequately supported in the study area. Furthermore. agricultural education has presumably not been adjusted to the new needs of the sector faced with a rapidly changing and more competitive situation. Legal constraints 9.42 Although the new Ley Agraria (Article 27) promotes a land market. land conflicts and limited resources have slowed down the land tiding process including its initial step. the issue of certiflcados de dereclws ejidales. There are still questions related to the right of use of communal land. especially forested areas. which have not yet been answered. This would also require new legal instruments. IV. Recommendations 9.43 Specific policies have been identified to minimize or eliminate the previously identified constraints to agricultural development. For much of the population. improved access to health and education is a top priority but this is not strictly an agricultural issue and is not discussed in this Memo. Decentralization 9.44 The public sector's role in fostering agricultural development remains essential and the void left by the retraction of SARH a concomitant strengthening of state capacity. The speed with which roles and functions could be transferred to state secretariats needs assessment; clear differences exist amongst the states. The transition period. which will end when SARH assumes an exclusive normative role. could be much longer than expected. Decentralization implies adequate numbers of trained staff in addition to resources and infrastructure compatible with the new functions. Due to the differences amongst the states' administrative capacities. it is important to involve local governments and communal representations in state agricultural programs. The Environment 9.45 Federal and state governments. and producers and nongovernmental organizations (NGOs). would play a key role in environmental protection and improvement. Due to the importance of soil erosion. specific actions to reverse soil degradation should receive top priority. 9.46 INIFAP. CNA and other institutions have studied technical approaches to promote the recovery of eroded soils. as well as soil and water conservation. which have already been applied over considerable areas of different catchments. The expansion of this positive. albeit limited, experience in the management of natural resources in the study area could be supported by financing a pilot watershed management program. The program could select 1-2 small catchments per state, based on previous in situ land management experiences. 9.47 These pilot activities on selected small cQ/ciuneras would have to be underpinned by appropriate strengthening of agriCUltural research and technology transfer and would depend on the availability of credit and/or an incentive fund, to compensate farmers for externalities.  160 Technology Generation!Adaptation 9.48 In general there are few technically proven and economically feasible crop production technologies for the tropical zones of Mexico. There are a number of topics (breeding, integrated pest management, intercropping, etc.) related specially to potential annual crops, perennial tropical crops and agroforestty which warrant further public research. These medium- to long-term research endeavors, due to their inherent cost and associated risks, are unlikely to interest the very limited and emerging private research groups in Mexico. 9.49 Specific research requirements could be integrated into adequately supported tropical research programs conducted by INIFAP. Additionally, the region possesses a valuable set of technological institutions which could also develop specific research lines on a contract basis. This system is already being used in the region (for example, in Oaxaca), although in a limited scale. Livestock research conducted by INIFAP and UNAM, should be continuously supported in order to provide ranchers with a continuous flow of sound technologies. Topics for research could be included as part of broader regional or state programs for the development of tropical agriculture or specific crop rehabilitation and diversification activities (see below). 9.50 Technology transfer programs for the large number of small producers with a heterogeneous agricultural potential could be developed and supported. Extension activities could be delivered (or financed) by the state or municipal government; the model developed by FIRCO seems to have some key elements which could be adjusted to the requirements of the different groups of producers. Additionally, programs directed towards upgrading the extension skills of private technical assistants could be supported. 9.51 SARH's restructuring has also disrupted the, albeit limited, linkage between technology generation and transfer. Tropical crops and agroforestty research programs for the region should be developed and/or strengthened; federal and state educational institutions, in addition to INIFAP, should be involved in this effort. There is also an urgent need to recover the valuable information and genetic materials generated/evaluated by institutions which have been dismantled by federal government, in addition to reincorporating dismissed, but highly trained, research staff into research institutions. Whether public or private, technology transfer must be continuously supplied with information generated/adapted and validated by research institutions. In the tropical region studied. mechanisms to better link research to public and private extension should be developed and supported. In this transitional period, technical assistance to producers is of paramount importance since the demands placed on the target population have changed. The GOM has a clear policy to withdraw SARH from public extension. Nevertheless, there are several actions and programs conducted either directly by SARH, CNA (through the Technical Cooperation Units [UCTs]), FIRCO, which provide producers with either wholly or partially subsidized technical assistance. In order to avoid sending different messages to growers in relation to the GOM's commitment to the phasing in of paid technical assistance, either through individuals or firms these actions and programs should be made mutually compatible and clearly defined. The success of technology transfer for livestock in Veracruz indicates that relatively simple, inexpensive and efficient extension systems can be developed. 9.52 Although the trend towards private technical assistance is desirable, privatization should protect the interests of poor producers. A rough characterization of producers' needs for technical assistance indicates that small producers with productive potential should be supported by Government-delivered or supported technology transfer programs (including producers' organization) which tend towards the development of private technical assistance. The model developed by FIRCO (for maize, beans, wheat and rice), subsidizing technical assistance in a decreasing proportion for a  161 determinate period, seems to have some key elements which could be adjusted to the requirements of some of the different groups within the study area. Commercial producers should pay for all technical assistance. Despite efforts to training private agents, their number is small and the states need to improve their training and accreditation programs. Infrastructure 9.53 The two key elements of infrastructure are rural roads and irrigation and drainage. Rural roads development has to consider two distinct areas. One is settled areas, where producers with commercial potential are taxed by high transport costs; in those areas, road development is needed to spur commercial agriculture. For example, there are several areas in Chiapas and Oaxaca where producers with potential are far as 70 kID. away from even a dirt road, thus prohibiting the access of basic public services (education, etc.). Priority must be given to the rehabilitation of the existing rural road network and to the development of all-weather dirt roads to allow isolated communities reasonable access to public Services and to markets. Labor-intensive schemes for road construction in hilly zones would be beneficial in creating employment and perhaps causing less erosion during construction. In other, more virgin areas, road development should proceed cautiously and in some instances not at all because of the potential for severe environmental damage caused by rapid new development. 9.54 As for irrigationJdrainage schemes, CNA would have to give priority not to the development of new infrastructure but to the completion of on-going work and rehabilitation of districts and units, always coupled with the establishment of water users' associations (WUO) which would be responsible for the operation and maintenance of the works. The process of transferring the responsibility to WUOs, slow and complex, would have to consider the need of a strong government- supported technical assistance component to the WUOs, as to permit WUOs to: (a) better organize themselves; (b) program their activities based on a realistic assessment of their opportunities as producers; and (c) better understand the formidable management task lying ahead of them. It might also be technically and economically feasible to support small irrigation for periurban horticultural production in parts of Chiapas, Oaxaca and Veracruz. 9.55 It is reasonable to suppose that more government inves~ent in infrastructure is likely to be needed to ease key social infrastructure and transport constraints in the rural areas. Furthermore, government's investment in infrastructure through activities and programs designed to intensely utilize manpower would create much needed job opportunities in the area, therefore providing small producers off-farm opportunities to increase household income. 9.56 In relation to irrigation and drainage, government could give priority to supporting completion of on-going work and rehabilitation of districts and units, coupled with the establishment of WUOs responsible for the operation and maintenance of the works. The WUOS would require a strong government-supported technical assistance program to improve their chances of developing a sustainable infrastructure management system. Additionally, the GOM could support the development of small irrigation infrastructure in sites where there is producers' interest in some form of cost- sharing agreements, and technical and economical feasibility. Institutional development 9.57 Public decentralization needs further support in order to consolidate the mostly normative role that federal bodies should have and the executive role expected of state, municipal and communal agricultural institutions. Furthermore, there is a strong need to improve interinstitutional coordination  162 at state level. The lack of regional coordination could be tackled through the creation of a regional board, or a similar type of forum, for discussion of relevant state policies and programs. Eventually, a technical unit subordinated to this board could be made responsible for preparing recommendations adapted to the study area. 9.58 Veracruz, due to the dynamic role displayed by its Secretariat of Agriculture and by federal delegations in the state, is likely to be a prime target for further development of state, municipal and communal capacities needed to support the development of sustainable agriculture. Crop rehabilitation and diversification 9.59 Government could support specific rehabilitation and diversification programs in the area which would be based on four pillars: more detailed agroecological zoning studies, availability of technologies and technology transfer (see above), an expansion of "shared risk" credit schemes to include demonstration and pioneering commercial scale production, and environmental considerations. It would be critical for these programs to finance part of the research on annual and perennial tropical crops and agroforestry which is required for the sustainable development of the zone. 9.60 It is sensible to suppose that a coffee and cocoa rehabilitation program should be supported by the GOM. Both crops are mostly in the hands of small farmers and have a significant social importance, in addition to contributing to the stability of tropical environment. Present experience with PRONASOL's coffee program, specially in Oaxaca, could provide basic elements to be adjusted for future rehabilitation and diversification activities. Further reforms in sugarcane 9.61 The social security payments to cane workers have been put on a more equitable basis with other labor sectors. A sunset provision should be put in the next Cane Decree to shift the social security payments for those workers on the same basis as other agricultural labor . (a) Uncompetitive mills. The Government has classified mills according to their degree of competitiveness. Some of the least competitive privatized mills have high unpaid, and perhaps unpayable debts, leading to further pressure on the Government for support to keep the mills open. The Government should consider the option of closing the mills, rather than keeping them open indefinitely and thereby losing money indefmitely, giving while giving generous severance payments to the workers. (b) Research and extension. The liquidation of IMPA has eliminated public research capacity in sugar cane. In principle, the sugar mills have agreed to assume the responsibility for sugar cane research and extension, but have not done so. (c) Economic incentives for cane production. The producer price of cane in Mexico will be above the world price and below the US price. This will give a generous incentive to cane production and is, any event, a higher rate of protection than that during the period when Mexico was a competitive exporter. Therefore, the post- NAFfA price of cane is more than sufficient incentive for cane production and the Government should give no additional support to cane production or milling.  163 Information requirements 9.62 The severe poverty and social inequalities of the tropics imply a continuing demand for public interventions at the federal, state, and local levels in the Mexican tropics. Many of those interventions have, until now, been based on incomplete or misleading information. Additional studies are therefore needed for better policy affecting agricultural growth and social development in the tropics. Such studies include: (a) improvement of baseline data, notably on land use, employment, and the environment, including means toXsize and mechanisms for future Governmental support.  164 10. SOIL CONSERVATION 10.1 Soil and moisture losses caused by erosion can lower crop yields, induce inappropriately short fallow periods, or lead to of encroachment of crops onto lands unsuitable for farming. With continuing erosion, farmland may lose all of its top soil, and no longer be suitable for crops. Though degraded land can be used for grazing, this sometimes exacerbates erosion by further depriving land of vegetative cover. Agricultural encroachment onto forest lands may reduce the availability of fuel wood and timber, and the wind protection and moisture-storing capacity of the trees. 10.2 An important risk in the strategy to develop agriculture in the center and south of the country is the greater erosion that can be caused by more intensive farming practices. While much information is available on the extent of soil erosion, less is known about the economic losses associated with erosion, about the costs and benefits of soil conservation measures, or the policy options available to limit erosion and to promote conservation. No comprehensive analysis has been done of the policy options available to the government to reduce erosion. This chapter analyses the economic losses to erosion and of the costs of benefits of erosion control; it then discusses the rationale for government action, if any, in combatting erosion. I. General Patterns or Land Degradation Caused by Erosion 10.3 Table 10.1 shows a recent estimate of the land area affected by wind and water erosion, as estimated by an expert panel from background information (The Global Assessment of Soil Degradation - GLASOD; Anaya-Garduno et aI., 1989). The GLASOD classification of areas according to erosion hazard at an original scale of 1:7 500 000 examined, among other questions, the type (water, wind or chemical), and degree (slight, moderate or severe) of erosion for major landforms of Mexico. This information is not detailed enough to permit an evaluation of productivity losses from land degradation, but it does give an overall impression of degradation. 10.4 Erosion damage from wind and water has been presented in Table 10.1 by estimates of how much of the affected land is chiefly suitable for irrigated crops, rainfed crops, livestock, or forestry. The extent of erosion is defined in five classes: 1 percent to 5 percent of the land affected by slight, moderate, or severe erosion; 6 percent to 10 percent; 11 percent to 24 percent; 25 percent to 50 percent; and more than 50 percent. Water erosion affects about 86 million hectares compared to 21.3 million hectares affected by wind. The two combined are a significant share of the national degraded area. 10.5 The GLASOD classification indicates that the principal land type affected by water erosion is rainfed crop land. Rainfed land is about 54 percent of all land surveyed (58.5 million hectares of the 107.5 million hectares identified as suitable for crops, livestock, or forestry). Rainfed land is more likely to show signs of moderate (19.4 million hectares of 53.8 million hectares classed as moderately eroded) or severe erosion (35.9 million hectares of 44.9 million hectares classed as severely eroded). Most of the land suitable for irrigated agriculture, livestock production, or forestry has moderate erosion. The extent of moderate erosion, the most frequent class on those three land types, is greatest on irrigated land.  165 Table 10.1 Land Degradation Caused by Soil Erosion from Wind and Water (millions of hectares) Atrected W'md Erosion Water Erosion All Erosion land (~) Slight Moderate Severe Slight Moderate Severe Slight Moderate Severe Rainfed land 1-5 0.0 1.7 0.0 0.0 1.3 16.8 0.0 2.9 16.8 6-10 0.0 4.6 0.0 2.2 2.8 0.0 2.2 7.5 0.0 11-25 0.0 0.8 2.3 0.0 7.1 6.3 0.0 7.9 8.6 25-50 0.0 0.0 0.0 1.0 0.0 3.5 1.0 0.0 3.5 >50 0.0 0.0 0.0 0.0 1.1 7.0 0.0 1.1 7.0 Total 0.0 7.1 2.3 3.2 12.3 33.6 3.2 19.4 35.9 Irriaated land 1-5 4.6 0.0 0.0 0.0 0.0 0.0 4.6 0.0 0.0 6-10 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 11-25 0.0 0.0 0.0 0.0 2.5 0.0 0.0 2.5 0.0 25-50 0.0 0.0 0.0 0.0 1.7 0.0 0.0 1.7 0.0 >50 0.0 0.0 0.0 0.0 2.7 2.7 0.0 2.7 2.7 Total 4.6 0.0 0.0 0.0 6.9 2.7 4.6 6.9 2.7 Livestock land 1-5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 6-10 0.0 0.0 0.0 0.0 0.9 0.0 0.0 0.9 0.0 11-25 0.0 3.9 0.0 0.0 0.0 0.0 0.0 3.9 0.0 25-50 0.0 1.2 0.0 0.0 0.7 0.0 0.0 1.8 0.0 >50 0.0 0.0 0.9 0.0 0.0 0.0 0.0 0.0 0.9 Total 0.0 5.1 0.9 0.0 1.6 0.0 0.0 6.7 0.9 Forestry land 1-5 0.0 0.0 0.0 0.0 8.4 0.0 0.0 8.4 0.0 6-10 0.0 0.0 0.0 1.0 1.4 3.7 1.0 1.4 3.7 11-25 0.0 1.4 0.0 0.0 8.9 0.0 0.0 10.2 0.0 25-50 0.0 0.0 0.0 0.0 0.7 0.0 0.0 0.7 0.0 >50 0.0 0.0 0.0 0.0 0.0 1.7 0.0 0.0 1.7 Total 0.0 1.4 0.0 1.0 19.4 5.4 1.0 20.7 5.4 AU land 1-5 4.6 1.7 0.0 0.0 9.6 16.8 4.6 11.3 16.8 6-10 0.0 4.6 0.0 3.2 5.2 3.7 3.2 9.8 3.7 11-25 0.0 6.1 2.3 0.0 18.5 6.3 0.0 24.6 8.6 25-50 0.0 1.2 0.0 1.0 3.1 3.5 1.0 4.3 3.5 >50 0.0 0.0 0.9 0.0 3.7 11.4 0.0 3.7 12.3 Total 4.6 13.6 3.1 4.2 40.2 41.8 8.8 53.8 44.9 Note: The table may contain rounding errors. Source: Anaya-Garduno et a1., 1989; and World Bank, LA2NR Mexico Agriculture Database.  166 II. Patterns of Agricultural Soil Erosion 10.6 Analysis of patterns of agricultural soil erosion should provide a consensus on the most- severely affected areas, the quantities of soil loss, and the practices that prevent loss. Analyses of agricultural soil erosion in Mexico usually use the FAO-Conservation Foundation method (1954). The method defmes five classes of erosion and corresponding mean productivity losses (Table 10.2): (a) Class A has no manifest erosion, with an overall loss of less than 25 percent of topsoil, and with no more than 10 percent of the area with B or C classes of erosion. This class has an estimated mean physical productivity loss of 17 percent compared to the situation without erosion; (b) Class AlB has slight erosion, with an overall loss of less than 25 percent of topsoil, but having between 10 percent and 25 percent of the area in classes B or C. This class has an estimated productivity loss of 22 percent; (c) Class B has moderate erosion; between 25 percent and 75 percent of topsoil lost, and up to 10 percent of the total area in classes A or C. This class has an estimated productivity loss of 37 percent; (d) Class B/C has severe erosion; between 25 percent and 75 percent of topsoil lost, and with 10 percent to 25 percent of the total area in classes of erosion A or C. This class has an estimated productivity loss of 38 percent; and (e) Class C has very severe erosion; more than 75 percent of the topsoil lost, though 25 percent of the area may be in classes A or B. This class has an estimated productivity loss of 75 percent. 10.7 The national weighted average estimate of the loss of productive potential is 35 percent (SARH, 1987a), based on these five classes. to.8 Table 10.2 summarizes information on erosion by state in 1982.'" "Rapid" to "severe" erosion is common. A rough grouping by the main agroclimate in each state shows little difference among the main agroclimates in terms of erosion. State date reported by Maas and Garcia-Oliva (1990) and by SARH (1987a) indicates that the tropics have the most severe and widespread agricultural erosion, with average losses of 3-4 mtlha in the dry and humid tropics and average losses of 1-2 mtlha in the temperate and semiarid areas. 10.9 These state data cannot serve to estimate the quantity of soil loss per unit of land. It is known that soil loss depends on rainfall amount and intensity, soil erodibility, length and steepness of slope, vegetative cover and land use practices. This relation can be estimated statistically with the universal soil loss equation (USLE), in which erosion is a function of the mentioned factors. Data from small experimental watersheds, in which conditions vary widely depending on the plot size, give differing results when analyzed with the USLE. For 10m2 plots, the annual loss has been estimated to 4.7 tlha; for 50 m2 plots, from 0.9 to 4.7 tlha; and for 7,000 r.rr plots 0.45 tlha (Slater, 1991). 93. Results of a 1962 survey are not presented separately because not all stales were I'C--surveyed in 1982.  167 Table 10.2 Severity of Soil Erosion in 1982 (Thousands of hectares) State No Manifest Light Moderate Severe Very Severe Eroded Total Baja California N. 1,003 1146 2364 1,791 860 7,163 213 Baja California S. 870 1015 2174 2,029 1,159 7,247 59 Chihuahua 4,427 4250 13188 2,109 681 24,561 973 Coahuila 301 1654 4662 6.911 1.504 14,731 292 Durango 3,459 247 4817 1,235 2,594 8,893 595 Nayarit 818 1219 607 76 7 1.909 303 Nuevo Leon 260 1693 1497 2.731 826 6,747 240 Sinaloa 1.170 1170 1111 1.228 1,170 4,679 1,116 Sonora 3,286 2373 4016 3,103 5,477 14,969 673 Tamaulipas 2,229 796 2388 1,353 1,194 5,731 1,345 Aguascalientes 27 82 192 110 137 521 128 Colima 125 21 88 234 62 405 155 Guanojuato 1,026 506 874 383 218 1,981 1,073 Hidalgo 501 63 397 522 605 1,587 604 Jalisco 1,864 81 1459 4,134 467 6,141 1,289 Mexico 516 919 611 28 7 1,565 838 Michoacan 961 120 1382 3,245 300 5,047 997 Morews 9 293 170 13 0 476 132 Puebla 102 1923 851 357 131 3,262 969 Queretaro 138 23 367 207 413 1,010 160 San Luis Potosi 1,145 1666 2811 575 32 5,084 810 T'laxcala 17 71 209 77 10 367 244 Zacatecas 367 2057 367 3,085 1,469 6,978 1,173 Campeche 1,630 2038 917 306 214 3,475 119 Chiapas 4,557 2123 477 64 0 2,664 1,196 Guerrero 2,449 127 1676 1,354 839 3,996 685 Oaxaca 1,413 1413 1894 1,884 2,826 8,017 922 Quintana Roo 1,611 1915 504 252 756 3,427 89 Tabasco 760 887 380 127 380 1,774 213 Veracruz 3.604 3462 96 2 1 3,561 1,182 Yucatan 578 193 1194 193 1,694 3,274 595 .... .vitv losS'" 17% 22% 37% 38% 75% 35% n.8. NOTE: The total eroded area does not equal the lurn of the erolion classes because of rounding errors. states are ordered using the classification in Table 1.1 of Chapter 1." The productivity loll is with respect to the no erosion situation. SOURCE: SARH, 19878 for erosion data and World Bank. LA2NR Mexican Agriculture Database.  168 10.10 Combining the information from the state level and from experiments should give a rough national estimate of the quantity of soil lost to erosion. Available estimates diverge greatly, however, despite their importance in setting national priorities for soil conservation measures. Maas and Garcia-Oliva (1990, at 29) report four calculations of the share of the national area affected by rapid erosion to range from 64 percent to 100 percent. They also observe that two such estimates differ in evaluations of the soil lost to erosion. One calculation of the average national loss is 2.7 mt per ha, with a minimum of 0.2 mt per hectare in the north and a maximum of 7.4 in the seriously affected Pacific Coast. Another study, based on the same data analyzed with a different method, found an average of 47 mt, with a range of 3 to 126 mt per ha. 10.11 The Mexican evidence may be biased downward because many experiments have been done on plots with slopes less than 15 percent. Studies from sites with steeper slopes (mentioned in Maas and Garcia-Oliva, 1990) have shown soil losses of between 30 and 130 mt per hectare, with some values as high as 150, 300, and 500 mt found in mountainous areas with very steep slopes. ill. Productivity Impact of Erosion 10.12 The on-farm costs of erosion can be classified as yield reductions, higher input costs and changes to less productive land uses including the extreme case of total topsoil removal. These costs have to be defmed with respect to the tolerance level of soil erosion. The latter has been defined as "the maximum rate of annual soil erosion that may occur and still permit a high level of crop productivity to be obtained economically and indefinitely" (Sinner, 1990, at 11).114 At the tolerance level, it is assumed that the net soil loss is zero and the yield costs of erosion are zero. Though the tolerance level is not an economic optimum because it does not equalize the marginal benefits and costs of soil conservation, it has agronomic significance and the practical virtue of being a target for conservation programs. The annual tolerance level has been estimated to be between 2.5 and 12.4 mt per hectare annually, depending on soil depth and texture. An average of 5 mt per hectare per year is used in this study. 10.13 Crop Yield Reductions. Crop yield losses to erosion result from reduced water-holding capacity and from decreased soil depth and nutrient load. Only a portion of the water potentially available for plant growth can be retained where surface runoff occurs. Water-holding capacity is further restricted at reduced soil depth. The diminished water-holding capacity of the soil and the loss of nutrients occasioned by decreased depth both reduce crop yields. Estimates in the United States for maize, wheat, and soybeans yield losses vary from 1.5 percent to 2.0 percent annually (Crosson, 1983). Erosion causes significant yield losses (Doolette and Magrath, 1990) for several Asian countries; Lal (1987) reported declining yields of 2 percent to 5 percent per year on hillsides in southern Nigeria while Nye and Greenland (1960) reported yield declines of as much as 83 percent for sorghum and 64 percent for pearl millet, in subhumid northern Nigeria. No corresponding national estimate exists for Mexico. 10.14 Increases in Production Costs. Increases in production costs due to erosion have several components. Farmers might apply more mineral fertilizers to replace nutrients lost to erosion, or more irrigation water to replace higher runoff caused by erosion. Crop management may become more costly due to the increased variability of soils that forces farmers to apply discrete management practices on different plots of land. Colaciccio and Setia (1986) estimated the additional costs of 94. Sinner does not define a -high" level, but presumably he means one that provides a competitive economic return to the farmen' resources.  169 mineral fertilizers to replace nutrients lost to erosion by: (a) calculating soil loss to erosion; (b) estimating the average nutrients in the soil lost; (c) calculating the nutrients lost to erosion; and (d) valuing those nutrients at the economic prices of replacement fertilizers. This has not been done here for lack of useful information on the quantity of nutrients lost to erosion. 10.15 Land Use Changes. A land use change is defined as a switch to a less productive use due to erosion, for example as shifting to a less-productive crop which can tolerate erosion. Productivity declines may lead to shorter fallow periods; these reductions in vegetative cover and nutritive value of grazing areas may lead to overgrazing and further land degradation, the acceleration of gully formation, or the destruction of forest and woodland faster than they can regenerate. Because we have no information on these shifts in Mexico, the economic analysis excludes their potential costs. 10.16 Off-farm Effects. Off-farm effects include sedimentation of waterways, dams and other collection sites. The sedimentation shortens the lives of reservoirs and irrigation systems. There is conflicting opinion about the economic costs of off-farm sedimentation. In the United States, Crosson (1983) argues that off-farm losses outweigh on-farm losses. A World Bank review (Doolette and Magrath, 1990) of Asian countries found that off-farm losses were probably less important than the on-farm costs because of the small proportion of watersheds with infrastructure compared to the total area affected by erosion. In Mexico, it has been calculated that 70 percent of eroded soil exits to the sea and 30 percent is deposited in lakes, rivers and structures. IV. Economic AnaJysis of Soil Conservation 10.17 Modest structural works and/or minimum tillage with maintenance of residual organic matter can reduce physical soil loss to acceptable levels. However, little is known about the economic benefits to society and the financial benefits to the farmer of using such soil conservation practices. The following sections review some of the management techniques available for arresting the on-farm effects of soil erosion in Mexico. A. Soil Conservation Techniques 10.18 Farmers have many traditional methods to arrest erosion. Wilken (1987) describes some of them in a semi-arid area of l1axcala state with severe erosion. They include: (a) check dams, built with earth and stone to block the flow of water and sediment in erosion gullies and then cultivated by planting the land formed behind the dam; (b) sloping terraces, built with stone and earth borders to control erosion and runoff overflow adjacent to ditches which conduct excess water out of the field; and (c) bench or flat terraces, in which flat cropland is carved out of steep slopes and cultivated. 10.19 An FAO report (Slater, 1991) further describes practices in four rural development districts in the states of Mexico, Michoacan, Nuevo Leon, and San Luis Potosi. They include contour banks with ditches, check dams, stone contour dams, contour plowing, land smoothing, minimum tillage, tie ridging, and such cropping practices as composting, manuring. and application of crop residue (see Table 10.3). Farmers practice these methods sporadically.  170 10.20 Considerable knowledge exists about soil and water conservation practices from previous work by the General Directorate of Soil and Water Conservation (DGCSA) and the continuing work of INIFAP, CNA, and Mexican universities. Starting in the 196Os, the Colegio de Posgrados initiated research into soil erosion and run-off, cultivation effects and the impact of structures on soil erosion. Most of this work was performed in the central highlands with a few discontinuous experiments in other agroecological areas. After the disbandment of the DGCSA in the mid-1980s, the Colegio de Posgraduados at Chapingo (CPCh), the Universidad Aut6noma de M6xico (UNAM) and a few other universities were the only active parties in soil conservation research even though INIFAP had as its mandate the "improvement, use and conservation of soil and water for agricultural use. " B. Effects of Soil Conservation Techniques 10.21 It is known that increasing the vegetative cover of land tends to reduce erosion, though there is wide variation in estimates of the magnitude of the effect. A SARH review (1987a) of the relative technical effectiveness of vegetative covers and mechanical practices concludes that a crop providing adequate soil cover, plus appropriate crop residue management, would reduce erosion losses to tolerable levels, even where the soil erosive potential is very high. The same study has estimated the value of vegetative cover in reducing erosion to range from 1 (the highest vaiue on bare soil) to 0.001 for a pine forest. A maize or sorghum crop would have a value of 0.1 to 0.9, or between 90 percent and 10 percent less than bare soil depending on type of soil and the crop management practices used; maize and sorghum are probably much less technically effective than pastures (0.001) or such permanent crops as coffee and oil palm (0.1 to 0.3). A general conclusion from the Bank's Asia study (Doolette and Magrath, 1990) was that vegetative covers were more technically effective than works or mechanical practices, but this is not clearly confirmed by the evidence given in tables 10.5 and 10.6 about soil erodibility indices with different conservation techniques. C. Cost-Benefit Analysis 10.22 The economic analysis concentrates on three major crops which are grown largely under rainfed conditions: maize, beans, and sorghum. In 1992 maize was harvested on 7.3 million hectares, beans on 1.37 million hectares, and rainfed sorghum on about 1.38 million hectares. The area of the three crops at serious risk of erosion in the 21 states is about 4.32 million hectares. 10.23 The decision to concentrate on these annual crops for the rainfed areas alone was taken for several reasons. Irrigated crops typically have less erosion because they are grown on level fields; in terms of the USLE, their slope factor is very small. Irrigated crop land is often more thoroughly covered throughout the year because more than one crop is grown; in terms of the USLE, their cover factor is low. The irrigated areas of Mexico are in more arid sites where rainfall is both lower and less intense than in the tropics, and are thus exposed to less frequent and less erosive showers. Rainfed permanent crops are often trees, shrubs, sugar cane or forages which provide good annual soil cover. thereby incurring less erosion. Moreover, such crops are often sown in drier areas like those where irrigated field crops are grown, and their erosion risk is less serious than in the tropics. The economic impact of soil erosion on these three crops in Mexico has not been well quantified partly because of methodological and data questions. Those questions are reviewed briefly below.  171 Table 10.3 Soil Conservation Techniques Type of Practice Best Environment Costs Over Time STRUCTURES vegetative barriers subhumid;humid moderate at outset bunds arid/semi-arid high at outset, then lower annual upkeep stone contours adaptable to all high at outset, then lower annual upkeep drainage ditches subhumid;humid high at outset, then lower annual upkeep flat terraces adaptable to all high at outset, then lower annual upkeep backslope terraces adaptable to all high at outset, then lower annual upkeep tanks arid/semi-arid high at outset, then lower annual upkeep check dams arid/semi-arid high at outset, then lower annual upkeep windbreaks arid/semi-arid high at outset, then lower annual upkeep TILLAGE PRACTICES contour plowing adaptable to all uniform tie-ridging arid/semi-arid uniform subsoiling adaptable to all uniform stone removal adaptable to all uniform minimum tillage unknown uniform CULTIVATION PRACTICES animal manuring arid/semi-arid uniform green manuring humid uniform crop residue use widely adaptable uniform crop rotation widely adaptable uniform forage crops arid/semi-arid variable SOURCE: Slater, 1991. 1. Methodological and Data Issues 10.24 The major methodological assumptions are as follows: (a) The estimate of a discount rate to weight future costs and benefits affects the net returns to soil conservation techniques. The real annual discount rate used here is 12 percent, which has been used in recent project analysis in Mexico by the Bank; (b) The use of fmancial or economic prices affects the net benefits of erosion control. Economic prices are used because they are the social opportunity costs of resources  172 used. The single exception is the discount rate applied to future costs and benefits; sensitivity analysis is presented comparing (higher) private rates with (lower) economic rates; (c) The number of years after which erosion begins to diminish crop yields affects returns. It is assumed that average erosion occurs every year; hence crop yields decrease continuously and the number of years before erosion is felt is zero. This assumption increases the costs of erosion by supposing them to occur earlier than if erosion manifested itself after a lag; and (d) The number of years until soil conservation techniques begin to reduce erosion and the number of years until those techniques have their full impact are related. It is assumed that conservation techniques have half of their maximum impact in the first year after they are practiced and their full impact in the second. This lag is due to the assumption that farmers would need time to learn fully the management of the new techniques. This assumption reduced the benefits of erosion control by supposing them to occur later, causing their discounted value to be lower; (e) The permanence of soil losses to erosion means the degree to which land lost to erosion can be regenerated. If soil losses were completely restored by soil regeneration, then they would not be permanent and the net annual loss (gross - regeneration) is zero. Since regeneration is not full, the net loss is positive (Le., gross > regeneration) and permanent; and 10.25 The data questions are: (a) On-farm crop yield losses to erosion: The yield costs were modeled from information provided by Mexican researchers or from external sources where national information was unavailable; and (b) The incremental costs of soil conservation techniques were derived from farm management information provided by Mexican specialists, as summarized in Table 10.4. These costs are calculated as a function of the quantity of soil lost to erosion. 10.26 Several important issues had to be excluded for lack of information. These were the off-farm benefits of soil conservation; this results in an underestimate of the total social benefits of erosion control. A second class of issues was the cropping pattern effects of erosion; it is not known what effects this would have on the total estimated economic cost of soil erosion. A third is fertilizer nutrient cost of erosion; excluding this results in an underestimate of the social costs of erosion and hence an underestimate of the benefits to any given conservation investment. 2. Plan of the Analysis 10.27 The economic analysis reviews evidence from eight states presenting typical tropical farming situations (Campeche, Chiapas, Guerrero, Oaxaca, Quintana Roo, Tabasco, Veracruz, and Yucatan) and 13 central rainfed states (Aguascalientes, Colima, Guanajuato, Hidalgo, Jalisco, Mexico, Michoacan, Morelos, Puebla, Queretaro, San Luis Potosi, TIaxcala, and Zacatecas). The national maize area sown in the 1992 spring/summer cropping season was about 6.31 million hectares, of which 3.38 million were sown in the 13 central rainfed states and 2.17 million hectares in the  173 Table 10.4 Costs of Soil Conservation Techniques and Soil Erodibility Indices Average Annual Costs Soil Erodibility Index, Type and name of technique (US$ per hectare) Relative to Bare Soil x 100 TROPICS Structures Check dams 15 50 Stone barriers 66 33 Cultivation Practices Contour furrows 26 20 No-tillage 34 20 Minimum tillage 54 33 Cropping Practices Mulching 10 33 Crop residue management 28 20 Spaced furrows 33 20 Planting in rows 54 33 Manure incorporation 65 13 Cover crops 85 47 Green manuring 201 13 CENTRAL RAlNFED AREAS Structures Contour banks with ditches 230/297/407/4921541 20 Check dams 221/2351332 50 Stone contours 179 50 Cultivation Practices Contour plowing 15117/18 20 Land smoothing 34 33 Minimum tillage 3/68/122 33 Subsoiling 17612251334/375 33 Tie ridging 8/59/67 33 Cropping Practices Compost 523 33 Cover crop 47164 50 Crop residues 64/69/81 20 Manuring 468 33 NOTE: Different costs for the same technique are variations for different crops, production locations, soil types. and land slopes. Costs - derived from CNA data and converted to US$ at US$1 =3.1 new Mexican pesos- include fixed and variable costs. A technique having an index of 100 would have the same quantity of annual soil erosion as bare soil. SOURCE: Costs are from CNA and Slater, 1991. The soil erodibility index is derived from SARH (1987a) and Maas and Garcia-Oliva (1990). 8 tropical states. Another 1.42 million hectares were sown in that season to beans, of which 0.93 million were in the central rainfed states and 0.14 million in the tropical states. About 0.41 million hectares were sown to sorghum of which 0.25 million were in the central rainfed states and 0.01 million in the tropical states. Because of uncertainties in the available information, the  174 analysis is limited to parts of the 21 states where erosion is manifest. Based on the data in Table 10.2, the areas in the three crops with greater than -no manifest' erosion are assumed to be proportional to the total area in each state with greater than -no manifest- erosion. The resulting areas are shown in Table 10.5. 10.28 The following steps were then taken for each situation: (a) The impact of the average physical soil losses on yields of maize, beans, and sorghum was calculated; (b) The annual economic value of the yield losses was estimated; (c) The average reduction in soil loss was calculated for each technique; (d) The costs of soil conservation were calculated for each technique; (e) The present value of net benefits for each combination of zone, state, crop, and technique were calculated to perpetuity; (1) The net benefits were subjected to sensitivity analysis by varying the private discount rate, the effect of erosion control on crop yield, the target level of soil loss; and (g) Estimates of the net present value (NPV) of benefits to soil conservation on maize, beans, and sorghum were aggregated to the national level by projecting them over the eroded areas in each crop in the tropical and central rainfed states studied. 3. Costs of Erosion Control Techniques 10.29 The costs of tillage and cultivation practices differ over time. Those of structures are highest when an initial outlay is made and they fall as only routine maintenance is done. Those of cropping practices are more or less the same in each year of the investment. The higher initial cost of structures would make them less profitable than cropping practices, other things being equal. Costs of seeds, plant cultivars, pesticides and agrochemicals are assumed not to be affected by the soil conservation control techniques and thus do not enter into the economic analysis of the incremental conservation investments. 10.30 Incremental costs of supervision and extension related to soil conservation differ among programs for combatting erosion. For supervising structures or cropping practices, such costs are estimated to be US$6 per hectare per year, based on the current costs of extension programs. The incremental supervision costs of a land reserve program are assumed to be SO percent of those of supervising field techniques, or US$3 per hectare annually, because verifying participation in the reserve program can be done more quicldy than teaching soil· conservation techniques and would need to be done less often. 10.31 The supervision costs of extending new soil conservation methods are assumed to be higher than normal extension costs. It is further assumed that cost per ton of erosion saved incurs increasing costs as farmers must work progressively harder to achieve those savings  175 Table 10.5 Values of Key Parameters for Economic Analysis Area cultivated of a crop, million hectares Rainfed spring/summer maize (32 states), total area 6.31 Tropical states (8 states) with erosion risk'" 1.41 Central rainfed states (13 states) with erosion risk'" 2.84 Rainfed spring/summer beans (32 states), total area 1.42 Tropical states (5 states) with erosion risk 0.09 Central rainfed states (8 states) with erosion risk 0.78 Rainfed spring/summer sorghum. (32 states), total area 0.41 Tropical states (5 states) with erosion risk 0.01 Central rainfed states (8 states) with erosion risk 0.21 Crop yields, mt hectare-l Maize, tropical 1.50 Maize, central rainfed 2.00 Beans, tropical 0.50 Beans, central rainfed 0.75 Sorghum, tropical 1.50 Sorghum, central rainfed 2.00 Crop prices, US$ mr! Maize 140 Beans 300 Sorghum. 125 Fertilizer prices, US$ mrl Nitrogen 380 Phosphorus 280 Discount rates, annual percent Economic 12 Financial 15/20 Supervision costs, US$ hectare,1 Soil conservation techniques Land reserve program Erosion and soil conservation parameters Loss of crop yield with erosion, annual percent maximum percent loss 20.0 Soil conservation productivity effect, gain in crop yield. total percent of yield with erosion low value 10.0 high value 40.0 NOTES: 01 65 percent of the area in the tropical states baa erosion greater than the "no manifest" category (Table 10.2). bI84 percent of the area in the central rainfed states baa erosion greater than the "no manifest" category (Table 10.2). SOURCE: Crop production areas and yields from SARR, Boletin Mensual, various years. World prices from IMP, InrernatiollfJl Financial Statistics, various years. Supervision costs are from Bank project supervision reports. Erosion and soil conservation parameters are from Slater, 1991 and Maas and Garcia-Oliva, 1990.  176 4. Results ror the Tropics 10.32 Tables 10.5 and 10.6 give the average annual costs and erosion indices of various techniques, ranging from 1.0 percent to 193 percent of the average gross value of rainfed maize production (a CIF price of US$I40 per ton times an average yield of 2 tons per hectare). 10.33 The tropics are divided into two distinct sub-regions, the dry tropics and the humid/subhumid tropics. The dry tropics cover 17 percent of the country, mainly in the states of lalisco and Oaxaca on the Pacific Coast and the states of Tamaulipas and Veracruz on the Gulf coast. The dry tropics have between 500 and 1,000 mm annual rainfall and a growing season of between 150 and 270 days. The humid/subhumid tropics cover about 20 percent of the country, south of Oaxaca city to Guatemala, parts of the Pacific states of Nayarit, Colima, Michoacan, Chiapas, and Guerrero; and parts of the Gulf Coast, from Veracruz to Quintana Roo with the exceptions of northern and western Yucatan. They are defined as having more than 1,000 millimeters annual rainfall in a crop growing period of 150 days or more. The total maize, beans, and sorghum areas in the five states analyzed are 1.7, 0.2, and 0.1 million hectares respectively, with SO percent having erosion greater than the "no manifest" category (Table 10.2). 10.34 The FAO classification (Table 10.2) defines about 23 percent of the total land area in the 13 mainly tropical states as rapidly or severely eroded. The tropics are subject to more rapid erosion than temperate zones because of higher and more intense rainfall, less fertile and shallower soils, and, in some cases, steeper slopes. Some tropical land management practices can also accelerate land degradation. Slash and bum agriculture leads to rapid degradation of cleared land. Burning vegetation, while it converts some nutrients into forms that can enter the soil and destroys weeds and pests, exposes the soil to physical damage from intense showers while destroying organic matter and microorganisms essential to natural soil processes. Mechanical clearing can damage the soil by compaction and removal of the thin top layer. 10.35 Table 10.6 shows the results for annual erosion losses of 5 mt and 1 mt per hectare. The values are the average benefits per hectare of the profitable techniques multiplied by the areas of each crop. Erosion control, even in the case when the "m" variable is assumed to be relatively low (at 10 percent) and the social discount rate high (12 percent), would be profitable in the tropical states for 34 techniques among the 3 crops examined, giving an NPV of US$103 million. The profitable techniques would nearly always be cropping or cultivation practices. 5. Results ror the Central Rainred Areas 10.36 The temperate highlands cover about 12 percent of the country in a central belt about 1200 meters above sea level from Aguascalientes in the north to Oaxaca City in the south. This area corresponds approximately to the isoyhets of 500-1000 mm annual rainfall and between 100-150 day growing period. The rainy season usually starts mid-May to early lune and lasts from three to four months. Year-to-year fluctuations in annual precipitation are considerable and within-season distribution is subject to periods of drought as well. The semiarid and arid areas occupy over 50 percent of the country. The southern limit is the 500-1000 mm isoyhet and the 100 day growing period. Much of the 13 states have wide areas classed as severely eroded (Table 10.2).  177 Table 10.6 Net Economic Benefits of Soil Conservation Techniques at Sodal Discount Rate of 12% (Net Present Value in USS Millions) Low Soil Conservation mgh Soil Conservation Productivity Effect Productivity Effect (10% maximum) (40% maximum) Target soil loss Target soil loss Target soil loss Target soil loss of5 mt of 1 mt of5mt of 1 mt TropiCDl Stotes Total 103 (34)'" 115 (29) 604 (77) 587 (74) Maize 92 (25) 104 (20) 579 (4O) 560 (40) Beans 11 (3) 10 (3) 22 (21) 24 (18) Sorghum 1 (6) 1 (6) 4 (16) 3 (16) Central Rainfed Stotes Total 675 (45) 645 (45) 2,170 (92) 2,135 (91) Maize 609 (4O) 585 (40) 1,731 (82) 1,673 (82) Beans 66 (5) 60 (5) 325 (9) 351 (8) Sorghum 0(0) 0(0) 114 (1) 110 (1) NOTES: '" Numben of profitable cases - • case is • combination of ltate, crop, and technique - are in parentheses. SOURCE: Generated by model from parameten in Table 10.5. 10.37 Farmers in the central and northern highlands use few soil and water conservation practices. The only widespread technique is some incorporation of animal manure, although this is not universal. Most land preparation is done with mechanical discs, frequently up and down the slope, not on the contour. Most producers do not use simple on-farm structures. There is the almost ubiquitous removal of maize stover from fields, exacerbating the nutrient loss to erosion and decreasing soil cover. Because of the severe risks of drought, frost and hail, maize and other crop cultivars with shorter growing seasons have been difficult to develop, and in consequence, little crop rotation is practiced and virtually no cover-crops are grown for soil protection. 10.38 SARH and FAO have analyzed four DDRs in the temperate highlands with severe erosion. The four are Ameca (Jalisco state), Atlacomulco (Mexico), Montemorelos (Nuevo Leon), and Morelia (Michoacan). Methods to control erosion in these districts would apply widely to the temperate highlands because the districts were chosen to be representative of maize farming in much of those Mexican highlands. Farmers generally know something of these techniques and could apply them at low incremental costs. Only three of the techniques - stone contour walls, contour banks with ditches. and check dams - require works. The others are crop management practices, requiring neither new equipment nor great modifications in farming practices. Development of such techniques to make them suitable to local farmers would require no more than extension to promote the most profitable techniques.  178 10.39 Table 10.6 displays the results for rainfed production of the three crops in the central rainfed states. Erosion control leading to a 10 percent increase in crop yields over a base situation to perpetuity would be profitable for 45 cases, with an NPV of US$675 million. The situation would be somewhat better for maize and beans in the highland and semiarid states than in the tropical states, but not so for sorghum, for which no technique was profitable. 6. Sensitivity Analysis 10.40 The profitability estimates given here are subject to uncertainties because of some doubts about the underlying data. Because of gaps in information about production locations, crops, or farming techniques, we have to use sensitivity analysis to model their most probable impacts. Sensitivity analysis with respect to the effects of the productivity effect of soil conservation and of the target quantity of soil loss to erosion is shown in Table 10.6. Further sensitivity analysis with respect to the discount rate, and its interaction with the productivity impact of soil conservation, is in Table 10.7. Table 10.7 Effect of Discount Rate on Private Profitability at Target Soil Loss of 5 mt (Net Present Value in US$ Millions) Low Soil Conservation Effect High Soil Conservation Effect Discount Rate 15% 20% 15% 20% Tropical StiUes Total 106 60 455 310 Maize 98 55 434 295 Beans 8 6 19 13 Sorghum 1 0 3 2 I Central Rainfed StiUes I Total 428 304 1,633 1,112 Maize 436 269 1,304 875 Beans 42 35 243 180 Sorghum 0 0 86 57 SOURCE: Generated from model using parameters in Table 10.5. 10.41 Productivity impact of erosion control. The higher productivity impact of soil conservation would obviously raise the returns to investments. This part of the analysis is highly speculative, however. While the national average estimates of the impact of erosion on crop productivity are at least 37 percent in much of the country, the more detailed local evidence from Mexico, and the estimates from the United States. show lower overall effects. While it is therefore improbable that the average impact could be as high as 40 percent there are apparently many sites where the losses could exceed that. The economic returns to erosion control at such severely eroded sites would be highly positive, as shown in Table 10.7. That the impact could be so great on severely eroded lands argues for accurate targeting on the choice of sites to promote erosion control measures since overall returns depend so much on choice of site.  179 10.42 Target quantity of soil loss to erosion. Cutting the target level of soil loss to 1 mt, which is considered to be the lowest practical quantity with an annual crop, would have little impact on the numbers of profitable techniques and the average return for profitable techniques (Table 10.6), These results mean that producers could cut erosion below an average target quantity of soil loss with only modest external fmancial incentives.~ 10.43 Private discount rate. The initial real private discount rate was 15 percent or 25 percent greater than the real economic rate. Increasing the real private rate to 20 percent had a strong negative effect on the financial profitability of soil conservation. At a target soil loss of 5 mt and a soil conservation productivity effect of 10 percent, the NPV of financial benefits was US$534 million at 15 percent and US$364 million at 20 percent, a decline of about one-third 10.44 The calculation of possible subsidies depends on the profitability of available soil conservation techniques. The latter can be organized into three groups: (a) those profitable at the lower economic discount rate and the higher private rate (Group A); (b) those profitable at the economic rate but not at the private rate (Group B); and (c) those not profitable at the economic rate or the private rate (Group C). If farmers have adequate information, they would never adopt group "C" techniques because their private profitability is negative. Group "C· techniques would not justify a subsidy because their social profitability is also negative. 10.45 Group" A" techniques. If adequate information is available about group •A' techniques, farmers would adopt them because they are privately profitable. No subsidy would be necessary to induce farmers to conserve soil to the tolerance erosion loss; a subsidy to induce them to conserve to the lower minimum erosion loss would be required, however. It is the difference between private profitability at the tolerance loss (5 mt) and the minimum loss (1 mt). This subsidy is unaffected by any divergence between the private and social discount rates because it is calculated using only those techniques that are profitable at specified private discount rates (Table 10.8). 10.46 The NPV of the subsidy, for the low soil conservation productivity effect, is US$32 million at a private discount rate of 15 percent and USS15 million at a private discount rate of 20 percent. At a given conservation productivity effect, the subsidy falls from the lower to the higher discount rate because average profitability falls. At a given private discount rate. the subsidy rises from the lower to the higher conservation productivity effect because average profitability rises. 10.47 Group "B" techniques. The group "B" techniques - positive social profitability and negative private profitability - might justify a subsidy equal because farmers would not adopt practices with negative private profitability, but society would not benefit from the use of such (socially) profitable practices. There were few techniques in this category; most were in group" A" or group "C," but there may be sites at which Group "A" techniques cannot be used. This subsidy does depend on a possible divergence between private and social discount rates; in Table 10.8, it can be seen to rise sharply with the private discount rate and to fall from the weaker to the stronger conservation productivity effect. Because so few techniques would be in class "B." and because of the high subsidies needed to compensate for the difference between the private and social discount rates, it would not be worthwhile to subsidize particular soil conservation practices in this manner. 95. The NPV of benefits is hi&her at the lower tar&et erosion rate in lome instances. This is because the incremental costs of some minimum tilla&e techniques are ne&ative (i.e.• they lave costs).  180 Table 10.8 Hypothetical Subsidies and Extension Costs to Achieve Annual Erosion Loss or 1 mt per Hectare with Group "An or nB" Techniques (Net Present Value in US$ Million) Low Soil Conservation Eff==U mgh Soil Conservation Effect Group A Group B Group A Group B Discount Rate 15% 20% 15% 20% 15% 20% 15% 20% Tropical Stotes Total 8 6 90 lOS 16 12 0 Maize 8 6 89 104 16 11 0 0 Beans 0 0 0 0 0 1 0 0 Sorghum 0 0 1 1 0 0 0 0 Central Rainfed States Total 24 9 51 322 29 9 207 261 Maize 19 6 4 259 29 6 207 261 Beans 5 3 51 63 0 3 0 0 SorRhum 0 0 0 0 n 0 0 0 SOURCE: Generated from model using parameters in Table 10.5. D. Is a Land Reserve More Efficient Than Erosion Control Practices? 10.48 One alternative to subsidizing soil conservation techniques is to pay producers to leave land idle. This is called a land reserve program and it would be more efficient than soil conservation practices if it cost less per quantity of soil saved. There are no economic benefits to the land reserve program. The land reserve produces no benefits in higher crop yields or lower fertilizer costs land is idle. 96 The only benefits are the environmental ones of the soil saved from erosion; hence, a land reserve program and a program of technical changes can only be compared on their respective net costs per ton of erosion saved. 10.49 The cost per ton of erosion saved in the reserve is the NPV of the supervision costs plus that of the reserve payment to producers, divided by the quantity of erosion saved. The quantity of erosion saved is the difference in erosion on the uncultivated reserve and the average erosion on the same land under cultivation. The reserve payment to producers is arbitrarily set equal to 20 percent of the gross benefits (yield times price) of output. Because a reserve program produces no gross economic benefits, its net benefits will always be negative. The cost per ton of erosion saved by technical changes is the discounted present value of net benefits, divided by the quantity of erosion saved. 10.50 If group" A" techniques could be generally extended, then they would produce much higher positive net benefits than would a land reserve in both the tropical and central rainfed areas 96. Excluding benefits from the off-faun effects of reduced erosion.  181 (Table 10.8). The conclusion is less clear with respect for group "B" techniques, because there are so few, but they would generally be inferior to a land reserve. The NPV of the land reserve costs would be about US$421ha in the tropics and US$821ha in the central rainfed areas; the higher cost in the latter zone reflects the higher productivity, and hence the higher opportunity cost, of land there. The land reserve would be superior to the group "C· techniques in the central rainfed areas-Le., though a reserve would have negative net economic benefits, those negatives would be smaller than the negative benefits of the group "C" techniques. A land reserve program would be similar in (negative) net benefits to soil conservation in the tropics. In general, a land reserve is only more efficient than soil conservation practices if the available practices are not profitable. This is essentially the finding of Sinner (1990) for the United States. Sinner's results have been criticized for minimizing the effects of supervision costs in promoting soil conservation practices, but at any rate, those costs (under Mexican conditions) have little effect on the decision to promote conservation practices or to reserve land. 10.51 There will be some farming areas where no soil conservation technique is profitable. However. in some of those areas farmers may continue to grow crops for subsistence because the cost of cultivation to them is lower than the cost of acquiring food through the market. Subsistence cultivation would therefore continue to incur soil erosion costs both on the farm and off and to cause serious land degradation. In such areas, the best alternative for the government is to buy the land. Because many such highly erodible areas will receive PROCAMPO payments, the costs of land purchase could be made a part of PROCAMPO. Once the land was purchased it could be converted to wild areas or to public forests. Table 10.9 Net Benefits per Ton of Soil Saved by Soil Conservation Practices (NPV in US$ million) Group A Group B Group C. Tropical states (NPV of land reserve net benefitslton of soil saved = US$-42Iha) Maize 11 -8 -53 Beans 13 no practices -39 Sorghum 11 no practices -41 Central rainfed states (NPV of land reserve net benefits/ton of soil saved = US Maize 82 no practices -103 Beans 27 no practices -124 Sorghum no practices no practices -21 NOTE: Target erosion loss is 5 mt. SOURCE: Generated by model using parameters in Table 10.6. E. Comparison to Other Results 10.52 The productivity impact of on-farm erosion has been estimated in several countries. While their methods and data differed, most case studies revealed significant economic losses on-farm due to erosion.  182 (a) In the United States, Crosson (1983) estimated the decline in wheat, corn and soybean yields to be from 1.5 percent to 2.0 percent from 1950 to 1980, costing farmers of the order ofUSS500-1,OOO million per year, or about 1.4 percent of agricultural GOP in 1980; (b) In Mali, Bishop and Allen reported (1989) a decline of farm income due to crop yield losses from erosion of between 2 percent and 10 percent; (c) In Haiti, an average estimated internal rate of return was 23 percent. To put it another way, the net benefits of erosion control were almost zero if farmers' discount rates were 20 percent (White and Jickling, 1991); and (d) A World Bank (1990) review took a slightly different approach and computed the cost of on-farm nutrient loss in different countries to be from USS2,500 million in Zimbabwe, to USS315 million to USS100 million in Java and the Philippines, respectively. 1. Fstimates of the Benefits of Erosion Control in Mexico 10.53 There is no clear understanding of the economic costs and benefits of soil conservation practices at the farm level in Mexico. One study in Aguascalientes and Oaxaca was done in 1916 (Schram, 1918). Aguascalientes is in a semi-arid region with relatively low crop yields, while Oaxaca is a mixed temperate and tropical one with somewhat higher yields. A full economic evaluation of these experiments was not done and only estimates of FRRs are available. At that time (1916) chemical and mechanical farm inputs were heavily subsidized and the exchange rate was overvalued, so fmancial results probably diverge substantially from the economic ones. Some important results are as follows: (a) The range of real financial rates of return (FRRs) to soil conservation practices was from -9 percent to 19 percent. Only Oaxaca had FRRs above 12 percent. These rates are probably lower than the real discount rates of farmers and hence would not encourage adoption of the practices studied; (b) The highest FRRs were with mechanically constructed terraces without vegetation. Deep plowing by itself, or with terrace construction, was uneconomical in Aguascalientes and Oaxaca; (c) About two-thirds of benefits from soil conservation accrue from soil nutrient savings and one-third from incremental crop output. The US study of Colaciccio and Setia (1986) found that soil nutrient savings were about 40 percent of the benefits and incremental crop output was 60 percent; and (d) Average yield increases in maize production were estimated to be 25 percent if only terraces were built, 20 percent if only subsoiling were undertaken and 40 percent if both terraces and sub-soil plowing were combined. Differences in FRRs between terraces and deep plowing were explained by their differing production costs. 10.54 Trueba and others (1983) conducted a detailed experiment with conservation practices in Michoacan state. They studied interactions among type of terrace (bench, wide base, spaced), tillage  183 (none, minimum, traditional), and soil fertility amendment for maize, beans. vetch and wheat. Their results were as follows. (a) One experiment showed no statistically significant effect of terrace type or of tillage method on maize yield. No economic analysis was done; (b) A second experiment, involving the interaction of wide base terraces, constructed with three spacings, with tillage type (none, minimum, and traditional), showed significant effects of the different treatments on maize yield and net benefits. All treatments except one repaid their initial fixed and operating costs in one year, and the remaining treatment repaid it in two years. The average financial rate of return of the eight treatments to perpetuity was at least 50 percent; and (c) The no-tillage treatment was the most profitable in the second experiment. The treatment would have cut erosion to very low levels though the effect was not measured with precision in the experiment. V. Policy and Institutional Issues 10.55 The necessary condition for actions to control erosion is that its social benefits exceed its costs. The preceding sections show that controlling erosion would produce significant net social benefits in terms of augmented crop productivity alone. The sufficient condition for public actions is that externalities exist which are intractable by private actions. Two major externalities have been highlighted: (a) the private discount rate of poor producers is greater than the economic discount rate; and (b) producers cannot capture the full economic benefits to soil conservation investments because some information about soil conservation techniques is a public good. As such, techniques to fight soil erosion would not be optimally supplied by the private sector and public investment in it would be necessary to reach optimal levels. A. Soil Conservation Policies 10.56 From 1950 to 1982 some 3.13 million hectares were the site of various public soil conservation activities (Table 10.10). These activities were quite varied and not always strictly related to soil conservation, including land clearing, restoration of waste lands, soil and water conservation practices. water for livestock. and establishment of orchards. The area accounted for about 5 percent of the potential crop land of Mexico in 1982. The costs of these activities was small, being at most about NS278.3 million (in 1992 pesos) in the most recent period and averaging about 0.24 percent of agricultural GOP at its highest. The rapid increase in area treated and in total cost beginning in 1971 was due to a change in the programs; before 1971. farmer participation was voluntary and after that year farmers were paid to participate. Since the 1985 reorganization of SARR, no detailed cost data is available but the total is known to be much smaller as than during the period 1971-1982 as many of the staff of the General Directorate of Social and Water Conservation (DGCSA) were let go and many activities curtailed. 10.57 There is no good analysis of the earlier programs. especially as far as economic effects in the field are concerned, but some partial information is available. Early conservation theory did not emphasize the tillage and cultural aspects of conservation. Much of the program used heavy equipment to create such structures as terraces and contour banks. These structures were not always maintained by the beneficiaries and did not yield the expected economic results. Little training in  184 conservation practices and techniques was given to the local extension agents. The DGCSA operated independently of other SARH operations, resulting in a transfer of soil conservation knowledge somewhat divorced from extension of other improved agricultural practices. With the cessation of a Table 10.10 Public Activities in Soil Conservation, 1950-1982 Cumulative Area Total Cost Average Percent or 1950-1958 0.17 28.34 0.01 1959-1964 0.25 29.96 0.02 1965-1970 0.31 35.86 0.03 1971-1976 0.66 114.79 0.16 1977-1982 1.74 278.83 0.24 SOURCE: SARH estimate provided to Bank mission. centralized agency, nearly all public effort to transfer soil conservation techniques stopped. Only limited government and academic work continues, however. 10.58 Field visits show several conservation practices to have lapsed while others are maintained and provide benefit. Since many of the early conservation practices were structurally oriented and were constructed directly by the government some degree of abandonment has occurred. This was not always so in some areas where maintenance and in fact some recent extensions to the original structures have been made. B. Current Soil Conservation Policy 10.59 There exists a broad legal framework for public actions in soil conservation. The most general laws for agricUltural production are: (a) Ley General del Equilibrio Ecol6gico y la Protecci6n del Ambiente (1988); (b) Ley de Conservaci6n del Suelo y Agua (1946); (c) Ley Forestal (1992); (d) Ley de Reforma Agraria (1992); and (e) the Ley de Distritos de Desarrollo Rural (1988). 10.60 The Ley General del Equilibrio Ecol6gico y Ia Protecci6n del Ambiente managed by the Ministry of Urban Development and the Environment (SEDUE) seeks to control environmental damage by requiring environmental assessment of proposed deforestation and other major land use changes. Only on approval of the required mitigation procedures is permission granted for land use changed. It was in part designed to put a stop to the indiscriminate deforestation that started during the period 1973-76 with the Programa de Desmonte when credit was provided to ejidos without access to lands to bring into cultivation new lands that were forested or unused. Many of these lands were marginal and are subject to erosion hazards. 10.61 The Ley General del Equilibrio Ecol6gico y la Protecci6n del Ambiente has precedent in most cases over other laws affecting environmental issues in agriculture. The Ley de Conservaci6n del Suelo y Agua has no promulgated regulations and is inactive. SARH administers the Ley Forestal for environmental and conservation activities in designated forest areas. The Ley de Distritos de Desarrollo Rural confers certain powers on the DDRs. including that of undertaking soil conservation activities in collaboration with producers' groups.  185 10.62 The National Plan for Agricultural Modernization defines agricultural strategies from 1990 to 1994. The program seeks to increase production of major food and animal products by improving the productivity and using natural resources more efficiently, with priority given to the appropriate use of water and soil. Within that plan, the government's soil conservation policy in the 1980s has been to view soil conservation extension and works as part of an effort to extend profitable and sustainable agricultural practices. There were several steps in this change. (a) In 1984 DGCSA was downgraded to a Subdirectorate and later eliminateq.; (b) In the remaining years the responsibility for soil conservation was decentralized to the individual states but without specialized budget and programs. Except where local interest was strong and had political backing most soil conservation programs ceased during these years; (c) The responsibility for water conservation passed to the CNA; and (d) In 1991 the government re-formed the Subdirectorate for Soil and Water Conservation (SDCSA) with responsibility for soil and water conservation within SARR. The SDCSA's initial approach will be to try to arrest erosion on crop and pasture lands before addressing communal and forestry lands. The SDCSA has only four professional staff and lacks resources to have a serious effect on soil erosion. C. Why have farmers not adopted better soil conservation practices? 10.63 Information costs. Private suppliers of soil conservation techniques might not invest fully because they cannot capture all the benefits from new practices that farmers can easily copy. This public goods problem would lead to an insufficient supply and therefore high costs of soil conservation technologies. Lack of information or expensive information does not seem to be the main barrier to wider use of soil conservation practices, however. Allover Mexico, farmers know of such practices but often do not apply them, or apply them only inconsistently, for reasons other than information costs. 10.64 Complex or indivisible technology. A related question to information costs is complex or indivisible technology that cannot be profitably applied by small farmers. Complex technology does not seem to bar soil conservation in the cases studied. Farmers already apply many of the methods. New techniques developed by research are often variants of well-known tillage and cultivation practices, and their wider use would not require dramatic changes in farming systems. With respect to the divisibility of soil conservation techniques, farmers would not use indivisible techniques because the average fixed costs of doing so are too high without incurring the costs of group action. Many known techniques are divisible (e.g., Wilken, 1987), so adoption by individual farmers is feasible without costly group action. Moreover, necessary equipment for soil conservation on individual plots can be obtained in rental markets, again obviating group action or costly purchases of equipment. 10.65 Ingh discount rates. It has long been contended that high private discount rates harm incentives to conserve soil because conservation benefits accrue over a long period while its costs are more immediate. Many soil conservation methods do not require high initial investments, however. Of the individual techniques shown to be privately and socially profitable, few required initial investments greater than those in subsequent years. This is obviously true of tillage and cultivation practices that are repeated annually or in a rotation; the incremental costs of minimum tillage may  186 even be negative if cheaper herbicides replace more expensive machinery. In addition, some conservation benefits accrue after a brief period, so the discount rate is inherently less forceful. Therefore, while there may be some theoretical justification for subsidies, the benefit-cost analysis given in this chapter suggests that good incentives exist for private adoption of socially profitable conservation methods without cash subsidies. . 10.66 There are two other aspects of discount rates that might make investments in soil conservation less privately profitable. First, if farmers' private discount rates exceed economic rates because they lack access to commercial capital, then a subsidy to the interest rate paid by farmers might be justified in order to equalize private and social benefits from soil conservation. However, such subsidies have tended to be captured by larger farmers, to lead to corruption, to bias the choice of technique toward soil conservation works (in which the cost of administering subsidies is less than in soil conservation practices, but the net benefits are lower) and to be expensive per quantity of soil conserved because of the cost of identifying farms at risk to erosion. A better alternative to specific soil conservation subsidies is to improve banking systems for farmers to eliminate the gap between economic and private discount rates. 10.67 Second, it can be said that investment in soil conservation differs from other investments, such as machinery, because soil loss is permanent - the soil lost beyond the tolerance level can never be recovered. Even if there is no divergence between the private and social discount rates, the soil lost to externalities causes producers to assign a lower value to the permanent losses in land productivity than does society does and to underinvest accordingly. A subsidy 0 the interest rate used to finance soil conservation might therefore promote investments beyond the private optimum and reduce or eliminate the additional social losses to erosion. Even if this divergence only applies to smallholders without access to private credit, it might still affect many Mexican farmers and is therefore of empirical significance. 10.68 Physical externalities. Even if soil conservation equalizes marginal benefits and costs on a given farm, there may still be additional costs of erosion, such as siltation of dams or waterways, or erosion on adjacent farms. These physical externalities are of no consequence to farmers whose land is the source of the externality and hence give no incentive for better soil conservation methods. While elimination of off-farm costs produces benefits to society, it gives nothing to the farmer investing in soil conservation. Even if farmers capture the full benefits of eliminating soil erosion on their land, they would capture no benefits from eliminating erosion off-farm. Farmers would accordingly underinvest in soil conservation and a subsidy would be required for them to invest enough to eliminate the off-farm effects of erosion. The effects of erosion of adjacent lands would justify government action only if the farmers involved could not contract among themselves; for example, by collective action for soil conservation. The costs of such physical externalities are likely to be site-specific, for example, in watersheds above hydropower plants and no general conclusion is possible about their deterrent effect on appropriate soil conservation measures. 10.69 Past policies. Another reason for inadequate conservation is that past government policies - high fertilizer subsidies, distorted product prices, official production targets, distorted rural fmancial markets - harmed economic incentives for soil conservation. 10.70 Land clearing resulting from subsidized credit was a major distortion. Government formerly gave direct and indirect credit for land clearing, thus providing additional stimulus for soil erosion. Maas and Garcia-Oliva (1990, at 28) note that the expansion of livestock production, promoted by subsidized credit, led to significant deforestation and soil erosion. That stimulus is now less  187 important, as the credit has been eliminated and the rate of growth of cropped land has fallen. Clearing of new land is now, moreover, subject to strict environmental controls. 10.71 Other distorted policies were producer price protection and fertilizer subsidies. The producer prices of maize and beans were typically above world price equivalents. The harmful effect of protection would have been to expand cultivation of those crops onto erodible marginal areas. But, this cannot have been very important in the past 25 years because the rainfed area under those crops has not expanded very much; hence, removing protection to such crops would not necessarily reduce the expansion of cultivation because it is already low. The beneficial effect would have been to raise the return to conservation practices. A beneficial effect of removing protection might be to shift incentives to less-erosive crops. Because the analysis has concentrated on rainfed lands, where the opportunities for shifts to other crops are much more limited than they would be on irrigated land, it is likely that a shift to less-erosive crops would have only a limited effect. 10.72 Fertilizer subsidies were disincentives to conserve soil because they reduced the private cost of mineral fertilizers necessary to replace nutrients lost to erosion. They also made it economically efficient to substitute mineral fertilizers for animal manures and mulches, which are known to slow erosion. These subsidies have disappeared and no longer discourage soil conservation. 10.73 Land tenure. A last reasons might be the consequences of historical land tenure policies on incentives to invest in soil conservation and other farm activities. Mexican law formerly did not give full private rights to users of ejido land, a form of communal tenure. This might have given ejidatarios insufficient incentive to invest in their land, including in soil conservation. A Bank review (World Bank, 1990) of land tenure in Mexico concluded, however, that data gathered after 1970 indicated that ejidos were "likely to be at least as productive on average as private farms". More specifically, the report presented no evidence about soil conservation, and other land investments, on ejido lands compared to those on other tenure types; The Government has changed its land tenure laws (see Chapter 8); these reforms will eventually reduce distortions in the rural land market and give better incentives for soil conservation, notably by improving security of tenure and by raising the collateral value of land. D. Are Public Erosion Control Measures Justified? 10.74 The recent cessation of public interventions to promote soil conservation is in contrast to SARR's policy from World War n to 1985. The government has a few clear alternatives to attack agricultural soil erosion.91 Immediate measures to reduce erosion are: (a) public works; (b) subsidies for farming practices or works via credit, direct payments or sharing part of soil conservation costs; (c) payments to farmers to reserve land from production; (d) extension; and (e) better incentives, including institutional reforms. 10.75 Public works. The government could construct public works on-farm for soil conservation, in essence reviving the historical activities of the defunct DGCSA. The main reason why public works would not be effective is that farmers lack incentives to maintain public works constructed, or might even destroy works to ease movement of field machinery. (Both phenomena were seen during the era of the DGCSA). A second reason is that rich farmers can corrupt public agencies and capture a disproportionate share of works and maintenance; there is only anecdotal evidence of this in Mexico. A third is that soil conservation practices are small and simple enough that private agents 97. It is considering a 20-year National Program for Soil Conservation in which a federal agency would manage planning. research. training, and extension. lome credit. and minor works. but it is not known if this Program will be implemented.  188 can build them with equal quality. While public financing, as subsidies to producers or to construction firms, may be necessary to promote some techniques, the initial investments are done more efficiently by competitive private agents. 10.76 Subsidies. Subsidies to promote soil conservation techniques would not be justified economically except in unusual circumstances. Those circumstances are the very few sites at which the economic benefits of soil conservation are positive but private benefits are negative. If the public authorities in Mexico had such preferences. then it would be much cheaper to pay farmers to reserve the erodible land from production. General subsidies to promote soil conservation techniques are not justified economically except in unusual circumstances. Those circumstances are the very few sites at which the economic benefits of soil conservation are positive but private benefits are negative. At such sites, it will be cheaper to pay farmers to reserve the erodible land from production or to buy the land. 10.77 The GOM should not give incentives through public financial institutions for soil conservation. This will lead to excessive credit flows to dubious soil conservation investments as a means of getting subsidies. FIRA would provide commercial finance to richer producers who would contract extension privately, but would not transfer erosion control subsidies. The main reason for not channeling subsidies through FIRA is to avoid them where erosion is light. Co-generation of projects is an important self-selection mechanism to maintain soil conservation structures. 10.78 Land reserve payments. Payments to producers to leave land idle are an alternative to subsidies or technical assistance for soil conservation. The comparison between a land reserve and subsidies to techniques depends on the availability of profitable techniques. If techniques which are both economically and financially profitable are available. then they produce much higher net benefits than a land reserve under the same conditions. If the only techniques available are not economically or financially profitable, then it would be better to pay farmers to do nothing with the land because the costs of reserve payments (including supervision costs) are less than the subsidies required to make farmers adopt such techniques. One possible means of implementing land reserves is to restructure the legal basis of PROCAMPO to use its funds to purchase erodible land for public reserves. If properly targeted to areas of severe erosion and low production value, a substantial area could be bought. 10.79 Institutional responsibility. The GOM should give clear institutional responsibility to one agency for soil conservation. except research on the issue, which should remain with INIFAP. This has not been done since the disappearance of the General Directorate for Soil and Water Conservation (DGSA) in 1985. SARH would have two roles in soil conservation: scientific and extension. The former would be maintained in INIFAP. The latter would be done by strengthening centralized support for soil conservation. 10.80 Extension. The GOM should target its remaining extension program to severely affected areas exclusively. It is economically beneficial to promote appropriate techniques through the extension channels. For small and medium scale farmers this would be done through the public extension service. For larger and richer farmers it could be done through private extension, whether contracted individually or through producers' associations. SARH's extension role would be to provide technical assistance through the DDRs and in collaboration with FIRA and FIRCO. 10.81 Research on soil erosion and conservation. Major remaining areas for research are: (i) the productivity effects of erosion and conservation, plus the costs and benefits of erosion control for different crops and farming situations; (ii) off-farm and on-farm effects. The sizes of on-farm and  189 off-farm effects are a major gap in knowledge because they determine the subsidies needed to elicit optimal on-farm soil conservation investments. The additional research required to estimate those effects may be relatively cheap, however, because the magnitude of erosive deposits in waterways can be estimated with mathematical models developed elsewhere. It is unlikely that further information about those relative magnitudes would change erosion control techniques, but it might affect the target areas and the public actions to control erosion; (iii) Risk. This comprises producers's attitudes to risk, the contribution of erosion to the variability of farm income, the effect of soil conservation on income variability, and the impact of random factors. This problem is difficult because: (a) the true variability effect of erosion is unknown; (b) the distribution of erosion over time is unknown; (c) if the effect on variability is large, it might be a significant share of agricultural sector income; and (d) interactions between erosion and other factors are obscure; (iv) Affected regions and activities. The most severely affected regions and crops are well-known, not only from field studies in Mexico, but from analogous research carried out in similar agricultural situations throughout the world. Therefore, further descriptive studies of, for example, the quantities of erosion loss per state or even per DDR, should not be apriority. One major unknown is the effect of livestock on soil erosion and the benefits of soil conservation practices in grazing areas. Mexico - Agricultural Sector Memorandum ANNEX 1 Page 1 Total Year Area Constant 1992 Value Constant 1992 Valu Area Yield Price Output Domestic World Area Yield 1992 price Output Domestic (000 hal (000 hal (mtlha) (N$/mt) (000 mt) (N$ 000) (NSOOO) (000 hal (mtlha) (N$OOO) (000 mt) (NSOOO) 1939 S,2S4 3,267 0.61 901 1,977 1,780,671 S63 0.16 I,S16 429 615,941 1940 S,388 3,342 0.49 187 1,640 1,290,943 601 0.77 I,SI1 464 703,561 1941 S,6S7 3,492 0.61 129 2,124 1,76O,S03 sa3 0.7S I,S03 434 652,859 1942 6,106 3,7S8 0.70 838 2,636 2,209,756 600 0.82 1,402 489 685,776 1943 S,357 3,083 0.59 1,091 1,808 1,983,026 SIO 0.71 I,SSI 362 561,765 1944 S,669 3,3SS 0.69 1,186 2,316 2,748,017 527 0.11 1,361 314 509,112 1945 S,707 3,4S1 0.63 1,221 2,186 2,669,853 461 0.14 1,361 341 474,473 1946 S,S16 3,313 0.12 997 2,113 2,31S,024 41S 0.82 1,431 340 481,003 1947 5,840 3,S12 0.72 1,020 2,S18 2,566,771 499 0.8S 1,430 422 603,240 1948 6,158 3,722 0.76 963 2,832 2,127,44S S71 0.83 1,459 471 696,ISO 1949 19S0 1951 1952 1953 1954 19S5 1956 1957 19S8 1959 10,802 6,324 0." 1,03S S,S63 5,7S9,IS4 S,529,926 937 1.3S 1,270 1,266 1,606,921 1960 10,107 S,5S8 0.91 I,OS2 S,420 S,703,096 S,273,262 840 1.42 1,153 1,190 1,490,941 1961 11,04S 6,288 0.99 1,040 6,246 6,494,348 S,673,082 831 1.68 1,266 1,402 1,774,841 1962 11,080 6,372 0.99 1,02S 6,337 6,494,463 5,683,653 741 1.95 1,201 1,455 1963 11,944 6,963 0.99 1,224 6,870 8,408,ISI 6,579,277 819 2.08 1,189 1,703 1964 1965 M:\mx\asm\data\NPC.wbl[NPCnew] 28-Jul-94 Mexico - Agricultural Sector Memorandum ANNEX 1 Page 2 Total Year Area Constant 1992 Value Constant 1992 Valu Yield Price Output Domestic World Area Yield 1992 price Output Domestic (000 hal (000 hal (mt/ha) (N$/mt) (000 mt) (N$ 000) (N$ 000) (000 hal (mt/ha) (N$OOO) (000 II ~~~~~!:::~::::::::::::::::~~::::::;:::;:::~::~~::::~:::::::~:~~~~~::::::~:;::~~*~~?::;~:~;::~::$:~~:;~~~~U~~~~:~~1~::;:;:i:~~~~?;~mmt~1~~~~;:~~::~~~~~~~~~~~~~l~~~W~~~~~~n~~~mlli?;2.~"1Will~WJ~~~~..~.lli1t.~m~~~;:~~~~~~~imlm~;~L~~~~~~}~~iii::: 1966 1967 1968 1969 1970 ,,-- .. ""'........""''"'" 1971 13,7S2 7,692 1.27 821 9,786 ',OOS,7" 7,202,717 614 2.98 786 1,831 1,438,320 1972 13,42S 7,292 1.26 773 9,223 7,128,212 6,168,409 687 2.63 730 1,809 1,320,663 1973 14,007 7,606 1.13 83a 8,609 7,212,516 8,311,106 641 3.26 672 2,091 1,405,748 1974 13,012 6,717 1.17 898 7,848 7,045,093 8,261,778 774 3.60 82S 2,789 2,299,739 1975 13,486 6,694 1.26 989 8,449 8,353,115 7,198,559 778 3.60 915 2,798 1976 1977 1978 1979 1980 1981 1982 1983 14,726 7,421 1.76 965 13,061 12,606,637 11,340,830 1S7 4.04 669 3,460 2,31S,211 1984 14,165 7,077 1.81 1,047 12,789 13,389,636 10,231,227 1,033 4.36 757 4,50S 3,409,263 1985 15,273 7,590 1.86 I,OOS 14,103 14,174,528 9,185,656 1,217 4.28 710 5,214 3,703,105 1986 13,644 6,417 1.83 998 11,721 11,697,450 8,100,932 1,201 3.97 681 4,770 3,248,011 1987 14,456 6,801 1.71 1,069 11,607 12,410,690 6,648,741 988 4.47 639 4,4 IS 2,820,233 1988 1989 1990 1991 1992 .......... "' .. __ • ______ • _______ • _________ •••••.• _._ •.•• _ .... __ •.•. •• •• ............. ..... _ ............... .. ___ .... ____ ••••••••••••••••••• • •••.....•.••• ···•·•·•·•·•· ____ __ ••• •• .. .. " " ......... " " " ~ ~ ~ ~ ~ ~ ~ _ . ~ _ . _ ~ ~ . _ " ' ~ " ' 9 " 9 " 0 ~ _ · T · T · . · ~ ~ ~ ~ . · ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ . . " " " y y ~ . M:\mx\asm\data\NPC.wbl[NPCnew] 28-Jul-94 Mexico - Agricultural Sector Memorandum ANNEX 1 Page 3 Year Constant 1992 Value Constant 1992 Valu World Area Yield 1992 price Output Domestic World Area Yield 1992 price Output Domestic (NS 000) (000 hal (mt/ha) (NSOOO) (000 mt) (NSOOO) (NS 000) (000 hal (mt/ha) (NS 000) (OOOmt) (NSOOO) 1939 139 0.65 658 90 59,484 1940 133 0.78 605 103 62,560 1941 147 0.63 683 93 63,251 1942 139 0.66 134 91 67,099 1943 136 0.58 838 78 65,724 1944 135 0.59 1,040 80 83,086 1945 165 0.68 981 112 109,596 1946 171 0.70 874 119 103,799 1947 171 0.68 876 117 102,315 1948 202 0.74 826 149 123,137 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1,318,769 107 1.67 819 179 146,556 136,618 240 0.75 1,018 179 182,070 1960 1,233,994 116 1.80 918 209 192,112 160,137 240 0.75 1,031 180 185,962 1961 1,416,539 117 2.49 783 291 227,551 235,113 233 0.75 1,038 174 180,811 1962 1,544,034 118 2.52 750 296 222,069 245,469 193 0.78 902 151 136,367 1963 1,7S3,SS2 198 2.04 759 402 305,152 342,365 232 0.&0 925 1&6 171,702 1964 1965 "''''''''''''''''~s''''''''_'''''.~.~_-'-'~''''''''''''''''''''''' .... wnn M:\mx\asm\data\NPC. wbl [NPCnew] 28-Jul-94 Mexico - Agricultural Sector Memorandum ANNEX 1 Page 4 Year Constant 1992 Value Constant 1992 Valu World Area Yield 1992 price Output Domestic World Area Yield 1992 price Output Domestic 1966 1967 1968 1969 1970 1971 1,300,361 ""'....... 936 _... _ .......... '"................ 2.69 621 2,516 1,563,303 1,8lS,I47 221 1.22 767 270 207,437 1972 1,338,738 1,109 2.35 631 2,612 1,646,953 1,894,233 217 1.43 755 310 234,071 1973 2,502,403 1,IIS 2.76 641 3,270 2,094,680 2,917,940 263 1.49 761 392 298,773 1974 3,522,327 1,IS6 3.03 718 3,499 2,722,774 3,377,942 173 1.44 765 250 191,440 1975 2,643,328 1,445 2.86 835 4,126 3,446,356 3,280,142 287 l.S4 836 440 368,218 1976 1977 1978 1979 1980 1981 1982 1983 3,025,595 1,518 3.19 582 4,846 2,820,645 3,832,711 303 1.84 818 557 455,565 1984 3,453,728 1,623 3.11 727 5,039 3,660,709 3,384,642 283 2.19 9lS 619 572,523 1985 3,555,237 1,862 3.54 653 6,597 4,306,626 3,865,266 211 1.91 924 536 495,343 1986 3,766,279 1,533 3.15 896 4,833 4,332,160 3,243,191 264 1.95 784 51S 403,S64 1987 3,111,765 1,8S3 3AO 702 6,298 4,418,703 3,400,760 286 2.16 627 617 1988 1989 1990 1991 1992 ____ W""·'·'·"_'·_."" ____ ·.·_T _______________ T'_._.·o ____ • __ •• ,··.·''· ••. ___ ._·.·._ •••• ·•· •• ·• __ .·._·.·. __ ,...."" ....... ........ _ .. " .. """""' __ ,, __ ._••••• _ ••• _. __ • _____ ." ..... ___ . _ . __ • • ·0· __ ·•· __ • • • • ___ • • • • • • • • • • _ _ _ _ _ • _ _ _ . _ • • • • • • _.n._.-.~_ .. _ ............ ~ ________·..·.·. ____ .·_·.·._·._. ___·.___.__ ........ _______._._.. _ ~~. ... _._.. _ " ~ _ ~ M:\mx\asm\data\NPC.wbl[NPCnew] 28-Jul-94 Mexico - Agricultural Sector Memorandum ANNEX 1 PageS Year Constant 1992 Value Constant 1992 Valu World Area Yield 1992 price Output Domestic World Area Yield 1992 price Output Domestic (N$ 000) (000 hal (mt/ha) (N$OOO) (000 mt) (N$OOO) (N$ 000) (000 hal (mtlha) (N$OOO) (000 mt) (N$OOO) 1939 40 2.57 1,256 103 129,460 1940 62 1.75 1,260 108 135,686 1941 53 2.06 1,357 109 148,395 1942 62 1.74 1,632 '108 176,539 1943 66 1.75 1,538 114 176,078 1944 68 1.S3 1,361 104 141,844 1945 56 2.15 1,533 121 185,686 1946 64 2.19 1,483 139 206,822 1947 72 1.91 1,492 138 205,608 1948 32 5.14 1,389 163 226,263 1949 108 1.71 1,277 185 235,801 ~~~~~~~~~::::~~~~~~~::~~~~~~~~~§:~~~m?:~::~~~~~~~~~~~~~t:*~~~~::~~~~~::::~~~::~~~lliU~§~m::?;~~l~~~~~~~~~~~~l:~*:~~l 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 199,947 127 2.05 1,264 261 329,919 868,466 1960 204,805 143 2.30 1,286 328 421,217 1,009,728 4 1.24 1,747 5 .,680 1961 191,619 146 2.28 1.288 333 428,907 755.866 10 1.99 1.766 20 34,851 1962 160,840 134 2.16 1,356 289 391,740 961,801 27 2.08 1,751 57 99,320 1963 189,854 l3S 2.20 1,372 296 406,614 946,999 27 2.05 1,781 56 100,208 1964 1965 M:\mx\asm\data\NPC.wbl[NPCnew] 28-Jul-94 Mexico - Agricultural Sector Memorandum ANNEX 1 Page 7 Year Constant 1992 Value Constant 1992 Valu World Area Yield 1992 price Output Domestic World Area Yield 1992 price Output Domestic (N$ 000) (000 hal (mtlha) (NSOOO) (000 mt) (NSOOO) (NS 000) (000 hal (mtlha) (NS 000) (000 mt) (NSOOO) 1939 161 0.45 1,116 119 139,944 1940 254 0.43 1,000 110 109,196 1941 316 0.43 1,143 137 170,649 1942 362 0.47 1,471 170 151,456 1943 <C09 0.47 1,611 193 311,731 1944 390 0.45 1,319 174 131,150 1945 366 0.44 1,541 161 148,117 1946 318 0.45 1,891 149 111,301 1947 333 0.47 1,795 156 179,446 1948 405 0.49 1,5n 199 313,171 1949 1950 1951 1952 1953 1954 195) 1956 1957 1958 19S9 751 0.85 1,045 636 665,079 1960 NA 26 1.15 1,797 32 57,590 899 0.88 I,OSS 787 130,917 1961 NA 33 1.26 1,825 41 75,563 794 0.97 1,183 767 906,567 1962 NA 37 1.27 1,707 47 79,831 787 1.05 1,091 825 899,613 1963 NA 37 1.29 1,699 47 80,141 847 1.01 1,094 908 992,914 1964 1965 ........................................................... . " .........,..,..., ........ ". M:\mx\asm\data\NPC.wb1 [NPCnew] 28-Jul-94 Mexico - Agricultural Sector Memorandum ANNEX 1 PageS Year Constant 1992 Value Constant 1992 Valu World Area Yield 1992 price Output Domestic World Area Yield 1992 price (N$ 000) (000 hal (mt/ha) (N$000) (000 mt) (N$OOO) (N$ 000) (000 hal (mt/ha) (N$ ~~~~::::;:::;:~~;:~::~~::;:::::::::~::~~~~~~:::::~1~~;;:t~~:~~~~~~U~~~::::~::::~~'::~~'::~~~~::;:~~~l:~~~~~::~~::~~~~~~~~~j~~~m~~~~~;;:;:~::-<f.*:!:$.~~::3.~~:::~'t~~~,:::~~~:~~m~~U::~~tw.mt~fu111~~~mmm~~~~::::m~~i~~~~l~~::i~:~ 1966 1967 1968 1969 1970 II!I iI" . .. "'" ""' ..... _ " " " - " _,"II _ " " .., , " _ " ...................... _ .. ,....... 1971 474,6S1 26S U5 1,419 411 512,7S1 451 1.36 1,104 624 688,358 1971 634,984 199 1.36 1,350 271 366,1l5 S23 1.28 958 670 641,632 1973 1,522,630 198 1.51 1,432 298 427,071 415 1.40 1,416 595 843,455 1974 1,118,229 192 1.42 2,300 271 626,541 578 1.43 1.343 826 1,109,481 1975 1.251.765 363 1.47 1.786 532 950.571 227 1.41 1,255 320 401.881 1976 1977 1978 1979 1980 1981 1982 .. ~.'II "'" UU"II .. ~~ ~ 1983 1,391.303 349 0.79 1,045 2n 289,060 232 U3 937 3S5 331,811 1984 1,389,366 117 0.91 1,066 209 123,107 30S 1.49 99S 454 451,631 1985 1,452,560 234 0.65 1,111 IS2 184,680 200 1.58 911 317 189,241 1986 1,245,951 203 0.79 1.763 161 283,528 1S7 1.42 571 114 127,815 1987 2,373,925 271 0.79 1,030 219 216,051 222 1.87 373 414 1988 1989 1990 1991 1992 M:\mx\asm\data\NPC.wbl[NPCnew] 28-Jul-94 Mexico - Agricultural Sector Memorandum ANNEX 1 Page 9 Year Constant 1992 Value Constant 1992 Valu World Area Yield 1m price Output Domestic World Area Yield 1992 price Output Domestic (N$ 000) (000 hal (mtlha) (N$ 000) (000 mt) (N$OOO) (N$ 000) (000 hal (mtlha) (N$OOO) (000 mt) (N$OOO) 1939 632 0.23 2,044 148 302,867 88 0.63 1,7lS 56 95,657 J940 635 0.15 1,757 97 169,988 100 0.66 1,599 66 104,864 1941 672 0.24 1,674 160 268,194 101 0.72 1,755 72 127,055 1942 750 0.24 1,491 183 272,559 124 0.67 1,313 83 108,667 1943 700 0.22 1,456 157 229,138 124 0.62 1,677 76 127,906 1944 734 0.25 1,394 183 255,434 126 0.71 1,409 90 126,752 1945 728 0.22 1,662 162 268,872 137 0.73 1,471 100 146,399 1946 734 0.19 2,385 139 330,677 141 0.75 1,336 106 141,575 1947 741 0.27 2,664 199 529,842 143 0.78 1,300 111 144,740 1948 788 0.27 2,187 210 458,426 144 0.77 1,399 III 154,741 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1,411 0.41 1,910 581 1,110,305 145 0.79 1,364 114 lS5,485 1960 1,326 0.40 1,937 528 1,023,132 147 0.78 1,413 us 162,027 1961 1,617 0.45 2,171 723 1,570,448 156 0.86 1,530 135 206,605 J962 1,674 0.39 2,2lS 656 l,452,l7l 146 0.88 1,755 129 226,592 1963 1,711 0.40 2,200 677 1,489,723 105 0.93 1,825 97 177,366 1964 1965 M:\mx\asm\data\NPC.wbl[NPCnew] 28-Jul-94 Mexico - Agricultural Sector Memorandum ANNEX 1 Page 10 Year Constant 1992 Value Constant 1992 Valu Yield 1992 price Output Domestic World Area Yield 1992 price Output Domestic (N$ 000) t~::::::;::~~::::::::::~~::~~:;~::::~::::~;~~::~::::~::::~::;;::~::::~::::~:::::;;:::::~)::::%~~~::~~::::;~:t~~:;§:~~~~~~~~~~~~::~~~~~~~::~::~~:::*~~:~~~~~~~:~t~~m~;~~tm~~~~~~z;@mm~~~~~~;:~f~~~~\§:~t~mmmi:s}~~~~~~~t~~m.~~;t.~m~~~~m.~~:~~~~~~~u~~~~~~~~~~~~~~~~$.~~ 1966 1967 1968 1969 1970 1971 1,96S 0.49 1,803 9S4 1,719,610 215 0.71 I .. 167 140,899 1972 1,687 0.52 1,740 110 1,51l,I85 247 0.92 1,149 228 262,045 1973 1,870 0.5.4 2,260 1,009 2,280,l42 216 L05 1,723 226 389,477 1974 I,H2 0.63 3,437 972 3,339,769 248 1.00 1,30S 249 315,108 . 1975 1,7S3 0.59 2,791 1,027 2,867,673 191 1.02 1,598 195 311,710 1976 1.316 0.S6 2.089 740 I 1977 1978 1979 1980 1981 1982 1983 1,996 0.64 1,489 1,282 1,908,487 137 0.41 2,m 57 IS3,155 1984 1,725 0.54 1,501 931 1,403,885 127 1.09 1,5l1 138 211,299 1985 1,782 0.51 2,990 912 2,716,1S7 125 0.93 1,572 117 183,489 1986 1,820 0.60 2,969 1,085 3,220,926 140 1.07 1,106 ISO 165,923 1987 1,787 0.57 2,245 1,024 2,298,160 140 1.07 1,275 ISO 191.227 1988 1989 1990 1991 1992 ................ -.~ .. -.-.~.---~--------------.---~~-~.-- .~ ........... -.-.... -~-~.-.~-.--------.---- ....---.-----------........_-- ...... ......." ............... M:\m:x\asm\data\NPC.wbl[NPCnew] 28-Jul-94 Mexico - Agricultural Sector Memorandum ANNEX 1 Page 11 Year Constant 1992 Value Constant 1992 Valu World Area Yield 1992 price Output Domestic World Area Yield 1992 price Output Domestic (N$ 000) (000 hal (mtlha) (N$ 000) (000 mt) (N$OOO) (N$ 000) (000 ha) (mt/ha) (N$OOO) (000 mt) (N$OOO) 1939 94 48.64 93 4,556 422,643 120 0.46 4,842 55 265,568 1940 911 50.56 116 4,973 417,372 116 0.45 4,898 51 256,572 1941 116 48.82 81 5,677 457,642 126 0.41 6,224 52 322,3114 1942 127 53.73 82 6.800 556,369 130 0.40 6,535 52 341,955 1943 139 49.45 94 6,1158 643.171 134 0.39 5,755 52 301,119 1944 140 48.39 79 6,754 535,733 135 0.45 4,717 60 2113,554 1945 141 47.SS 18 6,742 595,889 135 0.41 5,473 55 299,492 1946 1411 48.71 78 7,197 5511,036 135 0.42 4,820 57 274,180 1947 157 53.44 78 8,412 657,350 135 0.41 5,414 55 299,916 1948 173 55.12 SS 9,559 809,712 136 0.39 6,818 53 1949 1950 1951 19S2 1953 19S4 1955 19S6 1957 1958 1959 31S 56.32 69 17,765 1,234,606 1,779,718 285 0.34 11,487 98 1,120,673 1960 346 56.43 71 19,542 1,389,770 2,049,884 304 0.41 10,986 124 1,365,406 1961 348 55.14 74 19,168 1,418,500 1,799,993 307 0.41 11,035 127 1,397,1S8 1962 362 5a.39 76 21,116 1,605,607 1,959,446 314 0.44 10,283 140 1963 378 59.07 81 22,327 1,798,465 5,724,234 323 0.42 10,425 137 1964 1965 M:\mx\asm\data\NPC. wbl[NPCnew] 28-Jul-94 Mexico - Agricultural Sector Memorandum ANNEX 1 Page 13 Year Constant 1992 Value Constant 1992 Valu World Area Yield 1992 price Output Domestic World Area Yield 1992 price Output Domestic (N$ 000) (000 hal (mtlha) (N$ 000) (OOOmt) (N$OOO) (N$ 000) (000 hal (mt/ha) (N$ 000) (OOOmt) (N$ 000) 1939 7 0.17 17,133 20,354 42 41.06 1,736 1940 I 0.11 17,014 24,171 40 42.49 1,719 1941 9 0.11 19,331 2 31,471 42 43.41 1,106 1942 10 0.19 17,714 2 34,454 41 41.30 1,716 1943 11 0.11 17,731 2 36,161 47 41.17 1,964 1944 11 0.14 14,209 1 23,203 41 42.32 2,037 1945 12 0.21 14,331 3 38,097 41 43.39 2,100 1946 18 0.24 13,990 5 61,956 48 43.16 1,126 1947 17 0.24 9,Il7 6 62,465 49 43.20 1,121 1948 27 0.24 12,714 7 12,642 53 41.44 2,141 1949 1950 1951 1951 1953 1954 1955 1956 1957 1958 1959 926,175 67 0.31 8,413 12 181,714 114,905 93 46.69 110 4,361 514,191 1960 1,163,725 67 0.35 7,694 24 182,615 121,497 90 47.05 131 4,240 556,990 1961 1,121,454 61 0.40 7,738 27 208,438 197,544 91 46.50 140 4,230 593,014 1961 1,131,040 68 0.43 7,469 19 120,146 107,615 101 50.35 151 5,090 766,671 1963 1,076,766 69 0.44 7,186 30 119,068 139,589 102 50.37 lSI 5,132 773,448 1964 1965 M:\mx\asm\data\NPC. wbl [NPCnew] 28-Jul-94 Mexico - Agricultural Sector Memorandum ANNEX 1 Page 14 Year Constant 1992 Value Constant 1992 Valu World Area Yield 1992 price Output Domestic World Area Yield 1992 price Output Domestic (NS 000) (000 ha) (mOha) (NS 000) (000 mt) (NS 000) (NS 000) (000 hal (mt/ha) (NSOOO) (000 n t=:~~~~~~~;~~;;::::~~~~~::::~?:~~~?:~~~~~~$:~::::~::::::~~-;;~~:~~~~::::::::::::~~~~~~~;:::~~},~~~~l~~~~~1~~::~~~::~~:~~~~i~~~~~~~~;:~~~~n.~~I;ra~~~~~{~m~m~~~~mnm~~~~~~~!~~t1:}~~l:!:n~~~~~~?:~~~i~~~mBt~~~n~,-~~~~~ 1966 1967 1968 1969 1970 ...............""- -"""""" ......."-,,-,, . 1971 1,356.216 78 0.33 6.122 26 160,043 146,486 164 59.14 127 9,619 1,221,714 1972 1,624,971 78 0.49 6,60& 31 251,319 224,lSi 161 62.14 III 10,434 1,162.237 1973 2,054,992 10 0.41 6,669 33 211,135 321,292 III 61.76 102 11,151 1,137,191 1974 1,761,667 12 0.42 1,164 35 214,111 421,111 201 66.04 101 13,278 1,344,372 1975 1,773,741 72 0.46 7,537 34 253,151 268,091 204 70.04 III 14,260 1,679,986 1976 1977 1978 1979 1980 1911 1912 1913 3,623.060 59 0.57 6,314 33 211,499 340,22S 243 62.79 96 15.261 1.471,392 1914 2.577,730 71 0.51 6,741 36 . 243,810 410,094 237 61.92 103 14.611 1,511.435 1985 2,405,723 72 0.71 7,900 51 402,771 507,647 243 56.71 92 13,776 1,273,249 1916 1,242,919 74 0.63 3,293 47 154,842 623,015 251 53.94 95 13,906 1,321,640 1987 5,507,666 68 0.61 1,628 41 66,760 466,034 263 47.68 99 12,525 1,245,588 1988 1989 1990 1991 1992 _ _ ...... ,.,.........-.w,.". .. ________ •· __ H ___________ ·.~.~· • • • _ _ _ _ _ _ • • ·•· • • • • ·•· _ _ _ _ _ _ _ ~ _ _.. _ro ........ .......-.-.................... _"."........,.... __ .... _....·__________ • _____ • _________________ ••.•.•.....•..... _••. _............ __ •.•·.·.·"·_.·"._· ••••••• _•• _·. __ ·•• _·.· •• ·.·.n_.,..,."'~~",. .... .....~,. ...... ~ .... ~ ........................ "'9"0 ......... ~ .. _ . . . . . M:\mx\asm\data\NPC. wbl[NPCnew] 28-Jul-94  MEXICO AGRICULTURAL SECTOR MEMORANDUM Annex 1 - Table 1 &&ricultural Portfolio of the Bankin& SIstem (In millions ofN$) 1983 1984 1985 1986 1981 1988 1989 1990 1991 1992 Total Credit Balances of the Banking System 1,913 2,960 5,234 24,819 58,831 80,496 122,486 188,563 214,139 399,958 at year-end aI Agricultural Portfolio of the Banking System bl 421 801 1,261 1,818 3,691 8,862 15,381 21,956 26,844 33,393 % a/Total Credit Balances 12% 27% 24% 7% 6% JJ% 13% 12% 100/6 8% of which: BANRURALcI 168 295 540 844 1,823 4,035 4,609 3,223 2,931 4,011 FIRAdI 138 219 387 597 1,044 2,411 3,871 6,535 9,789 13,184 Commercial Banks eI 121 287 340 377 824 2,416 6,901 12,199 14,118 16,192 Share in Total Agricultural Portfolio BANRURAL 39% 37% 43% 46% 49% 46% 3()% 15% JJ% 12% FIRA 32% 27% 31% 33% 28% 27% 25% 300/6 36% 39% Commercial Banks 28% 36% 27% 21% 22% 27% 45% 56% 53% 48% SourceLilDd ooBiI: a/ 9IIIICO de Maico. IDdieIdInI EcaDomiCOl. Excllllb credit to the pjllic sector. IDcIudes craIi1 to ~ IIedOn, • -u • bouIing _ CDnIIIIID!ltica credit. bI 9IIIICO de Maico. IDdieIdInI EcaDomiCOl. Excllllb craIi1 to the public sector. cI BANRURAL's fiIIIIICiaI ~ dI F1R.A's AIImaI. Rqlorb. TIIbls 1 _6. eI Difl'enmoe ~ Total s.nmw 5yIIan A&ricu1tunl Pmfolio _ F1R.A _ BANRURAL's portfolio ~. RURFIN.WBI  MEXICO AGRICULTURAL SECTOR MEMORANDUM Annex 2 - Table 2 Aericultural Financine - Net Annual Credit Flows aI (In millions ofN$) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 Agricultural GOP aJ 1,392 2,533 4,307 7,467 16,825 30,691 39,246 54,811 66,682 87,295 BANRURAL bl cl 52 127 245 304 979 2,212 574 347 251 1,080 ! F1RA dI 48 81 168 210 447 1,366 1,460 2,664 3,255 3,395 I Commercial Banks el 48 165 54 36 447 1592 4485 5298 1919 2074 Total Net Financing 147 374 466 551 1,873 5,171 6,519 8,309 5,425 6,548 Participation in Net Agricultural Financing (%) BANRURAL 35% 34% 52% 55% 52% 43% 9% 4% 5% 16% FIRA 32% 22% 36% 38% 24% 26% 22% 32% 60% 52% Commercial Banks 32% 44% 11% 7% 24% 31% 69% 64% 35% 32% Net Financing as percent of AgGDP BAN RURAL 4% 5% 6% 4% 6% 7% 1% 1% 0% 1% FIRA 3% 3% 4% 3% 3% 4% 4% 5% 5% 4% Commercial Banks 3% 7% 1% 0% 3% 5% 11% 10% 3% 2% Total 11% 15% 11% 7% 11% 17% 17% 15% 8% 8% Sources and noles: '" Nel credil nOM reprcsenllbe _wl increueldeereuc in portfotio ballnCu. bI Due 10 Ihe IrMlo.fer ofOANRURAI:. o"",,rdue portfolio 10 rRONASOI.. and FIDEI..IQ, nel chang« in pot1folio were nega,i"" in 1990·91. Atl ...1 net credilllow. (new loans leu reco""rio.) wele u...d inslead. cJ DANRURAI..'I fmancial 1 •• lernonlS. rJ/ FJRA'. AMuaI Reports, Tables 1 lind 6. eI Difference between portfolio babncu of Ibe agri<:ullUrai portfolio of Ibe banlcina I)'llem and F1RA lind BANRURAL. RURFIN.WBl'  MEXICO AGRICULTURAL SECTOR MEMORANDUM Annex 2 - Table 3 Fiscal and Quasi-Fiscal Transfers to Develoument Finance Institutions (1983·1992) (Millions of 1992 US$) aI 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1983·92 BANRURALbl 1,230 1,004 1,017 1,031 277 1,091 833 3,017 324 423 10,246 ANAGSAI AGROASEMEX 283 357 492 450 426 600 456 706 1 21 3,790 FICART 138 147 166 87 63 136 5 12 69 22 846 FIRA cl 160 349 384 391 224 396 197 99 141 III 2,451 FIRA-Other dI 957 521 739 903 975 89 332 364 210 196 5,285 i Subtotal 2,767 2,377 2,797 2,861 1,965 2,312 1,824 4,198 744 773 22,618 ! Other trust funds and dev. banks 39 26 20 6 9 12 51 6 8 0 176 TOTAL 1..8QL~.403 2,817 2,867 1,974 2,325 1,875 4,204 751 773 22,795 (Millions of 1992 N$) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1983·92 BANRURAL bl 5,453 3,920 3,989 5,275 1,429 4,502 3,251 11,194 1,097 1,309 41,419 ANAGSAIAGROASEMEX 1,253 1,393 1,930 2,302 2,195 2,476 1,778 2,618 2 65 16,012 HeART 613 574 653 446 325 561 21 45 233 68 3,539 FlRA cl 708 1,363 1,505 1,999 1,153 1,635 770 368 476 345 10,322 FIRA-Othcr dI 4,244 2,033 2,898 4,619 5022 366 1298 1 351 710 606 23,149 Subtotal 12,271 9,284 10,974 14,642 10,125 9,540 7,117 15,577 2,518 2,393 94,440 Other trust funds and dev. banks 172 102 78 29 46 51 200 20 26 0 724 TOTAL 12,443 9,385 11,052 14,671 10,171 9,592 7,317 15,597 2,541 2,393 95,165 aJ Nominal pesos converted to nominal dollars, then deflated by US consumer price index, base year 1992. bl In 1990, the Government paid off a substantial amount of IlANRURAL's external dcbt (see 1990 audited financi,11 statcments, p. II). Outstanding debt balances decreased from US$2.4 billion ill 1988 to LJS$267 million in 1990. cI I,)X3-19IU~ data includes budgetary transfers 10 BANPESCA. BANPESCA assets transferred to FOPESCA (a FIRA trust fllnd) in 1989. dI it represents the cost to the Treasury of BAN XI CO's rediscounts ofFlRA portfolio and loans from other Government institutions at concessionary terms. RllRFIN.Wnt  MEXICO AGRICULTURAL SECTOR MEMORANDUM Annex 2~ Table 4 Fiscal and QUilsi~Fiscal '0 Transfers Development Finance Institutions (1983-1992) (Millions ofN$) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1983~92 % of total BANRURAL 105 125 200 493 310 2,090 1,811 7,900 949 1,309 15,293 53% ANAGSNAGROASEMEX 24 44 97 215 476 1,149 991 1,848 2 65 4,911 17% FICART 12 18 33 42 71 261 12 31 202 68 748 3% FIRA 14 43 76 187 250 759 429 260 412 345 2,174 100/" FIRA-Other 82 65 146 432 1089 170 723 954 614 606 4,880 17% Subtotal 236 296 55} 1,369 2,195 4,430 3,965 10,992 2,179 2,393 28,607 99% Other trust funds and dev. banks 3 3 4 3 10 24 112 14 22 0 195 1% TOTAL 239 299 555 1,372 2,205 4,453 4,077 11,007 2,202 2,393 28,802 100% Tolaltransfers as percentage of AgGDP 17.0% 11.7% 12.8% 18.3% 13.0% 14.4% 10.1% 20.1% 3.3% 2.7% Total transfers as percenl!tge of GDP LJ% 1.0% 1.2% 1.7% 1.1% 1.1% 0.8% 1.6% 0.3% 0.2% I _ _- ... .... - I RlIRFIN.WBl  MEXICO AGRICULTURAL SECTOR MEMORANDUM Annex 2 - Table 5 ,risc,,' Tnmsl'crs to Tl'lIst II'muls and ))evclopmcllt "anks l"inancing Agriculture. Livestock and Fisheries Production • (1983-1992) (In millions of 1992 US$) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1983-921 FIPESCO 10 --- --- --- --- --- --- --- --- --- 10 AGROINDUSTRIA 6 7 --- --- --- --- --- --- --- --- 13 FIPROR 2 1 I --- --- --- --- --- --- --- 5 FEDA 1 1 0 --- --- --- --- --- --- --- 3 PRODEL 6 7 4 4 7 10 16 --- --- --- 53 FINA 0 0 0 0 0 0 31 0 0 0 31 FIMAIA 14 10 14 2 2 3 4 6 8 0 62 Total 39 26 20 6 9 12 51 6 8 0 176 (In millions of 1992N$) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1983-921 FIPESCO 43 --- --- --- --- --- --- --- --- --- 43 ' AGROINDUSTRIA 28 25 --- --- --- --- --- --- --- --- 54 FIPROR 8 6 5 --- --- --- --- --- --- --- 19 FEDA 6 3 2 --- --- --- --- --- --- --- II PRODEL 25 27 17 18 34 40 63 --- --- --- 224 PINA 0 0 0 0 0 0 122 0 0 0 122 FIMAIA 61 41 55 11 12 11 15 20 26 0 252 Total 172 102 78 29 46 51 200 20 26 0 724 Source: "Ingresos'i Egresos de Enlidades Financieras Relacionadas con los Seclores Agropecuario 'i Pesquero, 1982-1992,· Direccion General de Programacion 'i Presupueslo Agropecuario, Pesquero 'i Abaslo - Direccion de Inlegracion 'i Evaluacion, SHCP. Sepliembre de 1993. • Excluding BANRURAL, FIeART, ANAGSA/AGROASEMEX. FIRA and BANPESCA. TRUSTJilJN.WBl  MEXICO AGRICULTURAL SECTOR MEMORANDUM Annex 2 - Table 6 Governmcnt Transfcrs and AgricuUurall.cnding Thrull~h FillA and DANRURAL Annual Lending 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 (Millions of 1992 US$) FIRA 1,433 1,805 2,034 1,131 1.109 1,403 1.720 2,163 2.826 3,358 BANRURAL 1.913 2,442 2,627 1,756 1,569 2,064 1,809 1,233 1,009 1.276 TOTAL 3.347 4.247 4,662 2,887 2,678 3,467 3,529 3,396 3,835 4.634 Government Tranden as percent of Annual Lending FIRA - US! per dollar discounted by FIRA 0.78 0,48 0.55 1.14 1.08 0.35 0.31 0.21 0.12 0.09 BANRURALIFICART - US! per dollar lent by BANRURAL 0.71 0.47 0.45 0.64 0.22 0.59 0.46 2.46 0.39 0.35 TOTAL - US! per dol/ar lent 10 the seclor 0.74 0.48 0.49 0.84 0.57 0.49 0.39 1.03 0.19 0.16 Government Transfers as percent of Annual Lending (including ANAGSAJAGROASEMEX) 1I1RA - US$ per dol/ar discollnted by FIRA 0.78 0.48 0.55 1.14 1.08 OJ5 0.31 0.21 0.12 0.09 BANRURAL/FICART/ANAGSAIAGROASEM - USS per dollar lelll by BANRURAL 0.86 0.62 0.64 0.89 0.49 0.89 0.72 3.03 0.39 0.37 TOTAL (bldudlne ANAGSAIAGROASEMEX) - US! per dollar lelltlo the sector 0.83 0.56 0.60 0.99 0.73 0.67 0.52 1.24 0.19 0.17 -_.- RURFIN.WDI I  MEXICO AGRICULTURAL SECTOR MEMORANDUM Anllex 2 - Table 7 Government Transfers to DANRURAL (In millions ofN$) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1983-92 Government Budget Transfers 96 47 96 338 523 950 1,029 795 883 1,309 6,066 GOM Assumption of External Debt *1 8 78 104 155 (213) I 140 783 7,105 67 NA 9,227 Total Transfers to BANRURAL 105 125 200 493 310 2,090 1,811 7,900 949 1,309 15,293 *' These figures represent payments accrued to cover annual interest charges and repayment of capital on BANRURAL's external debt In 1990, the Government paid off a substantial amount of BANRURAL's external debt (see 1990 audited financial statements, p. 11). Outstanding debt balances decreased from US$2.4 billion in 1988 to US$267 million in 1990. Annex 2 - Table 8 Table 8. BANRURAL's Operating Costs 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1983-92 Millions ofN$ 43 70 104 224 463 930 1,567 1,878 1,748 1,770 8,797 Millions of 1992 N$ 2,225 2,197 2,078 2,395 2,137 2,003 2,812 2,661 2,019 1,770 22,297 Millions of 1992 US$ 502 562 530 468 415 485 721 717 596 572 5,568 Sources: "Analisis Instilucional del BANRURAL para el Estudio sobre el Credito Agropecuario y Mercados Financieros Rurales ell el Agro-Mexicano." A Audited Financial Statements, and direct information provided by BANRURAL. BANRURAL.WBI  MEXICO AGRICULTURAL SECTOR MEMORANDUM Annex 2 - Table 9 Govemment T"ansfers to FICART (In million ofN$) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1983·92 I. Annual Transfers 6.2 7.0 7.8 6.5 9.6 15.5 11.7 31.5 35.7 J 1.4 162.8 2. Capilli I Contributions (External Loan Disbursements) 5.6 11.3 25.0 35.2 6\.0 245.1 0.0 0.0 166.0 36.5 585.7 (a) lOB 48 6.0 19.9 34.3 30.3 NA NA NA NA NA NA (b) IBRD ........ 0.8 5.3 5. t 0.9 30.6 NA NA NA NA NA NA TOTAL TRANSFERS. _ 11.8 18.3 32.8 41.7 70.5 260.6 11.7 3 \.5 201.7 67.9 748.5 Annex 2 - Table to F1CART's Opera'in! and Technical Assistance Costs 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1983·92 I ADMINISTRATIVE COSTS Millions of N$ 1.05 1.50 1.93 3.17 8.00 20.06 20.63 68.62 75.31 36.06 236.34 Millions of 1992 N$ 54.66 47.00 38.41 33.91 36.92 43.21 31.02 97.24 87.00 36.06 511.45 Millions of US$1992 12.33 12.03 9.79 6.63 7.17 10.47 9.49 26.21 25.70 11.65 131.46 TECHNICAL ASSISTANCE Millions of N$ 0.15 0.39 0.41 0.87 2.08 4.18 NA NA NA NA 8.08 Millions of 1992 N$ 8.05 12.33 8.09 9.32 9.59 9.00 NA NA NA NA 56.37 Millions of US$1992 \.81 3.16 2.06 1.82 1.86 2.18 NA NA NA NA 12.90 Sources: "Analisis Institucional del BANRURAL para el Estudio sobre el Credito Agropecuario y Mercados Financieras Rurales en el Agro Mexicano." August 1989 Direccion General de Programacion y Presupuesto Agropecuario. Pesquero y Abasto, SHCP. September 1993. FICART Audited Financial Statements. FICART.WBI  MEXICO AGRICULTURAL SECTOR MEMORANDUM Annex 2 ~ Table 11 Fiscal Transrers to ANAGSA and AGROASEMEX (1983·1992) (In millions of NS) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1983·92 CASH INFLOWS I. Tolal Govenunent premia 11 37.7 59.7 82.8 142.6 335.8 679.9 1,098.8 1,643.7 1.5 65.0 4,147 a) Agriculture 21 36.4 57.5 78.1 135.5 322.2 644.3 1,045.9 0.0 0.0 0.0 2,320 b) Olher 21 1.2 2.2 4.7 7.1 13.7 35.6 53,0 0,0 0.0 0.0 117 2. Total Farmers" premia II 10.0 22.9 51.3 99.9 253.5 485.2 467.4 0.0 0.0 0.0 1,390 a) Agriculture 21 8.7 20,4 46.3 92.4 239.6 448.1 406.0 0.0 0,0 0.0 1,26\ b) Other 21 1.2 2.6 5.0 7.6 lJ.9 37.1 61.4 0.0 0.0 0.0 129 TOTAL INFLOWS 47.6 82.6 134.1 242.6 589.3 1,165.0 1,566.2 1,643.7 1.5 65.0 5,538 CASH OUTFLOWS 3. Total Administrative Costs II 5.3 8.9 16.7 31.4 73.5 148.8 285.0 53.6 15.3 82.2 721 a) Agriculture 21 5.0 8.4 15.5 29.5 70.0 139.5 264,2 0.0 0.0 0.0 532 b) Other 21 0.3 0.5 1.2 1.9 3.4 9.3 20.8 0.0 0.0 0.0 37 4. Tolal Inderrmities II 28.8 58.4 1l1.5 283.8 656.0 1,485.8 1,172.9 0.0 88.5 147.4 4,053 a) Agriculture 21 27.5 56.2 126.7 275.3 642.8 1,451.1 1,102.1 0.0 0.0 0.0 3,682 b) Other 21 1.2 2.3 4.8 8.5 13.2 34.7 70.8 0.0 0.0 0.0 135 TOT AL OUTFLOWS 34.1 67.3 148.2 315.2 729.5 1,634.6 1,457.9 53.6 103.8 229.6 4,774 5, Other Govenunenl Transfers 31 (13.5) (15.3) 14.1 72.6 140.2 469.5 (108.3) 0.0 0.0 0.0 559 a) Agriculture 21 (12.6) (13.4) 17.9 76.9 151.1 498.2 (85.5) 0.0 0.0 0.0 633 b) Other 21 (1.0) (2.0) (3.8) (4.3) (10.9) (28.7) (22.8) 0.0 0.0 0.0 (73) 6. Total Cosl 10 Govenunent [I + 5] 24.1 44.4 96.9 215.3 476.0 1,149.4 990.5 1,643.7 1.5 65.0 4,707 a) Agriculture 21 23.8 44.2 95.9 212.4 473.2 1,142.5 960.3 0.0 0.0 0.0 2,952 b) Olher 21 0.3 0,2 \.0 2.8 2.8 6.9 30,2 0.0 0.0 0,0 44 Number of heclares insured (million) 6.8 6.1 7.0 7.1 7.3 6.6 4.9 NA 0.1 NA Number ofheclares inderrmifiedjmillion) 4.0 3.6 4.4 5.4 5.5 5.2 3.1 NA NA NA Source: Derived from P. Hazell and M. Gudger. World Bank internal memorandum, November 1. 1989. 11 From data provided by SHCP in 1989·90. 21 Division between agriculture and others is based on total indemnities paid each year. 31 Inflows less outflows. ANAGSA.WBI  MEXICO AGRICULTURAL SECTOR MEMORANDUM Annex 2 - Table 12 Fiscal Transfers to FIRA (In millions ofN$) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 . 1983-92 [Contributions to Capital (a) -FONDO 0.02 17.26 24.47 47.36 69.69 157.19 9.01 6.24 (1.81 ) 9.27 338.72 -FEFA 2.26 14.55 40.76 111.72 145.49 544.49 185.90 10.42 3.71 19.29 1.078.58 -FEGA 1.76 402 6.6] 11.84 26.22 40.13 66.00 124.20 108.76 0.00 ]89.57 Subtotal 4.04 35.83 71.86 )70.91 241.40 741.81 260.92 140.87 110.66 28.56 1,806.87 Subsidies (b) (c) - FONDO 0.08 0.00 0.00 0.00 0.00 0.00 0.00 0.71 30.00 30.64 6\.43 - FEFA 2.52 0.00 0.00 0.00 0.00 0.00 0.00 5.52 126.24 118.55 252.83 - FOPESCA· 5.78 5.66 n.a. 10.15 0.00 0.00 46.12 1.51 4.11 . 0.42 19.88 Subtotal 8.39 5.66 0.00 10.15 0.00 0.00 46.12 13.81 160.41 149.60 394.14 Transfer to cover payment ofFEGA guarantees (b) (c) 0.28 0.68 0.81 358 6.00 15.81 94.65 80.42 112.71 166.43 481.37 Transfers to cover debt payments (b) 0.91 1.20 2.90 2.34 2.61 1.59 21.18 24.57 2855 N.A. 91.90 TOTAL in millions of New Pesos \3.6 43.4 15.6 181.0 250.1 759.2 428.9 259.7 412.3 344.6 2,174.3 III millions of 1992 New Pesos 107.9 1,362.8 1,505.0 1,999.4 1,153.4 1,635.2 769.7 368,0 476,3 344.6 10,322.1 In millions of 1992 -- - -- US$ - .... - .. 159,6 348.9 383.6 390.7 223.9 396.3 197.2 99.2 140.7 1l1.3 2,451.5 Sources: (II) FIRA, Audited Financial Statements. (b) FIRA. direct infonnation, fax:es dated May 19 and 22, 1992. (c) For 1992, FlRA, Audited Financial Statements 1992-91. Notes: • Includes transfers to I3ANPESCA from 1983-1989. as reponed in "Ingresos y Egresos de Entidades Financieras Relacionadas con los Sectores Agropecuario y Pesquero 1982-1992" prepared by Direccion General de Programacion y Presupuesto Agropecuario. Pesquero y Abasto, SltCP, Sept. 1993 . •* Nominal pesos convened to dollars, then deflated by US Consumer Price Index. FIRA.WBt  MEXICO AGRICULTURAL SECTOR MEMORANDUM Annex 2 - Table 13 FIRA: FONFO-FEFA-FOPESCA Quasi-fiscal Cost of Concessionary Loans from Goverment (Millions of N$ ) Subsidy Subsidy Subsidy Subsidy Average Interest Interest Interest Interest Average Implicit CETES or Interest Differ. DitTer. Differ. Differ. Year Beg. Ending Debt Interest Interest Rate of Expense Current 1992 Current 1992 Balance Balance Balance Ex ense Rate Inflation Reinstated N$ N$ US$ mill. US$ mill. 1983 113 135 124 19 15.0% 80.8% 100 (82) (4,244) (680) (957) 1984 135 199 167 34 20.4% 59.2% 99 (65) (2,033) (386) (521) 1985 199 342 270 48 17.9% 71.7% 194 (146) (2,898) (566) (739) 1986 342 650 496 92 18.6% 105.7% 524 (432) (4,619) (706) (903) 1987 650 1,049 850 264 31.0% 159.2% 1,352 (1,089) (5,022) (790) (975) 1988 1,049 1,512 1,281 500 39.0% 52.3% 670 (170) (366) (75) (89) 1989 1,512 3,302 2,407 361 15.0% 45.0% 1,084 (723) (1,298) (294) (332) 1990 3,302 5,644 4,473 604 13.5% 34.8% 1,557 (954) (1,351) (339) (364) 1991 aJ 5,644 6,989 6,316 604 9.6% 19.3% 1,219 (614) (710) (204) (210) 1992 aJ 6,989 9,673 8,331 695 8.3% 15.6% 1,301 (606) (606) (196) (196) 1983-92 3!222 8,10 I {4!880) (23, 142).~ J4,23~) (5,285) aJ Outstanding loan balances do not include an outstanding debt with the Federal Goverment for N$I ,281.5 million and NSI,255.3 million in 1991 and 1992, respectively. bl Weighted average 28-day CETES. When the annual average CETES rates were negative, the rate of inflation was as the lower bound opportunity cost of funds. Sources: FIRA Audited Financial Statements. Source for interest paid: Table 5 of Annual Reports ('Costos Financieros' ). FIRA.WBl  MEXICO AGRICULTURAL SECTOR MEMORANDUM Annex 2 - Table 14 FIRA - Breakdown of Costs to Government (Millions ofN$) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1983-92 1 Technical Assistance Reimb. to Commercial Banks and Producers 0.9 2.0 3.8 5.9 9.0 33.9 24.3 51.6 54.9 47.6 234 2. Deficit in Guarantee Payments 0.3 0.7 0.8 3.6 6.0 15.8 94.6 80.4 112.7 166.4 481 3. Operating Costs 2.8 4.5 6.8 11.7 28.8 57.6 81.2 111.3 157.7 131.2 594 (Millions of 1992 N$) 1. Technical Assistance Reimb. to Commercial Banks and Producers 47 63 76 63 42 73 44 73 63 48 590 2. Deficit in Guarantee Payments 15 21 16 38 28 34 170 114 130 166 733 3. Operating Costs 143 142 136 125 133 124 146 158 182 131 1,420 (Millions of 1992 US$) 1. Technical Assistance Reimb. to Commercial Banks and Producers 11 16 19 12 8 18 11 20 19 15 149 2. Deficit in Guarantee Payments 3 5 4 7 5 8 44 31 38 54 200 3. Operating Costs 32 36 35 24 26 30 37 --- 42 54 42 360 aJ Excluding the cost of intermediation margins of commercial banks. Source: FIRA annual reports, financial statements, and direct information. FIRA.WBI .f  MEXICO AGRICULTURAL SECTOR MEMORANDUM Annes 2 - Table 15 BANRURAI.. Loss in Value of Government Transfers (Millions of NS) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 Beginning Balance 11 18 61 143 283 651 817 2,751 4,004 10,923 10,826 (+) Net Income/Loss Defore GOM Transfer (62) (42) (60) (126) (144) (157) (558) (982) (1,046) (1,228) (+) GOM Armual Transfers 96 47 96 338 523 950 1,029 795 88] 1,]09 (+) GOM Assumption of Banrural's Foreign Debt 8 78 104 155 {213) 1,140 78] 7,105 67 NA Ending Balance 61 143 2K3 651 817 2.751 4,004 10,923 IO,K26 10,908 Accounting Net Worth (reported) 19 30 99 211 438 2448 3953 NA NA ] 217 I (Millions of 1992 NS) Beginning Balance 1,399 3,649 6,238 9,026 12,958 13,724 17,889 20,138 29,942 29,830 (+) Net IncomeILoss Before GOM TransFer (3,203) (1,]31) (1,201 ) 0,343) (662) (338) (1,001) (1,]91) (1,208) (1,228) (+) GOM Armual Transfers 5,012 1,475 1,919 3,614 2,413 2,046 1,847 1,126 1,019 1,]09 (+) GOM Assumption of Banrural's Foreign Debt 441 2,445 2,070 1,660 (984} 2,456 1,404 10 1068 77 0 Ending Balance 51 3649 6,238 9,026 12,958 13724 17889 20138 29942 29,830 29912 (Millions of 1992 NS) Accounting Net Worth (reported) 825 8\0 1,655 1.715 1,375 5,063 6,831 NA NA 3,217 (-) Calculated Ending Balance 3649 6238 9026 12958 13,724 17 889 20138 29942 29,830 29912 Acc.umulated Capital Losses (1983-1992) (2,825) (5,428) (7,371 ) (11,242) (12,349) (12,826) (13,308) (29,942) (29,830) (26,694) Annual Capitall.osses (2,825) (2,604) (1,943) (3,871 ) (1,\07) (476) (482) (16,634) 112 3,\36 AIUlual Capilall.osscs + Annual Operating Losses (6,021) (3935) (3 143) (5215) (I 769) (814) (I 483) (18,025) (1,097) 1,908 Accumulated Losses of GOM Transfers (6027) (9962) (13 \05) (18320) (20089) (20903) (22386) (40,411) _(41,5011) Q?6_00) Millions of Millions of 1992 N$ 1992 US~ I. Tolal Equity Jan. I, 1983 1,399 21,066 2. GOM Tolal Transfers 1983-92 41,419 10,246 3. Net Operating Losses 1983-1992 (12,906) (3,227) 4. TOlal Net Intlows (2 + 3] 28,513 7,019 5. Tolal Available (I + 4] 29,912 28,085 6. Total Equity Dec. 31, 1992 3,217 1,033 7. Net Loss of GOM Transfers [6-5+3 ] 1-(:19,600)-- (30,279» 11 1983 represents lolal equity as reported in BANRURAL's balance sheet at year end 1982. BANRURAL.WBI  MEXICO AGRICULTURAL SECTOR MEMORANDUM Annex 2 - Table 16 FIRA. Loss in Value of Government Transfers 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1983-92 (In millions of 1992 N$) Beginning Balance 1,611 1,859 3,267 5,272 8,004 10,149 12,406 13,978 15,114 15,978 87,639 (+) Net Income before GOM Transfers 37 282 574 904 1,032 659 1, I 03 937 735 602 6,866 (+) GOM Annual Transfers aJ 210 I 126 1,431 1 828 1,114 1,598 468 200 128 29 8,13 I Ending Balance 1,859 3,267 5,272 8,004 10,149 12,406 13,978 15,114 15,978 16,608 --- Accounting Net Worth (reported) 1,044 1,823 2,693 3,355 2,670 3,832 4,601 4,543 4,605 3,467 --- (-) Caculated Ending Balance 1,859 3,267 5,272 8,004 10,149 12,406 13,978 15,114 15,978 16,608 --- Accumulated Capital Losses (1983-92) (815) (1,443) (2,579) (4,649) (7,479) (8,574) (9,377) (10,571) (II ,372) (13,142) --- Annual Capital Losses (815) (628) (1,136) (2,070) (2,830) (1,094) (803) (1,195) (801 ) (1,769) (13,142) Millions Millions of of1992 1992 HI US$bl 1. Total Equity Jan. 1, 1983 1,611 308 2. GOM Capital Contributions 1983-92 8,130 1,882 3. Net Operating Income before Operating Subsidies 1983-92 6,866 1,710 4. GOM contributions to income aJ 2 0 5. Total Net Inflows [2 + 3 + 4} 14,997 3,592 6. Calculated Equity at Dec. 31, 1992 [I +5] 16,608 3,900 7. Total Accounting Equity Dec. 31, 1992 3467 1,113 8. Loss in Value ofGOM Transfers [6-7] ! (13:142) (2,788>1 at Includes transfers to cover operating subsidies, payments ofFEGA guarantees, and paymentsofFlRA's debt. Excludes FOPESCA (equity data not available). b/ Nominal pesos converted 10 dollars, then deflated by US CPI index, base 1992. FlRA.WBl  MEXICO AGRICULTURA L SECTOR MEMORANDUM Annel 2 - Tilble 17 i!:s.imil.cd Itinnncitil nCUe". An'rllcd In Uorruwcrs un'IItA and BANIWRAL on Government Funds al On Millions ofN$) 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1983-92 FIRA IS Borrowers Average Portfolio of Discounts 105 179 315 449 825 1,881 3.007 4,833 7,901 10.906 Estimated Interest Paid by Borrowers to Conunercial Banks bl 24 47 87 198 560 893 1,141 1,460 1,262 2,027 'Estimated Amount of Interest Borrowers would have paid at Market Rates cl 91 116 218 500 1,359 1,376 1,507 2,057 2,217 2,647 Estimated Amount of Borrower Benefits 67 69 132 302 799 482 366 597 954 620 4,388 BANRURAl, I S Borrowers Average Loan Portfolio dI 222 354 608 969 1.893 3,743 4,763 5,360 5,487 5.587 Interest Paid by Borrowers eI 41 77 141 301 895 1.618 1,719 1.713 995 841 Estimated Amount of [nterest Borrowers 193 231 424 1,083 3,126 2,757 2,410 2,308 1,567 1.384 would have paid at Market Rates cl Estimated Amount of Borrower Benefits 152 153 282 782 2.231 1,138 692 596 572 543 7,142 Total Borrowers' Benefits on Credit Funded by Govenunenl 219 222 414 1,084 3,029 1.620 1,058 1,193 1,526 1,163 11,530 Benefit in millions of 1992 N$ 11,390 6,973 8,251 11,596 13,973 3,490 1,899 1,690 1,763 1,163 62,189 Benefit in 1992 US$ million ___ _ 1.~68 _L7~L_ ~.lOl __2,~66_ ~.713 846 486 456 521 376 14.120 p) It excludes commercial bank Ilarlicipalion ill financing (10-20% of the IImollnt of the credit) which. excel)t for some working cupilul lonns 10 ors. curries Ihe same interest rale of flRA funds. bl flRA's Interesllncome plus 5.5 percenlage points (average intermediation margin rIRA has given to commercial banks.) cl CPP or the rate of illOalion pIllS 5.5 percentage points for flRA borrowers, 6 percentage poiuts for 13ANRURAL borrowcrs. tV Average balances of BANRURAL's portfolio in 1986 through K9 arc anuual averages based on monthly balances; lor other yellrs, ligures representlwo-),car averages of outstanding loan balances. The balance for 1991 is prior to the transfer of part of BANRURAL's portfolio to FIRCAVEN and PRONASOL. e/ Equal annual interest income on portfolio. nIn') VI "'-I ''lInt  ANNEX 3 Fiscal Policies Background Information List of Tables Appendix 1: Rural Development Expenditures by Institution Appendix 2: Rural Development Expenditure Data Table 3: Rural Development Expenditures by Program (N$ millions) Table 4: Rural Development Expenditures by Program (in percentages) Table 5: Rural Development Expenditures by Category (N$ millions) Table 6: Rural Development Expenditures by Category (in percentages) Table 7a: SARH Expenditures by Program and Category (N$ thousands), 1985-88 Table 7b: SARH Expenditures by Program and Category (N$ thousands), 1989-92 Table 8: SARH Expenditures by Program (N$ millions) Table 9: SARH Expenditures Composition by Program (in percentages) Table 10: SARH Delegation Expenditures (N$ thousands) Table 11: Human Resources Evolution by Institution (SARlI, SRA, SEPESCA) Table 12a: SARH Transfers to Dependent Institutions (N$ millions), 1982-87 Table 12b: SARH Transfers to Dependent Institutions (N$ millions), 1988-92 Table 13: CNA Expenditures by Program (N$ thousands), 1982-92 Table 14: SRA Budget by Program and Category (N$ millions and percentages) Table 15a: SRA Expenditures by Program and Category (N$ thousands), 1985-88 Table 15b: SRA Expenditures by Program and Category (N$ thousands), 1989-92 Table 16a: SRA Transfers to Dependent Organizations (N$ thousands) Table 16b: SRA Transfers to Dependent Organizations (N$ thousands) Table 17: Procuraduria Agraria - Federal Expenditures (N$ thousands) Table 18: SEPESCA Expenditures by Program (N$ millions and percentages), 1982-92 Table 19a: SEPESCA Expenditures by Program and Category (N$ thousands), 1985-88 Table 19b: SEPESCA Expenditures by Program and Category (N$ thousands), 1989-92 Table 20a: SEPESCA Transfers to Dependent Organizations (N$ millions), 1982-87 Table 20b: SEPESCA Transfers to Dependent Organizations (N$ millions), 1988-92 Table 21: SEPESCA Delegation Expenditures (N$ thousands) Table 22: CONASUPO Expenditures and Revenues (N$ thousands) Table 23: Breakdown of CONASUPO Expenditures and Revenues (N$ thousands) Table 24: Solidarity and Regional Development Expenditures (N$ millions) Table 25: Solidarity and Regional Development Expenditures (percentages) Table 26: CFE Expenditures for Rural Electrification (N$ thousands and percentages) Table 27: SCT Rural Development Expenditures (N$ thousands and percentages) Table 28a: SCT Expenditure by Program and Category (N$ thousands and percentages), 1985-88 Table 28b: SCT Expenditure by Program and Category (N$ thousands and percentages), 1989-92 Appendix 3: Regression Results for the Distribution by State of Solidarity, Revenue Sharing and SARB Expenditures  ANNEX 3 Appendix 1 Page 1 of 14 Federal Rural Development Expenditures by Institution 1. Federal programs in the areas of agriculture, fisheries, rural credit, land refonn, rural roads, rural electrification, and other basic infrastructure are reviewed from 1982 to 1992.1 This involved collecting data from SARR, SRA and SEPESCA and their dependent organizations. In addition, infonnation was collected from SEDESOL (Solidarity), SHCP (credit), BANXICO (credit), SECOFI (CONASUPO), CFE, SCT, FERTIMEX and INEGI (land titling) for programs with a significant rural focus. The expenditure data is classified by program and functional category (recurrent or capital). Unless otherwise indicated, the data reflects actual expenditures taken from the national accounts (Cuenta PUblica) prepared by SHCP. All federal expenditures at the state and local levels as well as transfers to dependent institutions are included. All values are in constant 1992 prices. 2. This Annex examines expenditure trends for each institution and Annex 3 contains the data on rural expenditures. Chapter 7 evaluates the impact of rural development expenditures on efficiency and poverty alleviation, while the analysis of the fiscal and quasi- fiscal transfers to rural credit and insurance schemes is presented in Chapter 6. I. SARH 3. Given the importance placed on market interventions during the de la Madrid Administration, SARR's budget was reduced more severely than the overall rural development budget; the institution's budget declined by 56% between 1983 and 1988. In contrast, during the early years of the Salinas Administration it was relatively favored and rose by 22 % between 1989-1992. By 1992, SARR and its related agencies consumed 34% of the total rural development budget (Annex 3, Appendix 2, Tables 1 and 2.) 4. Investment and Recurrent Ex:penditures. 2 Since 1985, the share of SARR's expenditures allocated towards investments has declined, dropping from 63 % in 1985 to 48 % in 1992 (Annex 3, Appendix 2, Tables 7 a and b.) Falling expenditures for water investments is the principal source behind the drop, given its large share in the overall budget (approximately The definition of rural development expenditures used in this paper differs from the GOM definition which includes only expenditures by SARH, SRA, and DFIs. 2 The SHCP Cuenta PUblica presents two different data series for recurrent and capital expenditures. The data presented here is based on the expenditure breakdown by program and category (e.g., recurrent or capital) and is only available since 1985. It differs from the tables presenting aggregated capital and recurrent expenditure data.  ANNEX 3 Appendix 1 Page 2 of 14 one third). Research and technology as well as marketing investments have also decreased. since 1985, whereas infrastructure and education expenditures have risen. 5. Programs. Water development, administration and services accounted for an average of 75 % of SARH's total budget between 1983 and 1992 (Annex 3, Appendix 2, Tables 8 and 9.) While allocations for the remaining five programs, infrastructure, policy, education, research and technology and marketing each accounted for an average of 5 % or less of SARH's budget. 6. Water Development. Water development is the most important program in SARH's budget, consuming an average of 42% of total expenditures during 1982-1988, and slightly less, 36%, during 1989-1992. Water development has 5 subprograms: water management; drinking water, drainage; rainfed areas development; and irrigated areas development. Only 30% of the costs for water management and drinking water are assumed to occur in rural areas. As the water development program is implemented by CNA it is analyzed below (see para. 21.). 7. Administration. Administrative costs are the second most important category accounting for 28% of total SARH costs from 1983 to 1988 and for 24% during the first four years of the Salinas Administration. Nonetheless, administrative costs declined during the de la Madrid Administration dropping to 40% of their 1983 level in 1988, while in 1992 they were 140% of their 1989 level. The majority of administrative expenditures, about 75%, are recurrent costs, mainly for personnel. The decline in administrative expenses during 1983-1988 reflects the decrease in personnel expenditures (Annex 3, Appendix 2, Table 11.) The latter dropped from 70% of SARH's recurrent costs to 62% in 1988, due to a 16% decline in the number of SARH staff (excluding dependent agencies). In contrast, personnel costs were 7% higher in 1992 compared to 1989, while staff numbers have continued to decline under Salinas. Staff numbers in 1992 were 73 % of the 1989 level, suggesting more attractive remuneration policies and/or higher non-wage personnel expenses for SARH after 1988. 8. Services. The third most important program is services. The share of SARlI expenditures allocated to services rose from an average of 13 % during the de la Madrid Administration to 21 % for 1989-1992. Expenditures for plant and animal health, extension and promotion activities are included in this category. During the early 1980s, extension costs accounted for the majority of expenditures in this category, however, after 1985, extension costs appear to drop to zero and promotion expenditures became the largest item. While SARlI began restructuring its extension service in 1985, it is impossible that SARH incurred no extension costs since 1985. The probable scenario is that extension programs became reclassified and appear in the promotion category, which is fairly broad and non-descriptive. For instance, part of ASERCA's costs may be classified in the promotion subprogram, as the agency's costs are not clearly reflected in any given category. 9. Marketing. The marketing program traditionally accounted for less than 1 % of  ANNEX 3 Appendix 1 Page 3 of 14 SARH's budget with the exception of 1990-91, when it rose to close to 5 %. This brief surge reflected an increase in marketing assistance for agroindustrial crops which previously had been supported by the guarantee price system managed by CONASUPO. In 1992, marketing support was provided by ASERCA, but appears to be classified as a promotional program for the sector. The other component of the marketing program, marketing and storage infrastructure received only small budgetary allocations in 1989 and 1991-92. 10. Other programs. About 5% of SARH's budget is devoted to research and technology expenditures, approximately 85 % of which is classified as investment. Infrastructure expenditures for livestock, agriculture, forestry and rural roads combined consumed about 1 % of SARH budget under de la Madrid Administration and a slightly higher share, 3 %, during 1989-1992. Most of the expenditures within this category are for agriculture followed by livestock infrastructure; less than 10% of infrastructure expenditures are allocated to forestry and rural roads combined. The share of investment in the infrastructure program varies since 1985 but has averaged 87 %. Education programs have consumed less than 5 % of the SARH budget since 1983 and have mainly supported the four agricultural universities and the national program to train agricultural labor. Slightly less than 90% of education costs are for recurrent. The share of expenditures allocated to policy analysis has hovered around 1 %, with the exception of 1992, when it surged to 16%, reflecting increased expenditures to negotiate NAFrA and prepare the regulatory framework for the ejido reform of Article 27 of the Constitution and the new forestry and water laws. The majority (66%) of policy expenditures cover recurrent costs. 11. State Delegations. Real expenditures by the state delegations have declined significantly as a share of SARH's total budget (Annex 3, Appendix 2, Table 10.) In 1989 they were 29% of total SARH expenditures, while in 1992, they accounted for 11 %. It is not clear whether the lower expenditures by SARH's state delegations reflect a smaller role of SARH's state offices or an increasing decentralization of rural development responsibilities to the state governments. The share of expenditures in the states rises significantly, if the budgets of the dependent agencies are excluded. In a SARH report of October 1992, the state delegations are cited as spending 66 % of the budget and the central administration 34 %.3 12. Transfers. During the de la Madrid Administration, SARH transfers to dependent agencies remained constant at 30% of the total SARH budget,4 (Annex 3, Appendix 2, Table 12.) Transfers in 1988 were 60% smaller than their 1982 level, reflecting the overall decline in SARH's budget and the privatization or liquidation of many dependent organizations in the 1980s. Excluding transfers to institutions undergoing or having completed liquidation or 3 SARH, FederalizaciOn y ReestTUcturaci{m de fa Secretaria de Agricultura y Recursos HidraUlicos, (1992). 4 Includes full value of transfers to the regional water commissions and CNA for drinking water and water management.  ANNEX 3 Appendix 1 Page 4 of 14 privatization, the regional water commISSIons, INIFAP, FIRCD, INMECAFE and the Autonomous University of Chapingo (AUC) accounted for an average 80% or more of total transfers during the de 1a Madrid Administration. 13. In the Salinas Administration, transfers to dependent institutions climbed to 70% of SARlI's total budget by 1992 increasing by 120% over the 19891evel. The increase in transfers reflects the creation of CNA (1989) and ASERCA (1991), which by 1992 together accounted for 74% of total transfers. Transfers to other institutions decreased during this period and by 1992 INIFAP, FIRCD and AUC together consumed 16% of transfers, compared to 30% in 1982. Under de 1a Madrid, capital costs were an average of 54% of total transfers, while between 1989 and 1992 they were an average of 72 %. The decreased recurrent cost transfers reflects declining expenditures for marketing and processing parastatals, as well as for educational institutions and an increasing budget for CNA and water development. The creation of ASERCA in 1991 and more recently PRDCAMPO is likely to shift the composition of transfers towards recurrent costs. 14. CNA. An examination of CNA's budget follows, given the agency's budgetary importance. CNA was created in 1989 and charged with the development, operation and maintenance of major irrigation, drainage and drinking water infrastructure..5 Its budget has grown rapidly, and by 1992, it was a 100% greater than its 1989 level and accounted for 37% of SARlI's total budget. 15. In 1992, about 82 % ofCNA's budget was allocated for water resources development and 13% covered administrative expenses (Table 1 and Annex 3, Appendix 2, Table 13.) Research, promotional activities, policy and infrastructure each received 2 % or less of the total CNA budget. Within the water program, the development of irrigated areas consumes 56% of the total, which is down from an average of 70 % during the de 1a Madrid Administration. The second most important sub-program is drinking water with 25 % of the total. Management of the national hydraulic system accounts for 11 % of the budget, up from an average of 1 % for the period 1983-1988. In 1985, SARH initiated a program to expand small-scale irrigation in rainfed areas. However, resources allocated to the program have declined from the initial 15 % to an average of 5 % since 1988. Drainage programs also have accounted for 5 % or less of the water program. 16. CNA has sought to transfer the costs of operation and maintenance to organized producers, however, this policy has not been implemented evenly across all water programs. For example, in the irrigated areas subprogram, approximately half of the resources are devoted to rehabilitation and operations and maintenance. In contrast, in the drinking water, drainage and rainfed areas subprograms less than 1 % of their budgets are allocated to rehabilitation, The analysis of CNA's expenditures includes the full value of expenditures for drinking water and water management.  ANNEX 3 Appendix 1 Page 5 of 14 operations and maintenance, which suggests that the fmal users are responsible for these costs. While this is logical for drinking water, which is supported by municipalities, it is unusual for rainfed areas investments; producers in rainfed areas are apt to have fewer resources to cover operations and maintenance expenditures for irrigation than producers in irrigated areas.  ANNEX 3 Appendix 1 Page 6 of 14 Table 1 CNA Expenditures by Program, 1992 (percent or total) ! Program ShareofCNA Share of Water Budget Program Budget Administration 13 Policy 2 Research and Technology 0 Transfer Infrastructure 0 Water Resources Development 82 Water Management 11 13 Drinking Water 20 25 Construction &. 24 Expansion Operations &. 1< Maintenance Rehabilitation 1< Drainage 1 1 Construction &. 1 Expansion Rainfed Areas 4 5 Construction &. 5 Expansion Irrigated Areas 46 56 Large-scale irrigation 18 Small-scale Irrigation 9 Rehabilitation 15 Operations &. 11 Maintenance Protection of Productive 2 Areas Promotion <1 Other 4 Total 100 Source: Cuenta PUblica, SHCP  ANNEX 3 Appendix 1 Page 7 of 14 II. Agrarian Reform Secretariat (SRA) 17. Responsible for the land refonn program and the provision of technical assistance to ejidos, the importance of SRA has increased with the recent ejido land refonn program. Between 1982 and 1991, SRA accounted for 2 % or less of the rural development budget. With the passage of the revised Article 27 of the Constitution, SRA became responsible for the refonn of the ejido system, and in particular, the implementation of the land titling program and the satisfactory resolution of all pending cases of land refonn. As a result, SRA's budget increased by 133% in 1992 and its share of total rural development resources increased to 5%. 18. Programs. The three principal programs of SRA are administration, land tenure and extension (Annex 3, Appendix 2, Table 14.) Reflecting the policy refonns for the ejido sector, the share of resources allocated to land tenure issues has risen from an average of 64 % during 1983-1988 to 74 % for the fIrst four years of the Salinas Administration. Within the land tenure sub-program, land titling and conflict resolution consumed an average of 56% and 44% of the program's resources from 1983 to 1991. With the passage of Article 27, the share allocated to land titling increased to 75 % in 1992. The share of total resources allocated to administration has remained roughly constant at slightly more than 20%, while the share allocated to extension has decreased sharply from an average of 15 % during the de la Madrid Administration to 3 % for 1989-1992. 19. Recurrent and Capital Expenditures. Almost all of SRA's budget (90%) is allocated to recurrent expenditures as the secretariat has no infrastructure or other investment responsibilities (Annex 3, Appendix 2, Table 15.) Personnel services are the most important category for recurrent costs. However, personnel costs have declined signifIcantly and by 1992 were 50 % of their 1983 level, in part due to a 10% decline in staff numbers. 20. Transfers. Given the creation of two new agencies under SRA to implement the land titling program and the increased importance attribution to the resolution of pending land disputes, SRA transfers increased sharply in 1992 (Annex 3, Appendix 2, Tables 16 a and b.) Prior to 1992, transfers to dependent organizations and programs averaged less than 15% of SRA's budget, while in 1992 they reached 54 % of the total budget. The additional transfers support the Procuraduria Agraria (PA) and the Registro Agrario Nacional (RAN) to implement the land titling program (see Chapter 6 on land titling) and a sixfold expansion of resources for land purchases to resolve pending land tenure disputes. Other dependent organizations under the SRA had provided various types of support to ejidos, however, by 1992 they were no longer received budgetary support. 21. State Delegations. Excluding expenditures for the federal district, almost 30% of the PA's budget was to support state delegations in 1992 (Annex 3, Appendix 2, Table 17.)  ANNEX 3 Appendix 1 Page 8 of 14 m. Secretariat for Fisheries (SEPESCA) 22. The Secretariat for Fisheries (SEPESCA), responsible for aquaculture and marine fisheries development, consumes 2 % or less of the rural development budget. Moreover, SEPESCA's expenditures were significantly reduced during both Administrations and by 1992 they were 27% of their 1983 level. Similar to SRA and especially SARB, the decline in expenditures reflects the privatization and liquidation of many of the sector's production, processing and marketing parastatals and a decline in staff numbers. 23. Programs. Administration followed by infrastructure are the two most important categories of expenditure (Annex 3, Appendix 2, Table 18.) The share allocated to marketing, agroprocessing and extension each average less than 5 % and have declined to almost zero during 1989-1992. 24. Recurrent and Capital Expenditures. The share of recurrent expenditures averaged 70% between 1985 and 1992 (Annex 3, Appendix 2, Tables 19 a and b.) The share of personnel costs in SEPESCA's recurrent costs (excluding dependent agencies) declined from 72% in 1983 to 45% in 1992, in part due to a 25% cut in staff numbers during this period. 25. Transfers. Transfers by SEPESCA to dependent agencies have declined considerably dropping from 63% in 1983 of the total budget to only 5% in 1992 (Annex 3, Appendix 2, Table 20.) The three marketing and processing parastatals and the two credit trust funds were either privatized or liquidated in the 1980s. The National Fisheries Institute, which carries out fisheries research, remains the only dependent agency for SEPESCA. 26. State Delegations. Approximately, 12 % of SEPESCA's budget in 1992 was to support its state delegations, slightly higher than the 9% of 1989 (Annex 3, Appendix 2, Table 21.) IV. National Food Company (CONASUPO) 27. CONASUPO and its affiliates have been the principal GOM instrument to implement the guarantee price system. CONASUPO would purchase com, oilseeds, legumes, wheat and other basic grains from producers at subsidized prices, would process the commodities and then sell them to consumers or industry at subsidized prices. 6 CONASUPO also had a monopoly to 6 CONASUPO's affiliates include two storage companies BORUCONSA and ANDSA, which together own over 50% of Mexico's basic food storage capacity; and dairy products producer (UCONSA); and several nationwide chains of retail grocery stores (DICCONSA). In addition, oil manufacturing parastatal (ICONSA); a com four miller (MICONSA); and a wholesale food distribution network (IMPESCA) were previously part of CONASUPO but have been either liquidated or privatized.  ANNEX 3 Appendix 1 Page 9 of 14 import these products, as well as powdered milk, to ensure the country's basic food supply. The parastatal fmanced its operations with revenues from the sale of its products, domestic and external borrowing and fiscal transfers. The latter covered operating and investment costs as well as amortization on CONASUPO's debt as of 1985. 28. Fiscal transfers to CONASUPO were significant, reaching N$10.94 billion (US$3.5 billion) in 1984, or about 30% of rural development expenditures (Annex 3, Appendix 2, Tables 22 and 23.) The fiscal burden ofCONASUPO was unsustainable, especially given the economic austerity and reform program of the de la Madrid Administration and by 1988, fiscal transfers had dropped by 65% to N$3.8 billion (US$1.2 billion). During the first four years of the Salinas Administration, fiscal transfers have remained relatively stable at approximately 20% of the total rural development budget. The only exception was 1990 when they more than doubled, reflecting a N$8.3 billion (US$2.7 billion) transfer for amortization of the parastatal's debt. On average, GOM fiscal transfers and CONASUPO's own revenues have each accounted for 36% of total revenue, with the remainder mainly from borrowing. 29. CONASUPO's total expenditures have decreased steadily, especially during the early 1980s, and by 1992 were 36% of their 1983 level. Moreover, the overall composition of the budget has shifted away from oilseeds and basic grains, towards com, beans and powdered milk. By 1989, when the guarantee price system for oilseed and basic grains was eliminated, they accounted for only 17% of total CONASUPO expenditures compared to 66% in 1984. In 1992, almost all marketing and storage expenditures were for com, beans and powdered milk. Less than 1 % of CONASUPO's expenditures are capital, as the majority are for the purchase, transport, processing and retailing of crops. V. Social Development Secretariat (SEDESOL) and National Solidarity Program (PRONASOL) 30. Background. President Salinas' fITSt official action was to create Solidarity in December 1988. The National Solidarity Program (Programa Nacional de Solidaridad, PRONASOL) grew out of Mexico's Rural Development Investment Program (PIDER), and receives its funding from the same Chapter XXVI Regional Development of the federal budget. The principal objective of PRONASOL is to fight rural and urban poverty through a series of programs and funds that focus on three areas: (i) improving social welfare (health, housing, education and basic services); (ii) developing productive skills in agriculture, fishing, livestock, mining, micro-business and forestry, and; (iii) expanding basic regional infrastructure (roads, electrification, irrigation, drinking water, drainage, etc). Solidarity is a matching grant program that transfers resources to states and municipalities. States, and in some cases municipalities and users, provide matching contributions. 31. Solidarity has been coordinated by the Secretariat for Social Development (SEDESOL) and its state delegations since the Secretariat's creation in May 1992. Within SEDESOL, the  ANNEX 3 Appendix 1 Page 10 of 14 program is coordinated by the Under-Secretary for Regional Development, who is appointed by the President. The subsecretariat also coordinates the preparation of the Social Development Agreements (Convenios de Desarrollo Social, CDSs) with the state COPLADEs (State Planning Commission). The CDSs establish the legal framework for the coordinated federal-state investment program, including those activities fmanced through Solidarity (Chapter XXVI) and other joint federallstate investment programs. 32. With the exception of the four funds managed at the municipal level, specific project proposals are submitted by community representatives or municipal officials to the state COPLADE and SEDBSOL delegation for review within the guidelines established by Solidarity. Once resources for the projects have been approved, a Solidarity Committee is established and the relevant public institution (e.g., State Water and Sanitation Agency for drinking water investments, CFB for electricity projects, etc.) complete the technical design work and either contract out or carry-out by force account the agreed investment. The local Solidarity committees are responsible for organizing the communities' contribution, which may be in the form of labor, land, cash, or in-kind. In contrast, in the Municipal Funds program, resources are managed at the local level and are transferred directly to the municipal Solidarity committees. By end of 1992, there were more than 100,000 Solidarity committees in all 31 states and the Federal District. 33. Data sources. Given Solidarity's high level political support and its popular appeal, resource allocations to the program have more than doubled in real terms since its inception in 1989. However, it is difficult to derive exact expenditure figures on the program. While SHCP maintains a detailed breakdown on Solidarity expenditures in the cuenta publica, the overall figures are about 15 % lower than information provided by Solidarity. Furthermore, a different set of expenditure data for Solidarity appears in the 1992 Government Report (Informe de Gobiemo, Anexo IV). The expenditure data presented below for 1989-1992 are from Solidarity. All resources are considered as investment. Solidarity's operating costs appear in the overall budget for SEDBSOL and all participating public agencies cover their own operating costs. 34. Overall Trends. This section examines the following productive programs and funds for Solidarity that have a high incidence in rural areas: Municipal Funds; Solidarity Funds for Production; Solidarity Funds for Businesses; electrification; drinking water and drainage; rural roads and highways; indigenous communities; and regional development (Annex 3, Appendix 2, Tables 24 and 25.) For 1984-1988, expenditure data from the basic infrastructure and production categories of the Regional Development Program (Chapter XXVI) were included. On average, slightly more than half of total Solidarity expenditures support productive activities in rural areas, compared to an estimated 30 % in the mid-1980s under the Regional Development Program. Rural development expenditures under Chapter XXVI have increased significantly with Solidarity: average annual resource flows with Solidarity have been 180% higher compared to the Regional Development Program. Since 1989, expenditures have risen rapidly, and  ANNEX 3 Appendix 1 Page 11 of 14 allocations for the above-mentioned Solidarity programs were 195 % higher in 1992 compared to 1989. In contrast, resources allocated to the production and basic infrastructure programs of Chapter XXVI during the mid-1980s declined steadily dropping to 43 % of their 1984 level in 1988. 35. Resource flows to productive programs in rural areas under Category XXVI were less than 5 % of total rural development expenditures in the mid-1980s. With the creation of Solidarity they have risen to almost 20% by 1992. In reality, expenditure flows to Solidarity programs are much higher, as the state, and in come cases, municipal governments participate fmancially. In most cases, program beneficiaries also contribute fmancial resources and/or labor, materials, land, etc. Finally, federal secretariats also participate in the Solidarity special programs for irrigation, sewage, drinking water, drainage (SARHICNA), electricity (CFE), rural telephones and roads (SCT and SARR). A brief description of each of the Solidarity programs is provided below. 36. Drinking Water Program. In 1983, with the reform of Article 115 of the Constitution, municipalities were given responsibility for public infrastructure and services, such as drinking water and sewage. However, the decentralization of the administration of water and sewage services from the federal government to state and municipal governments without a complementary increase in financial support caused serious deterioration in the facilities, particularly in small rural communities and the poorer sections of urban areas. 7 This Solidarity program supports the construction, extension, and repair of drinking water and sewage systems in the rural sector and poor neighborhoods. CNA, BANOBRAS, Solidarity, state and municipal governments as well as users support the program: in 1991, 36% of the program's resources were from federal sources, 28 % from state contributions, 35 % from national and international credit funds, and 1 % from direct investment by the agencies in charge of water and sewage. Between 1989 and 1992, 7,284 rural water and sewage systems were supported by the program. The share of Solidarity rural resources allocated to drinking water and drainage averages 15 %. Resources for the program have increased by 240% since 1989. 37. The Rural and Urban Electrification Program is supported by Solidarity, CFE and state governments. It provides electricity to popular urban neighborhoods and rural areas. The program consumes less than 10% of Solidarity's rural budget. Solidarity's contributions to the program grew by 85% since 1989. In contrast, CFE support to rural electrification represents less than 1 % of CFE's budget and appears to be declining (Annex 3, Appendix 2, Table 26.) Between 1989 and 1992, 9,655 rural developments received support under this program. 38. Solidarity Production Funds (Fondos de Solidaridad para la Produccion or Credito a la Palabra) were created in 1990 to assist farmers not covered by BANRURAL, estimated at 7 Solidarity, Solidarity in National Development: New Relations Between Society and Government, SEDESOL, (1993) at 75. [Hereinafter Solidarity].  ANNEX 3 Appendix 1 Page 12 of 14 approximately 269,000 producers with 1.2 million hectares. s The main purpose of the Funds is to support the cultivation of basic products (mainly com and beans) in areas without irrigation and/or low productivity. The maximum surface for which a producer can receive aid is three hectares and in 1990 the average loan size was less than US$300. Producers who have not repaid their interest-free loan to the municipal treasury are not eligible for further assistance. Municipalities operate the funds and may use the resources for further loans or other community works. The State SEDESOL and SARB delegations as well as the COPLADE coordinators analyze the producer's profit level to establish the amount of support the producer receives. In 1990, 78 % of the Funds resources were from Solidarity and the remaining 22 % was from the state governments. During 1991, 40 % of the funds extended during the Spring-Summer 1990 cycles was recuperated. 9 A slightly higher amount, 55% was recuperated between 1990-92 in the states of Chiapas, Guerrero, Hidalgo, and Oaxaca. 10 On average, the Production Funds account for 15% of Solidarity's rural program. Between 1989 and 1992, resource allocations to the program grew by 14 %. 39. The National Fund in Support of Solidarity Businesses (Fondo Nacional de Apoyo Para las Empresas de Solidaridad) was created in November 1991. It aims to promote economic activity in the rural sector and low income urban areas by supporting viable productive projects which are "fmancially sound and make social sense." 11 Low income families unable to borrow from commercial banks or DFIs receive priority. The fund mainly supports agricultural, aquaculture, forestry, agro-industrial, mining, micro industries and other activities. It supports the creation and operation of local and regional fmancing and guarantee funds that are also supported by the state government and producers associations. After the initial contribution, NAFIN or BANRURAL open a line of credit for ten times the initial amount. Each regional trust fund is managed by a technical committee which authorizes the provisions of resources either in the fonn of credit or as temporary risk capital to support viable projects. 40. In 1992, the total fund resources were N$370 million (US$123 million), less than 1 % of which came from the state governments and the remainder were federal investments. Agriculture, agro-industry, aquaculture, forestry, livestock, commercialization and fishing accounted 94 % of the businesses, the remaining 6 % are for micro-enterprises and mining. 12 In 1992, the Funds accounted for 9 % of the total Solidarity rural budget. g Solidarity at 149. 9 Solidarity at 152. JO David Gow, "Small Fanner Production: The Achilles Heel of Decentralization in Mexico," World Bank (1993) at 32. II Solidarity at 130. 12 Solidarity at 130.  ANNEX 3 Appendix 1 Page 13 of 14 41. Solidarity participates in the construction, renovation and conservation of highways and rural roads to assist integrate the most scattered and low income communities and promote regional development. The community presents an investment proposal for evaluation to the state Coplade and if the project is approved it participates with labor, local materials or fmancial contributions. In 1989, the states provided 36% of the total Solidarity costs. SARHICNA and in particular, SCT also invest in rural roads; in 1992, SCT's programs for rural and bridge roads was 35% larger than Solidarity'S federal expenditures on rural roads (Annex 3, Appendix 2, Tables 27 and 28.) SARH investments were less than 1 % of Solidarity's costs. 42. Solidarity Municipal Funds (Fondos Municipales de Solidaridad). The Municipal Funds were created in 1990 to transfer additional resources to the poorest municipalities in the country and to strengthen their fmancial and administrative capacities. The funds support three types of projects: social welfare (sewage, drinking water, schools, parks, streets, etc.); basic goods to expand rural production (drilling wells for irrigation, construction and repair of canals, livestock units, ponds, drainage, etc.); and productive infrastructure (construction and renovation of roads, bridges, supply centers, electrification, etc). The communities, in coordination with the municipal authorities, decide what to fund and how the project should be organized and implemented. In 1992, the Municipal Funds received $N 541 million (US$180 million), which represented 14% of the Solidarity rural budget and amounted to an increased of 48% since 1989. 43. The Regional Solidarity Fund (Fondos Regionales de Solidaridad) was created in March 1990. It supports the creation of regional funds managed by indigenous groups, which provide resources for productive projects. Priority is given to viable projects with short reimbursement periods so as not to decapita1ize the Fund. The interest rate is set by the managing committee of the regional fund and is generally lower than that set by commercial banks. In general, the applicant must contribute a minimum of 25 % of the cost of the project. The contribution can be in labor, land, local materials and only minimally in monetary resources. Loans are for agriCUlture, livestock, marketing, the acquisition of productive infrastructure, fish production, etc. In 1992, the regional funds received N$158 million (US$53 million) from Solidarity or 4% of the Solidarity rural budget. Resource flows to the program decreased by 17 % between 1989 and 1992. 44. Regional Development. Solidarity supports regional development programs since 1989, and as of October 1992, there were 14 programs in 11 states. The programs seek to: improve the qUality of urban services; strengthen municipal structures; support decentralization and economic activity; and promote the rationale use of the natural and productive resource base. This study assumes that approximately half of the programs expenditures benefit rural areas. Resource flows to this program were N$281 million (US$94 million) in 1992, which was a 31 % decrease over 1989. This decline has caused its share of Solidarity's rural budget to decline from 31 % to 7%.  ANNEX 3 Appendix 1 Page 14 of 14 VI. FERTIMEX 45. Up until its privatization in 1992, FERTIMEX maintained a monopoly on the production, marketing, importing and distribution of fertilizers in Mexico. To promote fertilizer use in Mexico, FERTIMEX sold fertilizer at a unifonn subsidized price (US$45Iton or about half of import parity in 1986). High transportation costs and a large number of storage and marketing outlets coupled with the high per unit subsidy required GOM fiscal transfers to support FERTIMEX, which were N$2.5 billion (US$816 million) in 1986 or 7% of total rural development expenditures. Starting in 1986 the GOM began implementing a plan to cut costs and improve efficiency, although real refonns did not take place in until 1989, when changes in pricing occurred and FERTIMEX withdrew from retail distribution. Transfers to FERTIMEX decreased substantially beginning in 1986 and became almost zero (1 % of total rural development expenditures) with the sale of the company in 1992.13 Vll. National Indigenous Institute (INI) 46. The National Indigenous Institute provides support to indigenous communities for education, infrastructure, productive activities. It is assumed that half of INI's budget directly supports rural productive activities. Since 1986, INI's budget has risen gradually and by 1992 was almost 80% higher than its 1986 level. INI's expenditures in productive rural activities represent less than 1 % of he total rural development budget. vm. National Institute for Statistics, Geography and Informatics (lNEGI) 47. Following the refonn of Article 27 of the Constitution in 1992, the Government initiated an ejido and communal lands titling program (chapter 7). Under the national program to certify ejido rights (pROCEDE) , INEGI is responsible for the surveying and production of ejido and communal lands maps. In 1992, the total budget allocated to !NEGI was N$322 million (US$107 million). m:\mx\asm\green\anx3appl.txt August 24, 1994 13 Information and data presented in this paragraph are from: World Bank, Draft Project Completion Report, • Fertilizer Sector Adjustment Loan Project (Loan 2919-ME), November 1993.  ANNEX 3. ;ABLE 1 RURAL DEVELOPMENT EXPENDITURES BY INSTITUTION \8 NSMIWONS 1992 CONSTANT PRICES InstitutIon 1982 1 SARH\7 16801 10269 9751 9237 7344 6550 4504 5673 6957 SRA 1083 1090 1017 835 581 566 503 405 469 480 1121 SEPESCA 1436 1116 948 870 836 524 387 276 262 359 304 SOUDARIDAD \1 n1a n1a 1323 1369 788 764 567 1328 2550 3090 3920 CONASUPO\6 8516 6562 10936 9240 6793 4975 3815 4982 11052 3472 3570 DFIs \4 n1a 12443 9385 11052 14671 10171 9592 7317 15597 2543 2393 SCT (rural roads & telephon 37 3393 3675 664 925 975 679 1323 1564 2098 1325 CFE (rural electricity) 341 213 238 373 129 157 253 258 82 n.d. n.d. INI\s 118 75 88 92 52 60 56 56 66 86 93 INEGI (LAND TITUNG)\2 0 0 0 0 0 0 0 0 0 0 322 SEMIP (FERTIMEX)\3 n1a n1a n1a n1a 2524 2170 1346 1088 817 672 216 TOTAL 28331 35161 37360 33733 34644 26913 21701 22725 37470 18472 20220 FEDERAL BUDGET 433007 354413 367887 360685 358177 391960 336667 314567 298763 266998 254848 share I)f 7% 10% 10% 9% 10% 7% 6% 7% 13% 7% 8% AG.GDP 75509 72349 79572 85771 79844 77606 66102 70460 77669 77033 87295 share of 38% 49% 47% 39% 43% 35% 33% 32% 48% 24% 23% --- \ISolidarity expenditures for 1984 -1988 include RAMO 26 expenditures for production and basic infrastructure under the Regional Development Program. Expenditures for social programs are excluded. All operating costs for Solidarity programs are excluded. All data from Solidarity/SEDESOL. The following Solidarity programs are included for 1989-1992: Drinking water and drainage (75% of total costs); electrification (75% of total costs); Fund for Solidarity Businesses; Solidarity Production Funds; Regional Development Program(half of total expenditures); rural roads and highways; Municipal Funds; and indigenous communities. \2 INEGI data for land titling are from SHCP. \3Source: World Bank, "Project Completion Report Mexico Fertilizer Sector Adjustment Loan Project (Loan 2919-ME), • DRAFI', 1993, p.52. \4 See Chapter 5 for a breakdown of DFI costs \5 One half of INI's total budget is assumed to be for productive purposes in rural areas. \6. Sources: Cuenta Publica, SHCP, except fiscal transfers for 1986 are from World Bank, "Mexico Agricultural Sector Report, .. No. 7609 All years, except 1991-92, include transfers for current and investment costs as well as amortization of the debt. 1991-92 excludes costs for debt amortization. \7 Only 30% of the drinking water and water management program expenditures are included, remainder are assumed to occur in urban areas. \8 All data is from the Cue!lta Publica with the exception of SoHc:1.arity, INEGI, DFIs, and SEMIP. ~» ~"C 2: -"C 2: (1)at!I!J . . . . a.>< ;:4' ~ m:\mx\asm\data\rd1.xls ~  RURAL DEVELOPMENT EXPENDITURES BY INSTITUTIONS IN PERCENTAGES Instituci6n 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1 SARH 59% 29% 26% 27% 21% 24% 21% 25% 13% 31% 34% SRA 4% 3% 3% 2% 2% 2% 2% 2% 1% 3% 6% SEPESCA 5% 3% 3% 3% 2% 2% 2% 1% 1% 2% 2%1 N/A N/A 4% 4% 2% 3% 3% 6% 7% 17% SOUDARIDAD ("J CONASUPO 30% 19% 29% 27% 20% 18% 18% 22% 29% 19% ~ DFI N/A 35% 25% 33% 42% 38% 44% 32% 42% 14% 12% SCT 0% 10% 10% 2% 3% 4% 3% 6% 4% 11% 7%1 CFE 1% 1% 1% 1% 0% 1% 1% 1% 0% N/A N/AI INI 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% INEGI (LAND TITIlNG) 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 2%1 SEMIP (FERTIMEX) N/A N/A N/A N/A 7% 8% 6% 5% 2% 4% 1%1 TOTAL 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100%1 Sources: See Text Table 7.1 ~» ~ "CI Z - ('I) "CI ('I) 2!: ~ N&.~ r;:;' C.H m:\mx\asm\data\rd1.xls N  RURAL DEVELOPMENT EXPENDITURES BY PROGRAM N$MILUONS 1m CONSTANT PRICES 1982 1983 1984 1986 19B6 1987 1988 1989 1990 1991 1992i Administration and Policy Analysis 6719 4384 4298 2727 2631 2237 1760 1601 1468 1809 3251! Market Interventions 8,834 19,304 20,448 20,479 21,560 15,247 13,488 13,517 28,516 7,663 7,438 Credit \3 118 12,618 9,473 11,144 14,723 10,231 9,648 7,438 16,414 3.321 3.651 FERTIMEX nla nla nfa nfa 2,524 2,170 1,346 1,088 817 672 216 Marketing \4 8,716 6,786 10.976 9,335 6,837 6,016 3,840 4,991 11,285 3,670 3,570 1 Public Goods 9.145 9.313 10,568 7,944 6,017 6,118 3.619 5,124 5,962 7.419 7,045 Research and Technologv 584 385 384 609 419 412 312 372 350 348 365 Infrastructure\ 1 936 3,849 5,468 2,719 2.069 2.084 1,603 2.901 3,277 4.399 3,673 Water Development \2 5,929 3,634 3,372 4,296 3.327 2,362 1,447 1,573 2,050 2,380 2,700 Education 377 251 287 201 164 229 235 260 264 269 273 Extension 1,245 1,252 991 220 39 31 22 18 21 23 34 Plant and Animal Health 74 42 55 0 0 0 0 0 0 0 Oi Other Agricultural Activities 483 . 300 332 1,086 839 1,030 799 1,984 879 944 936 Land TenurefTitling 724 473 596 480 381 430 404 298 328 331 1,238 Other 2,427 1,360 1,130 1,016 791 681 296 301 317 306 313i TOTAL EXPENDITURES 28,331 3S,161 37,361 33,732 32,120 24,743 20.354 22,726 37,470 18,472 20,2201 \ 1 Includes agriculture. livestock, forestry. fisheries infrastructure as well as electrification, rural roads, and half of Solidaritv Regional Development expenditures, !the RAMO XXVI basic infrastructure and production programs (1984·1988' and municipal funds.1 \2 Includes all SARH water development costs, except for drinking weter and water management cost. where only 30% of expenditures are included. Also includes 75% of Solidaritv's drainage and drinking water program costs. \3 Includes DFI costs plus Solidaritv Production Funds, Indigenous Communities, and Solidaritv Fund fOI Busln•••es. \4 Includes marketing c08tS for SARH, SEPESCA and aU of CONASUPO's costs, Sources: See Text Table 7.1 ~» ~'CI Z _'CI ~ til Z _ w::S l-J e:>< ~ (".l ~ fwi: '.l'viX\ASM\lJA"fA\RDPROG .XLC  RURAL DEVELOPMENT EXPENDITURES BY PROGRAM IN PERCENTAGES 1982 1983 1984 1986 1986 1987 1988 1989 1990 1991 1992 Administration and Policy Analysis 24% 12% 12% 8% 8% 9% 9% 1% 4% 10% 16% Market Interventions 31% 55% 55% 61% 67% 62% 66% 59% 76% 41% 37% Credit nla 36% 25% 33% 46% 41% 47% 33% 44% 18% 18% FERTIMEX nla nla nla nla 8% 9% 7% 5% 2% 4% 1% Marketing 31% 19% 29% 28% 21% 20% 19% 22% 30% "20% 18% Public Goods 32% 26% 28% 24% 19% 21% 18% 23% 16% 40% 35% Research and Technology 2% 1% 1% 2% 1% 2% 2% 2% 1% 2% 2% Infrastructure \ 1 3% 11% 15% 8% 6% 8% 8% 13% 9% 24% 18% Water Development \2 21% 10% 9% 13% 10% 10% 7% 7% 5% 13% 13% Education 1% 1% 1% 1% 1% 1% 1% 1% 1% 1% 1% Extension 4% 4% 3% 1% 0% 0% 0% 0% 0% 0% 0% Plant and Animal Health 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% Other Agricultural Activities 2% 1% 1% 3% 3% 4% 4% 9% 2% 5% 5% land TenurelTitling 3% 1% 2% 1% 1% 2% 2% 1% 1% 2% 6% Other 9% 4% 3% 3% 2% 3% 1% 1% 1% 2% 2% TOTAL EXPENDITURES 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% \ 1 Includes agriculture, livestock, forestry, fisheries infrlSstructure ISS well as alectrlfication. rural roads, ISnd half of SolidlSrity Regiona. Development expenditures, the RAMO XXVI basic infrastructure and production programs, (1984-1988) and municipal funds. \2 Includes all SARH water development costs, except for drinking water and water management C08ts whare only 30% of expenditures are included. Also includes 76% of Solidarity's drainage and drinking water program coata. \3 Includes OFI costs plus Soliderity Production Funda, Indigenous Communltlel, and Solidarity Fund for Busines8es. \4 Includes merketing COltl for SARH, SEPESCA end ell of CONASUPO'I costl. ~» ~"I:I Z _"1:1 Z ~fDt:'j J:>.6.>< j;(' ~ N M:\MX\ASM\OATA\RCPROG.XLS  RURAL DEVELOPMENT EXPENDITURES BY CATEGORY· NS MUlions 1991 Constant Prices 1985 1986 1987 1988 1989 1990 1991 1991 RECURRENT SARB 3,424 2,919 2,705 2,239 2,758 n/a 2,423 3,639 8RA 680 529 555 500 379 n/a 427 936 SEPESCA 714 515 361 288 230 n/a 193 219 SOLIDARIDAD 11 0 0 0 0 0 0 0 0 CONASUPO/5 9,240 6,793 4,975 3,815 4,982 11,052 3,472 3,570 [DFIs 21 11,052 14,671 10,171 9,592 7,317 15,597 2,543 2,393 SCT (roads, rural telephones) I 0 0 1 181 nla 165 0 CFE (rural electricity) 0 0 0 0 0 0 0 0 0041 46 26 30 28 28 33 43 47 INEGl (land titling) 31 0 0 0 0 0 0 0 64 SEMIP (Fertimex) 51 n/s 2,524 2,170 1,346 1,088 817 672 216 Total Recurrent 25,157 27,977 20,967 17,809 16,963 27,499 9,938 11.084 Sbare Recurrent Costs 75% 81% 78% 82% 75% 91% 54% 55%1 CAPITAL SARB 5,824 4,425 3,845 2,265 2,934 n/a 3,250 3,3181 SRA 152 55 12 3 26 n/a 54 184 SEPESCA 152 326 164 98 46 n/a 166 85 SOLIDARIDAD 11 1,369 788 764 567 1,328 2,550 3,090 3,920 CONASUPO 51 0. 0 0 0 0 0 0 0 • I OFls 21 {I 0 0 0 0 0 0 0 SCT (roads, rural telephones) 663 925 915 678 1,141 n/a 1,'0.5 1, 312 1 CFE (rural electricity) 373 129 157 253 258 82 n/a n/al INI41 46 26 30 28 28 33 43 47 INEGI (land titling) 31 0 0 0 0 0 0 0 258 SEMIP (Fertimex) 51 0 0 0 0 0 0 0 0 Total Capital 8,519 6,614 5,941 3,892 5,161 2,665 8,508 9,124 Share Capital Costs -'-- ~% 19% 22% 18% 2~% 9% 46% 45% - • See Table 1 for source of data. ~» l! All SOUDARlDAD expenditures In: capital ~'C ~ -'C ...", ('1)('1)..0:. 21 All DFb cxpc:nditllcClI In: reculIClll. VlI:Jt:-'J ~~ 31 Breakdownbued"" 1993 'xukdo~'11. 41 Assvmca III oftolallNl r.:-, dr-elopme..t cosIJ In: 1U'Un'cnt. 51 All cosIJ In: dasaificd as Je(.Ulr.'I'Il. .... m: \mx\asm\data\ruraldev .XIS  RURAL DEVELOPMENT EXPENDITURES BY CATEGORY· IN PERCENTAGES Institution 1985 1986 1987 1988 1989 1990 1991 1992 RECURRENT SARH 14% 100/. 13% 13% 16% N/A 24% 330/. 8RA 3% 2% 3% 30/. 2% N/A 4% 8% SEPESCA 3% 2% 2% 20/. 1% N/A 2% 2% SOLIDARIDAD 11 0% 0% 0% 0% 0% 0% 0% 0% CONASUPOI5 37% 24% 24% 21% 294'10 40% 35% 32% DFIs21 44% 52% 49% 54% 43% 57% 26% 22% scr (roads. rural telephones) 0% 0% 0% . 0% 1% N/A 2% 0% CFE (rural electricity) 0% 00/. 0% 0% 0% 0% 0% 0% INI41 0% 0010 0% 0"'. 0% 0% 0% 0% INEOI (land titling) 31 0% eel. 0% 0% 0% 0% 00/. 1% SEMIP (Fertimex) 51 N/A 9% 10% 8% 6% 3% 7% 2% Total Recurrent 100% lOG-I. 100010 100% 100-10 100% 100% 100% CAPITAL SARH 68% 66% 65% 58% 51% N/A 38% 36% SRA 2% w. 0% 0% 0% N/A 1% 2% SEPESCA 2% So;. 3% 3% 1% N/A 2% 1% SOLIDARIDAD 11 16% 12% 13% 15% 23% 96% 36% 43% CONASUPO 51 0% 0% 0% 0% 0% 0% 0% 0% DFIs 21 0% 0% 0% 0% 0% 0% 0% 0% scr (roads. I111lJ1 telephones) 8% 14% 160/. 17% 20% N/A 22% 14% CFE (1UIlll electricity) 4% 2% 3% 7% 4% 3% N/A N/A INI41 1% 0% 1% 1% 0% 1% 1% 1% 1NE0I (land titling) 31 0% 0% 0% 0% 0% 0% 0% 3% SEMIP (Fertimex) 51 0% 0010 0% 0% 0010 0% 0% 0% Total Capital 100-10 100% 100% 100"'. 100-10 100% 100% 100% ~» ~"O Z ·See Table 1 for source of data. -1 z II All SOUDARIDAD expcadiIweIare apitaI G = {!!j O\CI.~ 1J All DFIa cxpalllituna are recw:rcnl. ;:;' CM 31 Blcakdown buecI 011 1993 breakdown. N 41 A.ssumfs 112 oftata! INI rural development COlts are nc:umnL ~I AII.-neI ...... dlW:ifietf •• , ..... ,""""  SARB EXPENDITURE BY PROGRAM AND CATEGORY NS THOUSANDS 1992 CONSTANT PRICES ]985 19" 1987 1988 Program Current Capital Total Current Capital Total Curftllt Capital Total Cu.rre.nt Capital Total Administration 1,352.982 623.817 1.976.799 1.336.996 384.376 1.721.371 1.295.789 463.376 1.759.165 1.102,613 251.394 1,354,0071 Policy Analysis 16,2'0 134,183 ISO,03 36,399 12,318 48,718 26,702 13,819 40,'21 40,082 3,879 43,961 ~ RenArch &. Technology 172,120 304.869 476,989 38,025 351.377 389.401 45.844 337,809 383.653 44,633 244,237 288,869 InfrasINcl.Ure 3.724 98AS7 102.181 2,577 91,266 93,843 27,136 61,458 88,593 9,569 30,683 40,252 AgJicullUrc 0 0 0 0 0 0 0 0 0 0 0 0 Uvmock 3,146 83.003 86,1'0 1,978 79,354 81,332 26,817 S6,'SSl 83,376 9,'S6 28,012 37,'681 Forestry S78 10,3" 10.973 599 9,902 10,SOI 318 4,691 ',009 13 2.671 2,684! Rural Roads 0 S.OS8 5,058 0 2.010 2,010 0 208 208 0 0 0 Water 446.737 3,848,658 4,295,395 342.051 2,984.747 3,326.798 35,387 2,326,645 2,362,031 35.170 1,412.230 1,447.400 Water Management \1 0 1,207 1.207 0 6 6 0 60 60 1,199 74,8SS 76.084 Drinlcing Water \1 317 281,187 281,'03 138 19'.851 19$.989 548 148.463 149.011 0 0 0 Drainage 0 14,179 14,179 0 12.'" 12.575 0 19,5'7 19.557 0 0 0 RainfedAreas 427,322 342.'27 769,849 277,466 263.661 541,126 17,569 287,292 304.862 11,525 59,690 71,215 lITigated Areas 19,098 3,209,558 3,228,656 64,448 2,512.653 2.577,101 17,269 1,871,273 1,888,542 22.447 1,277.655 1,300,102 Education 175.724 2S,371 201.095 146,560 17,237 163,798 204,751 24,576 229,327 219,311 15.408 234,719 Markeling 16,409 26,128 42,537 19,077 9,870 28,946 12.002 8,612 20,613 9,748 12,236 21,984 Services 826,127 260,858 1,086,986 677,S92 161,702 839,294 751.172 280,120 1,031.291 623,354 175,917 799,270 Promotion !:LS.829 259,763 1,085,592 677,592 161,'31 839,123 751,172 279,110 1,030,281 623,354 175,677 799,031 Extension 299 1.095 1,394 0 171 171 0 1,010 1,010 0 239 239 Plant and Animal Health 0 0 0 0 0 0 0 0 0 0 0 ° Other Prognum 403,525 501,424 904.949 319,671 412,242 131,913 306.273 328.9S3 635,226 154.409 IIB,940 273,349 TOTAL 3,413,599 _ ~8%3.765 ',137~63 • _2,918, N !I _ 4,~5,134 7,3.....~ 2;!05.054 3,845,367 6,550,421 2,238,18L 1,164,913~ 4,50l,812 ao.-: Cuenle PUIIIaa 1919-1992 NlA HoI:..",..... .. the CuonIa PubIc:a 1/ A.8umes 30% ot the costs of thi8 subprogram are In rural ar8a8. 1-1» ~"CI Z _"CI (11 «'II Z ....,J=~ I»c,YI ;:r ~ II-' M:\MX\ASM\DATA\SARHPRG 3.XLS  SARH EXPENDITURE BY PROGRAM AND CATEGORY NSTHOUSANDS 1992 CONSTANT PRICES 1989 1990 1991 1992 Procnun CUrrent Capital Total Curent Capital Total c.m-t Capllal Total Current Capital Total Adminis1ration 942,146 285,458 1,227,604 N/A NlA 1,148,600 1,010,501 330,177 1,340,677 1,367,286 346,052 1,713,338 Policy Analysis 33,988 15,438 49,426 N/A N/A 58,700 73,994 4.5,171 119,164 1,148,4.52 52,600 1,201,052 Research & Technology 48,8S5 299,76.5 348,621 NlA NlA 325,100 45,360 275,137 320,497 48,054 281,045 329,099 Infrastructure 31,398 162,865 194,264 N/A NlA 124,700 10,972 265,681 276,654 11,292 223,162 234,454 Agriculture 11,806 124,310 136,115 N/A NlA 64,100 140 237,038 237,178 0 205,211 205,211 Livestock 17,920 32,834 50,754 N/A N/A 48,200 10,833 17,407 28,240 10,518 10,832 21,350 Forestry 0 1,317 1,317 N/A N/A 1,600 0 260 260 774 1,966 2,740 Rural Roa<b 1,673 4,404 6,077 N/A NlA 10,800 0 10,976 10,976 0 5,1S3 5,153 Water 38,485 1,317,402 1,355,887 N/A N/A 1,669,340 220,008 1,685,660 1,905,668 221,965 1,740,037 1,962,002 Water Management \1 1,981 49,955 51,936 N/A NlA 67,170 0 95,489 95,489 0 89,382 89,382 Drinking Water \I 0 86,230 86,230 N/A NlA 137,370 3,879 207,967 211,845 5,338 171,846 177,184 Drainage 0 11,432 11,432 N/A NlA 15,600 0 . 37,698 37,698 0 32,583 32,583 Rainfed Areas 1,459 62,854 64,313 N/A NlA 112,100 0 125,045 125,045 0 124,808 124,808 Irrigated Areas 35,045 1,106,930 1,141,975 N/A NlA 1,337,100 216,129 1,219,462 1,435,591 216,627 1,321,418 1,538,045 Education 237,259 22,256 259,515 N/A N/A 264,200 234,559 33,941 268,500 241,747 31,477 273,224 Marketing 4,756 3,4:46 8,202 N/A N/A 232,800 . 192,808 5,701 198,509 0 652 652 Services 1,294,307 689,673 1,983,980 N/A NlA 878,800 582,953 360,858 943,81\ 560,130 375,378 935,508 Promotion 1,294,307 689.673 1,983.980 N/A N/A 878.800 582,953 360,858 943,811 560,130 375,378 935,508 Extension 0 0 0 NlA N/A 0 0 0 0 0 0 0 Plant and Animal Health 0 0 0 N/A N/A 0 0 0 0 0 0 0 Other Programs 126,831 137,704 264,535 N/A N/A 308.S00 51.946 247,269 299,215 40.125 267.636 307.761 TOTAL 2,758.027 2,934,007 5,'92,033 5,010,740 2,423,101 3,249,595 5,672,6" 3,639,051 3,318,039 6.957,090 8cuea: cu.nt. PI'iflIca 1888-1192 NlA Hal IIVIIiIaIIIe In tI. C40MIIa I'uIIIea 1/ A.,um•• 30% of the co.t. of thi•• ubproor8l11 .r. In rural .r•••• t-i» ~"C Z -"= ~ Z I'll t'!I . M:\MX\ASM\DATA\SARHPRG3.XLS . O!6.X .... W I ..\,  SARH EXPENDITURES BY PROGRAM N$ MILLIONS 1992 CONSTANT PRICES Category 1982 1983 1984 1986 1988 1987 1988 1989 1990 1991 1992 Administration .5,382 3,287 3,271 1,977 1,721 1,159 1,3S4 1,228 1,149 1,341 1,713 Policy Analysis 88 70 10 ISO 49 41 44 49 59 119 1.201 Re6earch & Technology 498 3S6 3S7 477 389 384 289 349 325 320 329 Infraestructure 114 19 91 102 94 89 40 194 125 277 234 - Agriculture 0 0 0 0 0 0 0 136 64 237 205 - Livestock 151 62 88 86 81 83 38 51 48 28 21 • Forestry 15 13 4 II II 5 3 1 2 0 3 • Rural Roads 9 5 5 5 2 0 0 6 11 11 S Water 5,929 3,534 3,372 4,295 3,321 2,362 1,441 1,356 1,670 1,906 1,962 • Water Management· 1 1 0 1 0 0 76 52 67 9S 89, i - Drinking water· 622 472 392 282 196 149 0 86 137 212, l77i - Drainage 2S 26 36 14 13 20 0 11 16 38 33! • Rainfed Areas Develop. 0 0 0 770 541 305 71 64 112 125 12S1 • Irrigated Areal 5,277 3,035 2.944 3,229 2,577 1,889 1,300 1.142 1,338 1,436 1,538 Education 1,881 1.37S 1,273 1,130 868 1,052 821 1,992 1,112 1,142 936 Marketing 192 223 36 43 29 21 22 8 233 199 I Services 1.615 1,110 1,181 1,081 839 1,031 799 1,984 879 944 9361 I • Plant and Animal Health 74 42 S5 0 0 0 0 0 0 0 0 • Extansion 1,057 768 193 I 0 I 0 0 0 0 0 - Promotion 483 300 332 1,086 839 1,030 799 1,984 879 944 936 Other programs 2,548 1,358 1,140 90S 732 63S 273 265 309 299 308 Tobll 11,306 11,3'3 10,736 10,1" 1,04' 7,373 5,090 7.415 5,85' 6,547 7.620 n.d.: Hac .,......In.". CuenII. ~ . . . . . . 30~ of . . . . ."'.... _ _ ...,......... ~» &oura.: DINcdcln 0 . . . . . Pw---'6n Y ~A~.......... y.u.... ~-= Ciil z Z oo::spj ~~ N M:\MX\ASM\DA TA \SARI~!"'~'l1 ..l/'lS  SARH Expenditure Composition by Program In Percentages Catagory 82 83 84 86 88 87 88 89 90 91 92 I Adminbotration 29% 29% 30% 19% 21 % 24% 27% 11% 20% 20% 22%1 NorlTUltivity 0% 1% 0% 1% 1% 1% ,% ,% 1% 2% 16% , R•• earch " T.chnology 3% 3% 3% 6% 6% 6% 6% 6% 6% 5% 4%1 I InfrllOlltructure 1% 1% 1% 1% 1% 1% 1% 3% 2% 4% 3%1 • Agriculture 2% 1% 4% 3%] • Uveetock 1% 1% 1% 1% 1% 1% 1% 1% 1% 0% 0% • Foreetry 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% • Rural Roade 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% Water 32% 31% 31 % 42% 41% 32% 28% 18% 29% 29% 26% • Water Management 0% 0% 0% 0% 0% 0% 1% 1% 1% 1% 1% • Drinkable Wat er 3% 4% 4% 3% 2% 2% 0% 1% 2% 3% 2% • Draina". 0% 0% 0% 0% 0% 0% 0% 0% 0% 1% 0% • Rainfed Areas Develop, 0% 0% 0% 8% 7% 4% 1% 1% 2% 2% 2% • Irrigated Are.. 29% 27% 27% 32% 32% 28% 28% 15% 23% 22% 20% Education 10% 12% 12% 11 % 11 % 14% 16% 27% '9% 11% 12% Marketing 1% 2% 0% 0% 0% 0% 0% 0% 4% 3% 0% Services 9% 10% 11% 11% 10% 14% 16% 27% 15% 14% 12% • Sanity 0% 0% 1% 0% 0% 0% 0% 0% 0% 0% 0% • Extention 8% 7% 7% 0% 0% 0% 0% 0% 0% 0% 0% • Promotion 3% 3% 3'" "'" '0'" 14% 18% 27% 15% 14'" 12% Other programe 14% 12% 11 % ,% ,% ,% 5% 4% 5% 5% 4% Tat .. 10O,," 100% 100% 100% 100"" 100% 100% 100,," 100% 100% 100% 1-1» ~"= Z Source: Cuenta PubUc. ISHCP) -"= Z ~ n. t!!j \O8.~ •• ... t/.I N M:\MX\ASM\OATA\SARHPRG 1.XLS •  SARH DELEGATION EXPENDITURES (Thousands of New Pesos) CownaDtPr~e&l991 Stlte 89 90 91 92 93 Agua.caliente. 20,991.17 13.877.99 14.902.60 16079.5 13,583.38 Baja California 77.586.33 34,527.97 28.406.10 26612.7 23,331.60 Baja California Sur 26,325.74 16,460.26 16,714.93 16842.8 13,380.37 Campeche 21,819.61 13,348.44 19,637.32 16529.9 14.363.65 C,.. .. h.!il .. 32.272.12 18.718.71 21.912.55 23480.7 19,385.11 L.olima 24,410.08 1 !,444.62 12.490.48 14467.7 12,246.30 Chiapas 80,164.60 49.666.99 57.340.60 58269.8 60.381.28 Chihuahua 79,002.63 64.730.32 80.179.41 47702 37,290.06 Durango 39,822.26 32.872.45 49.253.52 31658.9 22,968.40 Region L8Qunera 48,345.64 23,875.15 22,800.92 24514.2 19,148.49 Guanl.jl'·'~'" 46,822.30 29,789.13 32,805.34 33061.9 28,892.51 Guerrel. 60,816.47 30,920.07 40,404.57 32445.1 28,660.00 HidallJo 41,104.38 ~3,OO6.37 22,462.67 23674.9 21.596.36 Jaliacu 72,198.64 42,367.56 47,068.81 47189.6 39,714.79 Mexico 49,663.13 29,112.78 28,746.43 32983.4 26,840.37 Micho8ch 88,621.26 60,489.14 49.361.88 52759.3 44.987.96 Marelos 36,078.14 16,478.95 16.802.26 16070.2 14,208.49 Navarit 42,604.61 20,396.34 22,646.23 22886.4· 18.636.71 Nuavo Leon 37,157.12 21,228.13 23,313.96 23098.1 19,964.43 Oaxaca 56,639.67 30.456.14 40,298.98 34440.8 30.567.21 Puebla 44,507.31 28,028.47 28.867.50 31306.6 26,699.91 Queretaro 38,761.04 17,644.74 20,705.10 20860.2 16.887.12 Quintana Roo 26,915.29 20,892.86 18,674.86 19243.2 17,074.89 San Luis Potos( 69,696.22 26,916.38 29,926.11 30456.7 26,198.81 Sinaloa 126,390.62 47,066.a3 47.287.90 42142 36,338.17 Sonora 98,681.24 42,969.88 37,561.49 37936.7 30,800.65 Tabasco 43,667.39 29,160.19 32,387.49 30247.9 27.916.80 Tamaulipas Centro 124.131.70 . 62,694.76 51,316.66 47662.3 43,121.10 TIaxcala 22.179.63 14.507.58 18,447.89 19094 14,970.41 Veracruz 79,318.76 49,568.36 52,420.71 63046.6 49,031.23 Yucatan 37,852.03 19,961.39 20,875.96 22882.6 21,338.81 Zacatecaa 39.699.70 23.629.26 27.256.07 31065.6 26,017.99 Diatrito Federal 17.903.93 12,681.83 13,449.90 14964.9 12,325.76 TOTAL 1,739,850.65 848.167.94 1,046,606.17 996,665.10 847,757.99 ac.-8AIIH ~» I» "0 ~ s:: "0 n I'D ~ L, .... =t!I!!! Oc.>< ~. c.l N m:\mx\asm\data\sarhdell.xls  HUMAN RESOURCES EVOLUTION BV INSTITUTION 1983 1984 1986 1988 1987 1988 1989 1990 1991 1992 1993J SARH 140.262 138.276 137.618 129.838 122.342 117.201 99.301 89.632 79.633 76.301 73. 347 1 SRA 15.038 16.038 16.047 16,341 16,049 20,418 17.913 16.249 16.122 14.086 13.426,i I PESCA 6.422 6,422 6,346 6,163 6,836 6,698 6.204 6.141 6.198 6.077 _ 4.8431 Souro.:SARH.SRA. SEPESCA 1-1» I} -,:, Z --,:, Z til l!!!! ... = ('j) IyI ..... Q. ...... ;:;0 ~ ~ ..,,,,,,"1Mm1. . ., , . ' - 1  SARH TRANSFERS TO DEPENDENT INSTITunONS Nt MUION8 1992 CONSTANT PRICES 19 82 1183 1184 198 Ii '98S 1981 ENTIDAD REClJRRENT CAPffAL RECURRENT CAPn'AL RECURRENT CAPn'Al RECURRENT CAPITAL RECURRENT CAPffAL RECURRENT CAPffAl OIlOANOS DESCONCENTfIADOS INSTITUTO NACiONAl DE INVESTIGACIONES FORESTAt£S V AGROPECUARIAS 3 322 302 1 COMISION DE AGUAS DEL VAu.E DE MEXICO 919 679 984 664 I COMISION DEL LAGO DE TEXCOCO 93 60 68 68 COMISION DE AGUAS DEL VAu.E DE MEXICO V LAGO DE TEXCOCO 536 381 COMISION HACiONAl DEl AGUA \1 INSTITUTO MEXICANO DE TECNOLOGIA DEL AGUA 48 COt£GIO SUf'£R/OR AGROP, DEl ESTADO DE GUERRERO 8 8 8 0 6 0 6 5 0 UNiVERSIDAD AUTONOMA AGRARIA 'ANTONIO HARRO' 69 58 47 16 53 12 66 68 5 APOVOS V SERVICIOS A LA COMEftCIAUZACION AGROPECUARIA SUMA 17 1,072 66 729 53 1,067 60 744 65 859 73 7421 ENTIDADES DE CONTROL OIRECTO AZUCAfI S.A. DE C.V. FORESTAL VICENTE GUERRERO 22 15 13 10 11 3 5 6 4 3 6 2 INSTITUTO MEXICANO DEl CAFE 395 92 217 231 178 67 149 50 110 5 97 3 PRODUCTOS FORESTAt£S MEXlCANOS 8 11 4 11 4 4 3 2 3 5 4 4 SUMA 423 118 234 262 193 64 157 57 111 13 107 9 ENTIDADES DE CONTROL INDIRECTO COt£GIO DE POSTGRADUADOS 108 61 74 26 68 15 68 5 62 7 76 4 FIDElCOMiSO DE RlESGO COMPARTIDO 770 4 496 178 99 139 96 131 16 192 18 FONDO GANADERO 8 49 14 8 0 8 1 6 1 5 PROOUCTOfIA NACIONAL DE BIOLOGICOS VETERlNARlOS 4 2 0 1 2 1 1 0 1 0 UNIVERSIOAD AUTONOMI>. DE CHAPlNGO 214 31 188 8 166 11 164 19 149 7 180 12 SUMA 1.102 162 740 48 400 139 373 129 344 31 448 40 ENTlDADES EN PROCESO DE DESINCORPORACION 149 347 138 40 102 36 90 30 72 11 59 127 ENTIOADES CONCLUYSION SU PROCESO DE DESINCOfIP. 422 116 290 314 2'" 378 332 614 266 405 233 186 SUMAN LAS TRANSFERENClA6 AL SECTOR PARAEST. V OROANOS DESCONCENTfIAOOS 2,113 2,405 1,4116 1.443 960 1,683 1,012 1,474 854 1,331 921 1,104 ,,~I_ OTRAS TRANSFERENC lAS 41 53 64 0 0 0 18 118 74 11 TOTAL 2.2'13 _ _ 2.~ _ 1,620 1,013 1,683 1,012 1,474 932 1.449 995 1,115 -...: c.....111 ""'_ !SHC" \1 Includft .ntlra co.u. for drInIdng water IIIId water rIlIIIlIIgefIle JIIOIII'tmII. ~» ~'O Z (i"'1 z . . =~ .NCo""" ~ .. ~ ~ C.H N M;\MX'~SMIPA TA\SARHTAN3.XLS  SARlI TRANSFERS TO DEPENDENT INSTITIJTIONS NSMW.JONS ' ••Z CONSTANT PRICES 1988 un u,. u ,. 19 91 INSTmJTlON REClIJlJI.ENT I CAPITAL RECURRJ:NT I CAPITAL REC1.lJUlENT I CAPITAL RECURRENT I CAPITAL RECUllRENT I CAPITAL ORGANOSDESCONCENTRAD05 lNS1TIVfO NAClONAL DS INVESTIOACIONES FOR.ESTALSS Y AOROPECUAIUAS 230 252 23S 24S 0 246 COMISION DE AQUAS DEL VALU! DB MEXICO 7 COMISiON DEL LAOO DE TEXCOCO COMISION DB AQUAS DEL VALLE DB MIOOCO Y LAOO DE TEXCOCO 114 COMlSlON NACIONAL DEL AQUA \1 78 1,280 57 2,~ 66 2,127 III 2.142 lNS1TIVfO MEXlCANO DB TECNOLOOlA DEL AOUA 38 0 64 54 2 68 102 COLEOIO SUPERIOR AOROP. DEL ESTADO DE OUERR.EltO S 0 S 0 6 0 1 0 7 0 UNlVERSIDAD AtITONOMA AORARlA 'ANTONIO NARRO' 78 88 87 I 94 2 90 2 APOYOS Y SERVlClOS A LA COMEllClALIZAClON AOROPBCUARIA II II 1,082 21 SUMA 8l 441 171 l,,04 150 1,384 180 3.054 U'. 3.114 ENTIDADES DE CONTROL DIRECTO AZUCAllS.A DBC.V. 102 291 199 FORESTAL VICENTE OUElUiERO 4 I 13 I <4 I 0 0 INS1TIUTO MEXlCANO DEL CAFE 80 2 16 427 12 17 S6 6 S2 31 PRODtlCTOS FORESTALES MlOOCANOS SUMA 2 87 , 1 12 804 n5 0 2 II 117 U 0 , 51 31 ENTIDADES DE CONTROL INDIRECTO COLEOIO DE POSTGRADUADOS FIDEICOMISO DE IlIESOO COMPARTIDO rONDO OANADERO 76 S9 5 II .. ,. 82 2 S4 <4 17 177 3 197 6 81 7S 2 I 317 10 81 110 7 2 291 2 PRODtlCTORA NACIONAL DE BlOLOOICOS VETERINARIOS 1 0 1 I 1 2 UNlVERSIDAD AurONOMA DB CHAPINOO 191 189 194 SUMA 33. 9 ZI 210. 311 10 70 443 20 U6 ,)54 ,....14 201 40.. 30% 8 ENTIDADES EN PIlOCESO DE DESlN.CORPOAACION 52 10 52 10 59 15 16 122 47 9 ENTIDADI'3 QVE CONCLtn'l.RON $V PROCI'3O DE DESINCORP. 119 72 143 <IS 71 25 21 41 lIS 10 SUMAN LAS 1'IlANSFEJlII:N1AS AI. SECTOR PAR..\ESTA.TAI. Y ORGANOS DESCONCENTRADOS 135 "8 1,541 2,4S4 110 2,867 688 l,S67 1,809 M77 OnmR TRANSFERS 11 22 61 )7 <44 163 69 86 TOTA.L IN 510 1,541 2,4" I'D 2,MS 7U - 3,730 1,17' 3,563 s..... "'TMaI s.utB E.. ~"I_ ~cuagru"~B~ t-i» ~'C :z \1 IDduda fIII ............ ,ar ......11 I . , .... ~_......- (i'" "g :z ...... ::I~ NQ.""" c:r •• n 11'1 ~ M;\MX\A$M\OATA\SARHTRN3.'9-S N  CNA EXPENDITIJRE BY PROGRAM (1989 - 1992) N$mOUSANDS 1992 CONSTANT PRICES Program 1989 1990 1991 19921 Administration 230,910 257,574 346,210 354,7651 Normativity 13,916 45,162 39,211 48,000 Research & Tech. 5,343 20,390 122 0 I Infraestructure 280 393 0 0 Livestock 208 273 0 0 Rural Roads 72 119 0 0 Water 1,050,664 1,792,768 2,358,032 2,308,8031 Water Management \ 1 138,816 207,734 318,299 297,940 Drinking Water \ 1 282,716 455,382 693,161 569,951 Drainage 8,024 15,134 37,698 32,583 Rainfed Areas 58,146 101,318 110,619 113.548 Irrigated Areas 562,962 1,013,201 1,198.256 1.294,781 Promotion 2,629 4,251 0 450 Other Programs 41,314 29,853 49,536 105,842 TOTAL 1,345,056 2.1S0,390 _ • • _w:a 2,793,112 2,817~ft60 AS . . . t-lj;ll-j;II- ~"C Z \ 1 Entire program costs art'! included. An estimated 70% of expenditures occur in urban areas. -"C ~ ~ ... = ('I) Source: Gerencia de Programacion y Presupueslo, CNA wQ.~ ;::;. C.H N M:\MX\ASM\DATA\CNAPRG01.XLS  SRA EXPENDITURES BY PROGRAM N$ MiUioos 1992 CoastaDt Prices Cllqory 81 13 84 IS 86 .., 88 89 90 91 92 ADMlNISTRATION 271 160 254 162 184 123 89 92 115 12:1 171 LAND TENURE 714 473 596 480 381 430 404 298 321 331 916 - LAND TITLING 618 400 484 230 18l 250 241 119 161 142 69S - LAND CONF. RES. 106 22~1 73 111 150 199 110 162 166 166 189 • NOltMAT1VITY 0 0 0 0 0 0 4 0 0 EXTENSION 140 444 173 ° 119 19 13 10 14 14 II 29 OTHER. PROGRAMS 2 0 0 I TOTAL 1136 1077 0 1023 132 0 584 0 ° 567 SOl 40S 469 3 480 7 S 1121 lb:peuditure Composition Calesory 81 13 84 15 16 87 l& 89 90 91 92 ADMlNISTRATION 24" IS" 1$" 20" 31" 22" II" 23" 27 'ill 26'i11 IS" 65" 80" 82" LANDTENURB -Ut.ND TrrLINO 64 'ill S4" "" 37" S8" 47" 58" 28" 31" 76" 48" 74" 32" 70" 3S" 69" 30" 62" - LAND CONF. RES. 9" 7" 11" 30S 34" "" 31" 32" 41 " 35'1 39" 20" - NOltMATIVITY 0$ O'ill 0" 0" 0" 0" 0" I,. O'ill O'ill 0" EXTENSION 12" 41 'ill m, 23" 3,. 2" 2" 3" 3" 4$ 3$ oTHER PR.OGRAMS O'ill O'ill Olio Olio Olio Olio 0,. I,. TOTAL 100" 100" 0" 100" 100" 100" 100" 100" 100" 100" '" 100" -- 0" 100" Sowec: CucaIa Publica (SHCP) ...;» Bog ~ ~'" -= ~~~ N M:\MX\ASM\OATA\SRAPRG02.XLS  BVDGET BY PROGRAM AND CATEGORY N$11a--. 1"2 CautaDt Prie_ 5RA 198& 1988 1887 11188 Prov ra _ Com.nu Caplt" 1..... Conlent. C.pItaI Tot.. Com.n'. Capit.. To'" Conle"t. e.plt" Tote' I Administration 161.646 767 162.302 134.766 48.868 183.623 116.112 7.117 123.228 68.281 464 88. 736 1 Lind Tenur. 476.282 4.182 480.474 317.736 3.476 381.211 426.827 4.410 430.337 401.679 2.217 403.866 1 land Tltlino 227.741 2.808 230.348 180.714 1.027 181.741 246.127 2.366 260.484 241.066 1.176 242.234 Land Coni. Rei. 246.661 1.673 260.124 197.021 2.448 189.470 177.800 2.043 178.843 180.621 1.101 161.622 Policy Analyais 0 0 0 0 0 0 0 0 0 0 0 0 Extension 41.681 147.167 188.846 16.724 2.374 19.096 12.880 0 12.980 10.226 181 10.407 Other Program. 0 (I 0 (l 0 Q 0 Q 0 0 0 0 TOTAL 878.617.84- 162,106.81 831.823.86 628.214.;./ls 64.71'.81 68l11."J109 &66.018.40 11.6~e:'1i 688.646.16 600.OS8.46 2,811.11' 602,818.37 Sown: Cu.".. ''''ro. ,SHe,. ~» ~"C 2: crl 2: ..... =1:!I!j U\Q."'" s:» •• ~ Ii'4 1'.1.1 tv WdWXlASWIDATNSIlAI'l(l',1U.I  BUDGEI' BY PROGRAM AND CATEGORY N$'l1a_da 1m Ceutaat Prie. 1988 '880 1991 1882 P",gr.ma Cortlente Capital Total Cortlenta Capital Total Conlenta Capital Total Conlente eepltal Totel Administration 88.726 2.108 81.836 124.670 89.426 36.482 124.817 142.266 28.716 170.880 Land Tenure 216.469 21.890 298.349 327.910 316.601 14.487 330.988 761.919 163.171 916.690 Lind Titling 121.667 7.202 128,759 162,130 133,486 8.422 141,906 661,373 143,619 694,982 Lend ConI. ReI. 164.902 10.796 166,698 166.780 183,016 6,066 189.081 210.646 10,162 220,696 Policy Anllly_11 0 3,893 3,893 0 0 0 0 0 0 0 Extllnsion 12,182 1.879 14,061 13,630 16,622 1.992 17.613 27,077 1.611 28,688 Other Programs 443 662 1.006 3.280 6,187 1.616 6.682 4.866 115 4.971 TOTAL 378,811.00 28.439.63 406.260.62 481,270.00 4ge,816.41 __ 480. 1()C).3g _ 63,484.90 8~8,117.00 184.412.00 1.120,629.00 10....: CUI.1a , ....... "HeP, ~>i ~"O - ":I (\\ ('I -8.:>< VI .... c:rJloll~ N 1oI.,IoIXlASIoI/DATAlSItAPf(J).XUI  SRA TRANSFERS TO DEPENDENT ORGANIZATIONS !'liS 'J1pqe1MDdf 1m COIUtaDt PriC'leS l!l8l 1983 l!l 84 l!l85 1986 l!l81 Inltilution CUmllll Capital CUlTenl Capilal Currenl Capllal C..n-enl Capital C..rreal Capital C_I Capital Proc:uraduri. Aguri. Regisu-o Agrario Nacional Oelegacionel Agrariu Comision Agrarias y Mixtas U40.11 1,60212 14,4U.OO 12.,703.4S 11,4110.63 Nuevo Vallam 1,602.71 35,167.16 2,333.79 1,876.39 Fid. "Cumbre.s de U.no Larso' 4,732.31 5,760.70 13,04B.95 836.84 4,71211 3,763.81 Fid. BabiA de Bandera, lS,322.86 30,530.16 12,315.24 3,215.95 13,281.15 7.507.96 12,804.94 6,014.14 1,475.6S 2,373.88 Fid. Fondo NAI. de FomcnIO Ejida\ 10,271.37 161.49 6,670.17 IUS4 4,837.77 94.24 3,664.24 96231 Fid. Centm de Est. Hisl. del Agr. at Mexico 1,S13.!ItI 104.93 760.85 83.16 1,068.08 6283 935.98 99.S7 844.76 1.185.42 46.13 Ccnu-41 Campe.sin. Indcpcndienlc: 209.86 103.95 94.24 Eslado., MWlicipiDi y Dislrilo Fed_1 14.171.86 2,2$9.81 16,209.66 31.41 1,924.77 830.26 Compra de bien.... aervicio. 'J Remunccac.. 8,124.16 3,903.15 213,932..64 6,000.09 111,614.28 aobicma Federal 491.86 10,234.97 Ayuda • campesinos y OfS.•gr.riu 3,6S7.0S 1,154.11 4,139.62 4,181.02 2,975.09 Subsidio p.r. comprillexpr. de tierra. 11,623.12 37,410.65 30,785.79 14,059.54 12,040.48 103,219.55 atms 213.10 486.82 6,973.92 1,387.56 318.63 224.56 419.74 TOTAL ---.~ _ 70,S49.~3,348.16 49,317.13 319,761.04 53,184.02 166,086.19 55,202.62 86,348.68 32,432.30 4,298.64 ~80~ 876.38 Source: CuenCa Publica (SHCP) t-i>i ~'CI -'CI -= til O\Q.Y ~ I'D __ II""t ~ W N m;\mx\a.m\data\sratmOl  SRA TRANSFERS TO DEPENDENT ORGANIZATIONS N$1bousands 1991 CoosImt Prices 1988 1989 1990 1991 1992 Enlidad CUrTCIII Capilal Cunent Capital Currml Capila. Currenl Capital Currenl Capilal Procuradwia Agraria 53,296.00 33,929.00 Rcgistro Agrario Nacional 22,606.00 18,394.00 Dclegacioncs Agrarial 231.05 95,611.00 Comisi6n Agrariu y Mixeu 10,803.37 10,502.59 19,292.89 1S,307.97 28,289.00 Nuevo Vallarta Fid. "Cwnbn:s de Uano LarIO· Fid. Bahla de Banderas Fid. Fondo Nal. de Fomcnlo Ejidal Fid. Centro do EsL HUt. del Asr. en Mexico 1,934.10 2U4 1,468.06 17.95 1,992.35 42.51 1,926.93 12. 71 Central Campcsina Indcpcndicnte Estado.. Municipios y Distrilo Federal Compra de bienca, scrvicios y RcmUDerac. Gobicmo Federal Ayuda a campcsinol y org. agrari.. 2,431.62 8,672.24 Subiidio para compralexpr. do lierra. 110,262.83 34,S46.19 51,306.48 52,89S.74 324,S47.00 Otro. 437.22 409.22 3,953.72 470.21 52S.72 25S.07 9,000.41 30,642.00 TOTAL 125,869.13 430.76 '0,470.'6 418.16 11,789.61 297.58 79,131.04 243.75 459,380.00 147,934.00 Source: Cuenla Piablica (SHCP) >-i» 2:" S» "tt CD ~ ;;:~~ 0'* •• .. CM N m:\mx\alm\data\8'atmO 1  -, PROCURADURIA AGRARIA FEDERAL EXPENDITURES N$ Thou.sands 1991 f'1~ Prices 1992 1993 State Cunene Capita. Tota. Cunent Capital Tota. Aguascalientes 418 78 496 -1,982 129 2,111 Baja California 306 68 374 2,768 266 3,023 Baja California Sur 234 80 314 1,666 131 1,787 Campacha 361 80 431 3.446 244 3,689 Coahuila 427 129 666 4,181 246 4,426 Colima 203 73 276 1,749 126 1,876 Chiapas 640 232 872 12,983 967 13,940 Chihuahua 870 227 1,096 7,906 791 8,697 Oistrito I=ati.. ral 32,030 26,631 67,661 37,773 6,133 42,906 DUranll" 924 -171 1,096 8,660 607 9,167 Gultl'liljuato 873 216 1,088 7,941 600 8,641 Guerrero 804 297 1,102 8,296 604 8,899 Hidelgo 740 246 986 7,676 684 8,260 Jalisco 799 136 936 8,166 641 8,806 M6xico 1,113 270 1,383 9,769 799 10,667 Michoac6n 817 203 1,020 9,468 826 10,284 Morelos 843 61 904 3,016 128 ~,146 Nayarit 323 73 396 3,666 263 3,818 Nuevo le6n 673 141 716 3,712 246 3,967 Oaxaca 1,061 192 1,243 14,769 1,240 16,999 Puebla 687 162 849 7,941 621 8,662 j Queretaro 249 82 332 2.638 133 2,671 ! Quintana Roo 186 120 304 2,226 126 2,362 San Luis potosr 484 186 670 7,190 629 7,819 Sinaloa 796 204 999 7,328 486 7,814 Sonora 967 218 1,186 6,762 468 7,230 Taba.co 479 116 696 4,068 266 4,313 TamauliplI. Centro 634 166 790 6,083 604 6,688 TlaxcaJa 223 89 312 2,106 129 2.234 Veracruz 1,636 322 1.968 16,886 1.464 17,340! Yucatan 7~ 180 886 4,468 388 4,846! ~» Zac.tecal 337 109 446 6,196 370 6,666 ~"CI ~ - "CI <» ...,. 1"11 .c. TOTAL 61,122 30,448 - 82,187 231,231 20,000 261,231 _=I.!I!J Source: Procuraduda _\.lra~la ~S:~ ~ w ~ m:\mx\asm\data\padaIO 1.xls  I , . .' SEPESCA EXPENDITURE BY PROGRAM 1'11-1"1 NSMiIlioDS 1"1 Constant Prices C.tegory 82 83 84 8& 86 87 88 89 90 91 92 ADMINISTRATION 978 861 763 438 678 314 263 132 135 224 166 RESEARCH 8& TECH. 86 29 27 32 30 28 23 24 25 28 36 INFRAESTRUCTURE 384 164 136 209 133 100 63 79 88 101 97 MARKETING 8 2 2 52 15 20 2 0 EXTENSION 48 40 25 29 20 17 12 4 7 6 5 AGROPROCESSING 3 1 1 105 66 47 23 36 8 TOTAL 1606 1102 963 866 841 626 386 276 262 369 304 Expend"rue ComPHltion by Program C...gorx 82 83 84 86 86 87 88 89 90 91 92 ADMINISTRATION 66% 79% 80% 61% 69% 60% 68% 48% 51% 62% 55% RESEARCH 8& TECH. 6% 3% 3% 4% 4% 6% 6% 9% 10% 8% 12% INFRAESTRUCTURE 26% 15% 14% 24% 16% 19% 16% 29°,," 33 0,," 28% 32% MARKETING 1% 0% 0% 6% 2% 4% 1% 0% 0% EXTENSION 3% 4% 3% 3% 2% 3,% 3% 2% 3% 2% 2% AGROPROCESSING 0% 0% 0% 12% 8% 9% 6% 13°,," 3°A. 0% OTHER 'VALUEI TOTAL 100% 100% 100% 100% 100% 100% 100% 'VALUE I 100% 100% 100% ac.n.: CUInta PubIcIt 18HCP1 ..,» ~"'CS ~ -"'CS ~ -= (I) ooQ.~ ;:c. "' W N M:MX\ASM\DAT~\PESPRGjl.XLS  SEPESCA EXPENDll'URES BY PROGRAM AND CATEGORY N.T......... '89Z CoM_a ...... 1986 1986 1987 1988 Program Current Capital Total Current Capital Total Current Capital Total Current Capital Total Administration 422,663 16,493 438,166 332,963 246,076 678,039 214.161 99,460 313.612 189.302 74,140 263,442 R. & T. 29,633 2,689 32.21~ 21.695 1,860 29,646 26.664 1,361 27,924 22.908 187 23,095 Infraestructure 116.046 94.314 ",(,q, 'l6C 67,496 66, HI6 132,691 63,201 36,361 99,662 44,794 18,492 63,287 Marketing 62.166 0 62.166 16.196 0 16,196 20.023 0 20,023 2,298 0 2,298 Extension 17,406 11,869 29.274 16,099 6,047 20,146 12,094 6,083 17,177 9,270 2,302 11,672 Agroprocessing 77,248 27,641 104,889 66.524 8,640 66,104- 24,802 21,]30 46,631 19,427 3,146 22,672 Other programs 0 0 0 0 0 0 0 0 0 0 0 0 Total 714,149 161,907 888,066 614,871 326,&68 aMO,780 -- 360,836 163,994 624.829 287.999 98.266 386,266 Sour".: Cuenta Publica 18HCP) o-i» ~"O ~ _"0 ~ ('0 I't I.!!.:I .... IO~ =:>< I» 11"1 ~ N m:\mx\lIsm\dlltll\peaprgI3  SEPESCA EXPENDl11JRES BY PROGRAM AND CATEGORY 1989 - 1992 N$ 'IbousaIIcIs 1992 CODStant Prices 1989 1990 1991 1992 Program8. Current Capital Total Current Capital Total Current Capital Total Current Capital Total Administration 127,360 4,788 132,138 0 0 134.70 129,913 94,269 224,172 146,827 20,407 166,234 R. & T. 23,266 666 23,812 0 0 26.13 26,764 1,061 27,816 32,696 3,208 36,803 Infraestructure 39,868 39,630 79,186 0 0 87.68 30,731 70,309 101,040 36,176 61,626 96,701 Marketing 0 293 293 0 0 0.00 0 0 0 0 0 0 Extension 3,847 776 4,422 0 0 7.06 6,814 0 6,814 4,933 0 4,933 Agrolndultry 35,914 0 36,914 0 0 7.69 0 0 0 0 0 0 Other Programs 0 a 391 0 0 0.00 0 0 0 0 a 0 Tota. 229,822 46,943 276,166 0 0 262.18 193,212 186.828 368,840 218.630 86.141 303,871 80..0.: ClHlnta Pllblo. (SHep, >-i» ~ "CI 2! -"CI ~ -= ('0 l"D \Oe:~ .., eJ"1iOIc,u m:\mx\asm\data\pesprgi3  SEPESCA TRANSFERS TO DEPENDENT ORGANIZATIONS N$MiI6oos 19t1 CODstant Pric:el .,.1 .,.3 I"" ENTIDAD ~I CAP. cn.1 CAP. cn.1 CAP. CTE. I 1"5 CAP. CTE. U8' I CAP. CTE. I U87 CAP. - ORGANOS DESCONCENTIW)()S INSTITUTO NACIONAL DE PESCA 6.B4 5.60 3.94 SUMA 6.84 6.60 3.94 ENTIDADES DE CONTllOL DIRECTO PRODUCTOS PESOUEROS MEXICANOS E INDUSTRIAS PESOUERAS PARAESTATALES DEL NOROESTE 1-' 61.89 627.78 76.80 630.32 186.88 361.41 PRODUCTOS PESOUEROS MEXICANOS S.A. DE C.V. 462.66 46.60 294.93 266.60 142.32 123.64 (PROPEMEX) 1..1 SUMA 61.B9 527.78 78.80 830.32 186.BB 361.41 462.66 46.60 294.93 266.60 142.32 123.54 I ENTIDADES DE CONTllOL JND.lJlEcto FIDEICOMISO FONDO NAl. PARA El DES. PESOUERO 3.00 23.84 0.77 4.97 4.61 FIDEICOMISO PARA EL FOM. Y APOYO DEL DES. PESO. 1-' 164.06 36.84 SUMA 164.06 36.84 3.00 23.84 0.77 4.97 4.61 SUMAN LAS TRANSF. AL SECTOR PARAESTATAL 61.89 691.84 113.64 630.32 189.88 386.26 469.40 46.60 301.19 271.66 160.87 123.64 o T RA S TRANSFERENCIAS 2-' 1.33 0.94 0.83 I TOTAL 83.22 891.84 113.84 830.32 190.82 386.26 470.03 48.80 301.19 271.66 160.87 123.64 601lfGII: CUlntll NIIio.IISHCPI ~» ~'C 2: -'C .., CD f'D """ N:ll.!II!J O~ .... ~ ~. ~ W rn:\rnx\a8rn\data\p••trni02  SEPESCA TRANSFERS TO DEPENDENT ORGANIZATIONS N$MlIIloas .992 Constant Prices I Utt Ut, I 19" 1991 19 91 CTE. CTE. ENTIDAD I CAP. CTL CAP. CAP. CTE.I CAP. CTE. I CAP. ORGANOS DElICONCEN'J'lW)09 lNSTITUTO NAC10NAL DE PESCA 3.03 3.69 6.96 8.20 11.56 SUMA 3.03 3.69 6.96 8.20 11.66 ENTIDADElI DE CONTROL DIB2Cl'O PRODUCTOS PESQUEROS MEXICANOS E lNDUSTRIAS PESQUERAS PARAESTAT ALES DEL NOROESTE 1.1 PRODUCTOS PESQUEROS MEXICANOS S.A. DE C.V. 109.94 76.37 86.08 IPROPEMEX) 1-' SUMA 109.94 78.37 66.08 ENTIDADElI DE CONI'ROL INDIllECl'O FIDEICOMISO FONDO NAL. PARA EL DES. PESQUERO FIDEICOMISO PARA EL FOM. V APOVO DEL DES. PESQ. 1.1 SUMA SUMAN LAS TRANSF. AL SECTOR PARAESTATAL 112.97 78.37 88.86 8.96 8.20 oT RA S TRANSFERENCIAS 2) 2.33 1.76 TOTAL 11&.30 78,1' 88.8& 8.86 8.20 11.68 --- --- ~ ao-: c...... P..... f8HCPJ I-i» g."C:I ...,. Z -"C:I ~~"'"" N::Il!I!j o=. ..... cr ;;. ~ . m:\mx\a.m\data\p ••trni02 . N  SEPESCA DELEGADON EXPENDITURES N$1bousaods 1m CoostaDt Prices Shlte 84 86 86 81 88 89 90 91 92 Aguascalientea 204 809 386 89 96 172 39 309 51 Baja California 7,860 2,390 1,206 176 68 490 954 336 244 Baja California Sur 2,111 3,035 1,385 998 263 1,399 298 523 1,205 C8Illpeche 735 1,201 461 146 230 202 407 288 1,555 Coahuila 151 372 366 426 296 101 147 16 200 Colima 1,929 2.563 3,265 6,292 2,008 2.372 7,082 3,596 1.758 Chiapas 2,664 10,903 4,699 23 1,306 232 945 146 1.180 Chihuahua 1,002 888 507 242 543 503 128 291 11 Distrito Federel 27,852 16,149 3,449 14,804 4,527 13,912 33,534 96.640 41,460 Durango 264 432 231 281 541 309 708 515 593 Guanajuato 9 196 170 81 202 153 77 140 259 Guerrero 104 4,931 1,348 168 1,806 193 1,212 844 643 Hidalgo 86 918 670 133 250 487 251 130 162 Jalisco 91 368 1.206 0 736 96 1.529 1,921 1.182 Mexico 16 880 154 128 280 55 589 538 17 Michoacan 1.787 4,262 1,966 963 368 431 773 448 1.622 Morelos 296 711 314 46 620 339 284 16 200 Nayarit 8.063 1.474 1.261 1,588 621 1,146 1.990 1,030 2,170 Nuevo LeOn 0 0 84 1 68 21 21 29 36 ! Oaxaca 983 1.312 458 2,027 3.403 3,574 6,001 7.281 9. 262 1 Puabla' 226 584 331 108 116 102 58 133 161 Queretaro 314 0 84 0 65 17 388 88 97 Quintana Roo 2,426 2,167 434 311 138 1,685 1,481 1,844 1.474 San Luis Potos' 19 442 167 32 191 139 121 196 1.516 Sinaloa 4,184 3,616 2,972 2,212 837 1,778 9,1,22 12,006 4.469 Sonora 3,783 4,831 706 243 812 372 4,307 4,035 901 Tabasco 4,219 1.954 601 143 279 183 100 560 425 Tamaulipas 1.438 4,098 2,182 1,239 217 4,582 4,091 2,753 3,756 Tlaxcala 562 358 254 211 919 98 310 1,816 14 Veracruz 4,671 2,496 710 311 818 159 168 531 4.424 Yucatan 7,291 11,212 3,860 181 126 3,813 2,108 1,713 924 Zacatecas 0 191 78 0 163 189 43 118 0 Subtotal 84,403 86,684 36,830 32,689 22,366 39,184 79,261 140,763 82.667 Progr. Cobartura Na!. 8,699 i.liP'; :7.591 0 0 0 0 0 0 Proyectos sin Danom. 682 914 0 0 0 0 0 0 0 ~» TOTAL Soure.: SEPESCA 93,784 94,084 43,421 - 32,889 22,385 39,184 79,287 140.783 82,567 ~'= ~ -,= ~ tt ~ N= . . . . o -Q,,11"'0 ~. ~ m:\mx\asm\datll\pesdeiO 1.)(ls t-.)  CONASUPO EXPENDITURES AND REVENUES N. THOUlMD. l . .Z CON.TMT I'NCP Program 82 83 84 86 88 87 88 89 10 91 82 EXPENDITURES \1 17,570,043 18,946,889 18,759,888 12,832,178 7,414,507 7,289,633 8.713.372 9,175,064 8.923.684 8,961,860 6.920,740 Purchase, Siorage. Marketing 17,570,043 18.946.889 18.759,888 12.832,978 7.414,607 7,289,533 8,713.312 8,633.318 6.923.584 5,338,001 5.101,537 Corn 0 0 4,741,671 4,499,889 3.240,887 2.831.564 3.588.680 4,098.448 3.191.599 n/a nla Wheat 0 0 2,572,184 2.850.144 660.182 687.386 888,330 413.714 117.211 nle nla Grain. 0 0 3,622,294 2.869,461 617.326 1.415.518 2,015,842 1,001.666 166.820 niB niB legumBI 0 0 639,151 618.132 670.083 199.808 485,811 746,446 815,528 n/. n/. 011 leads 0 0 6,104,878 1,281,828 888,864 672,824 618,186 168,323 70,288 nJa nla Powdered Milk 0 0 687,560 600,681 636,338 841.701 1,336,026 2.144,143 1,883,968 niB nI. Other 0 0 392,142 113,034 8911.868 84.648 143.898 60,533 118.176 nla n/. Admini,tr.tlon 0 0 0 0 0 0 0 489.928 0 378.839 359,480 Construction 0 0 0 0 0 0 0 11,633 0 10,528 5.251 Social Ani,tanea 0 0 0 0 0 0 0 40,174 0 1.224.181 1,460,881 Buildingl 0 0 0 0 0 0 0 0 0 2,312 3,581 INCOME 26,8118,222 26,302,251 21,054.619 18,886.864 6.793,353 14.733,220 13,442.866 13.1184,651 11,488.188 1,392,669 6,820.283 ReVllnUIIS of goods & .arvlce. 9,666,317 8,512,241 8,618,644 1,466,537 nla 6,888,207 6,330,184 3,846,837 4,065,026 3,903.780 3.220,334 Other Income 254,250 987,629 1,419,478 298,116 nla 451,430 172,667 249,205 630,942 17,204 30.436 Intarnal Financing 3.032.424 8,1185,003 2,1156,488 458,030 niB 758.815 351,799 135.141 0 nla niB External Financing 4.627,177 575.862 3,423,767 1.533,406 niB 2.848.810 3,773,078 4,771,507 -4.140,235 nla nfa Government Trenlfer. \2 8,616,055 8,5111,616 10,938,274 9.240,286 Jt,793,353 4.974,889 3.815.238 4,882,087 11.052.456 3.411,865 3,£1119.613 1 Source.: CUlnta Publica, SHCP, except fllcal tranarert ror 1988 are rrom World Bank. -Meldco Agricultural Sector Report •• No. 71100-ME. 1988, p. 18 and Income data for 1891·92 are from CONASUPO. \ 1 Exclude, amorllzatlon on the debt. \2 Includes transfall for debt amortization after 19B5. 1-:1;;';;' ~" ~ 0-1 t-.)I:I t-.)S:~ ~ eN N m:\mx\88I1\\data\conprQO 1  BREAKDOWN OF CONASUPO EXPENDITURES AND REVENUE IN PERCENTAGES Program 82 83 84 86 8. 87 88 89 90 91 92 EXPENDITURES \1 100% 100% 100% 100% 100% 100% 100"", 100% 100% 100% 100% Purchase, Storage, Marketing 100% 100% 100% lOa,," 100% 100% 100% 94% 100% 77% 74% Corn 0''- O'*' 25% 36% 44% 39% 41 % 45% 46% nla nla Wheat O'*' 0% 14% 23% 9'" '''' 8"4 5% 3% nla nla Graina ...,". U'lI> I;:' ::4 23% B% 20% 23% 11 % 11% nla nla LeQumea 0 0,.- 0% 3% 4% B% 11% 6% 8% 8% n/a nla Oil aeeds 0% 0% 33% 10% 13% 8% 6% 2% 1% nla nla Powdered MHk 0% 0% 4% 4% 9% 12" 16% 23" 29% nla nla Othar 0% 0% 2% 1% 9% 1% 2% 1% 2% n/a nla Administration 0% 0% 0% 0% 0% 0')(0 0% 5% 0% 5% 5% Construction 0% 0% 0% 0% 0% 0')(0 0% 0')(0 0% 0% 0% Socia. Al8iltance 0% 0% 0% 0% 0% 0% 0% 0% 0% 18% 21% Buildings 0% 0% 0% 0"- 0% 0"- 0"- 0% 0% 0% 0% 1NCOME 100% 100% 100,," 100"- nla 100')(0 100% 100% 100% nla nla Revenues of goods &. service. 37% 34% 32% 39% n/a 40% 40% 28"- 36% nla nla Othe, income 1~ 4% 5')(0 2% nla 3% 1% 2"- 6% nla n/a!I Intelllal Financing 12'" 34" 10"- 2"- nI,. 6% 3" 1% 0% n/. 'o/a! Edama' Financing 18% 2% 13% 8')(0 nla 18')(0 28% 34% -36% nla nla! GovernmeD! Tranafers \~ --~ - - -- ............... -.- ........ .... 3~~_ 26% 40')(0 48% nla 34% 28% 36"- 86% nla nJa Sourca: S•• Taxt Tabia 8.7 ~» ~-= Z (Dog Z N=l'!!J we:~ ~ ~ m;\mx\a.m\data\conprgO 1 t-.;t  SOLIDARITY AND REGIONAL DEVELOPMENT CRAMO XXVI) EXPENDITURES 1992 Conetant Million N$ 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 Drinking Water and Drainage \ 1 0 0 0 0 0 0 0 217 380 474 738 Electrificetion \1 0 0 0 0 0 0 0 122 186 232 226 Fund for Solidarity Bueineeus 0 0 0 0 0 0 0 0 0 0 368 Solidarity Production Funds 0 0 0 0 0 0 0 0 560 437 639 Regiona' Development Programs \2 0 0 0 0 0 0 0 411 344 459 281 Rural Roads and Highways 0 0 0 0 0 0 0 613 526 864 970 Municipal Solidarity Funds 0 0 0 0 0 0 0 0 366 369 641 Indiganous Communities 0 0 0 0 0 0 0 65 191 255 1581 Basic Infrastructure \3 nla nla 839 699 300 236 168 \4 \4 \4 \4] Productive Programs \3 nla nla 484 771 488 528 399 \4 \4 \4 \41 Tota' Rural Productive Programs nla nla 1,323 1,370 788 764 567 1,328 2,550 3,089 Total Solidarity/RAMO XXVI n/a n/a 4,576 3,527 2,149 2,349 1,853 2,943 4,644 5,991 3,92:1 6,817 Shara of Total Solidsrity n/a n/a 29% 39,*, 37% 33$ 31% 45°111 65% 62$ 58% \ 1 Includes en estimetod 75% of total costs a18umed to occur in rufel aroas. \2 Includes an estimated 50% of total costs aesumed to occur in rural areae. \3 A subprogram that exieted in the Regional Development Program, the predecessor to Solidarity. Source! IV Government Report 1992 /Informs IV de Gobiernol \4 After 1989, the Solidarity classification of RAMO XXVlexpendituree is ueed. PROGRAM SHARE OF SOUDARITY AND REGIONAL DEVELOPMENT (RAMO XXVI) In Percentages 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 Drinking Water and Drainage \1 0% 0% 0% 0% 0% 16% 15% 15% 19% Electrification \1 0% 0% 0% 0% 0% 9% 7% 8% 6% Fund for Solidarity Businesess 0% 0% 0% 0% 0% 0% 0% 0% 9% Solidarity Production Funds 0% 0% 0% 0% 0% 0% 22% 14% 16% Regional Development Programs \2 0% 0% 0% 0% 0% 31% 13% 15% 7% Rural Roads and Highways 0% 0% 0% 0% 0% 39% 21% 28% 25%1 Municipal Solidarity Funds 0% 0% 0% 0% 0% 0% 14% 12% 14% Indigenous Communities 0% 0% 0% 0% 0% 6% 7% 8% 4% Basic Infrastructure \3 63% 44% 38% 31% 30% \4 \4 -\4 \4! Productive Programs \3 37% 56% 62% 69% 70% \4 \4 \4 \4 Total Rural Productive Programs 100% 100% 100% 100% 100% 100% 100% 100% 100% t-i» ~"= ~ -,,= ~ = Oft>""" N ~ .j:l.Q.~ ... ;::;. ~ ~ N c.. m:\mx\asm\data\solid.xls N Ln  en EXPENDrn.rD8 J'OIl RURAL l1'.LECTIlD'ICATlON N$~ 1"2 eo.....at Prie.. Pro Rural EI"trili~.l.ioD 81 340,920 113 212,796 ... 237,553 as 373.096 16 129,280 .., 157,260 118 252,514 ., 257,988 90 82,165 n.d. 91 n.d. 9% Total!!xp~illlrc of ~ 22,0'-2,291 23,638,794 24,684.045 24,249,.5!2 _ 2,2.6E1 ,:'16 I ~. 099 .n1.s_ 1~;1.h46() _ _1'5. i79,057~6-,-714~ 18 17,591,073 D.d. Pro,ram Rural ElectrificaliOD ,. 2S 113 IS 84 IS 85 2S " IS .., IS II 2S .9 2S ,. OS D.d. 91 n.d. 91 I Total Expe:adilurc of CFE 100S 100S 100" 100" l00S 100" l00S 100S 100" 100" n.d. I Source:: Cue:1I1Il Public,a (SHCP) ~>~ 2::g 2! ('I)tDl'!l N[~ 0\ •• 1004 ~ N m:ImlI\oMmldatala,.",QOt  SCT RU1lA.L DEVELOPMENT EXPENDJ11JBJ!S NSTII_adI 1m c-taa' I"riea PlOllun at 83 84 89 86 87 88 89 90 91 92 Tot.lonio Awol 311,021 21,392 0 1,434 128 1,053 3,3C19 2.303 4,291 27.770 12.667 Carrelar... Troncale. 0 2,881.245 3,031.298 0 780,635 146,701 664.211 986• • 1.239,452 1.561.061 965.422 Calf, AI....ol.do,.. y Cam.R..... 0 478,817 643.579 882,691 1«,443 221,259 I 14,BS3 333,689 320,363 509,770 346,670 Subt..u! 36,621 3.393.454 3.614.877 864.026 926.106 916.018 682,432 1.322.637 1.564.095 2.098,599 1,324.159 _ __ - .... .... TOT A l GLOBAL 11,218.245_ _12.208.783 12,971.260 12,871.101 If.2119 c31! __10,294.6.1 6,269.613 4,631.23. 4,481,573 5.890,090 6,651.866' &pondit... C..........1Ilon "'OQRm.. 82 83 84 85 86 87 88 8. 90 91 lIZ T0"'000. Rwol 0% 0% 0% 010 0'" 0% 0% 0% 0% 0% 0% Cam,leraa T,oncale. 0", 2ot'" 23% 0'" 8% 1% 9% 21% 28% 26% 11% Carr. Alim.ntado,.. V Cam,RUt.I.. 0% 4% Ii% Ii% 2% 2% 2% 1% 7% 8% 6% ToTAL I~~% 100% 100'" __ lOO~_ 100% 100% 100'" 100% __ 100% ___ 100% I Sour c.; C... nta Publica ISHCPI ~>i ~", ...... ", ~ ftI N= ~a:~ N m;\m"IMmId.t.\oclprgOl .....  SCT EXl'END:rnJIlB BY PIlOGRAM AND CATEGORY N$n...... 19J2Coattut ..... 1985 1986 1981 19881 Procram Cumat Clpill!. Clamat Clpilll Currcot ~ilal Cumat C.pilal Tclcfc:ufll Rural 0 1,414 128 0 7,05:1 0 :I,369! c.m:t.etu Troocalcl 0 0 710,407 37 746,670 12 564,129 Carr. Alimalladoru y c.m.Rurall:a 691 661,194 143,940 401 220,858 625 114,128 TOTAL - 4.429,692. !,:441,914 5,912,790 2,183,007 7,411.688 1,462,562 4,801,0«)1 Expenditure Ccmpoaition I 1985 1986 1987 1988 1 Program I r.umat capilli Cuma& Capilli Clamot C~ilal Clarrent C~ital Telcfc:ufll Rura! I 0"' "' 0 .. 0" 05 05 0" 1 Clmleru TroaWea 0" 0.. 13 .. 0.. 10" 0 .. 12"! Cm. AIimcaLldoru >' Com. Rur.lca 0" I" 2" 0" 3 .. 0 .. al ___TOTAL _- .. .. 100" 100.. 100,. 100,. 100,. 100,. 100" Sow...: Cuont<I PublIca (SHCPI ~>i ~"= -,,= ~ til N=>e 005:: l» ~ ~ m:'mx'-nVJ.t......lprg02.:da N  SCI' EXPENDI'lVRE BY ftlOGKAM AND CATEGORY N$ ............ 1m c..tut Pria:I 1919 1990 1991 1m Prosram Clamat Capital CwTcaI. Cipital OImat Capital OImat Capilal T...oaIa Jtun1 22 2,211 0 0 1,037 26,740 161 12,SOO CamIcru TrOllClliet 11O,21S 106,311 0 0 163,241 1,391,810 195 965,221 Carr. Alimcntadoru y CIID.llunIeJ 513 333,156 0 0 2,682 S07,oaa 39 346,631 TOTAL 1,693,025 2,931,214 0 0 2,071,222 3,926,868 2,291,695 3,253,171 19.9 1990 1991 1m ProerllDU OIrreat OIrreat Capital Clarreat Capit.U Telefoaia !.ural o~ . Capital o. Clanaat Capital 0" •• 0" 0" Cll1'fW;ru TrOl1\:&1ca 11" 21" '" 36" 0" 30r. Carr. Alimaltadoru y CIID.llura1ea 0" n. or. 13. 0" 11" -- .... _-- TOTAL --_- ...... 1001 100" NIA -- N/A -- -_ _-.... 100" 100" 100" _ ....... - 100" ...,» _'C ~ go'C G tt !V= oo~>< c::rll'lc".l N m:Vnx'..m\Ot.\,mPfg02.lde  ANNEX 3 Appendix 3 Table 1 Distribution by State of General Fund Revenue Sharing Dependent Variable: State Allocation of General Fund Revenue Sharing (in N$) Explanatory Variables 1990 1991 1992 Populationll 232.8 292.4 333.2 (7. 12f** (8.63) .... (10.31)--- CONAPO Marginality Index -IS0.3 x 106 -191.9 X 106 -189.5 x lQ6 (2.50) (2.57)" (2. 67r* Per Capita21 State GNP -1.6 x 106 -1.7 x lQ6 -1.6 x lQ6 (0.64) (0.64) Constant 93.2 x 106 122.5 X 106 124 X 106 (0.47) (0.60) (0.64) Adjusted R-Square 0.66 0.74 O.SO Notes: Statistically significant at the 10 percent confidence level. Statistically significant at the 5 percent confidence level. Statistically significant at the 1 percent confidence level. 11 All years use 1990 INEGI Census Data. "JI All years use 1992 GNP/Capita by state from Excelsior, November 29, 1993. M:\mx\asm\greeo\am:3app3.tab  ANNEX 3 Appendix 3 Table 2 Per Capita Collections of Water Charges by State Dependent Variable: State Per Capita Collections of Water Charges (in N$) Explanatory Variables 1990 1991 1992 Populationll 3.8 6.0 15.9 (0.87) (0.93) (1.66)* CONAPO Marginality Index 6.9 15.9 27.2 x 1()6 (0.71) (1.13) (1.28) Per Capita21 State GNP 1.0 1.3 1.9 x 1()6 (2.86)*** (2.58) (2.57) Constant -19.8 -19.2 x 1()6 -27.2 x 1()6 (0.74) (0.50) (0.39 Adjusted R-Square 0.15 0.12 0.17 Notes: Statistically significant at the 10 percent confidence level. Statistically significant at the S percent confidence level. Statistically significant at the 1 percent confidence level. 1/ All years use 1990 INEGI Census Data. '11 All years use 1992 GNP/Capita by state from Excelsior, November 29, 1993. M:\m.x\asm\greeo\anx3app3.tab ~ •  ANNEX 3 Appendix 3 Table 3 Per Capita Collections of the Property Tax (Predial) by State Dependent Variable: State Per Capita Collections of the Property Tax Explanatory Variables 1990 1991 1992 Populationll -1.9 x 106 -3.14 X 106 -5.6 X 106 (0.40) (0.45) CONAPO Marginality Index 1.49 3.1 5.2 (0.14) (0.16) (0.19) Per Capita21 State GNP 0.9 1.7 2.4 (2.5)** (2.47) (2.42)"- Constant -28.7 -54.7 -74.8 (0.99) (1.04) (1.00) Adjusted R-Square 0.11 0.10 0.11 Notes: Statistically significant at the 10 percent confidence level. Statistically significant at the 5 percent coDfidence level. Statistically significant at the 1 percent confidence level. IJ All years use 1990 INEGI Census Data. '11 All years use 1992 GNP/Capita by state from Excelsior, November 29, 1993. M:\m:x\asm\arem\anx3app3.tab  ANNEX 3 Appendix 3 Table 4 Distribution of Municipal Fund Revenue Sharing Dependent Variable: State Allocation of Municipal Fund Revenue Sharing (in N$) Explanatory Variables 1990 1991 1992 Populatio~lI -0.3 5.6 6.9 (0.61) (2.49)·· (2.61)·· CONASUPO Marginality Index 1.49 -5.0 x 106 -2.9 X 106 (0.14) (1.01) (0.50) Per Capita2! State GNP -0.1 x lij6 -0.2 X 106 -0.1 x 1~ (0.39) (0.90) (0.51) Constant 11.4 x 106 -54.7 13.2 x 1~ (1.04) (0.83) Adjusted R-Square 0.33 0.10 0.14 Notes: Statistically significant at the 10 percent confidence level. Statistically significant at the 5 percent confidence level. Statistically significant at the 1 percent confidence level. 11 All years use 1990 INEGI Cel18Us Data. '1J All years use 1992 GNP/Capita by state from Excelsior, November 29, 1993. M:\Jnx\asm \grem\aDx3app3.ta.b •  . ANNEX 3 Appendix 3 Table 5 Distribution of Aquasealientes Revenue Sharing to Municipalities in Aguascalientes Dependent Variable: Per Capita Revenue Sharing by Municipality (in N$) Explanatory Variables 1990 1992 Population1 -1.2 x 10' -2.0 X 10' (0.56) (0.78) CONAPO Marginality Index 0.16 0.25 (0.94) (1.29) Population Density 0.18 x 1()3 0.61 x 1()3 (0.15) (0.44) Constant 0.28 0.36 (2.87)·· (3.26)" Adjusted R-Square 0.44 0.51 No. of Observations 9 9 Notes: Statistically significant at the 10 percent confidence level. Statistically significant at the 5 percent confidence level. Statistically significant at the 1 percent confidence level. II All years use 1990 INEGI Census Data. '1J All years use 1992 GNP/Capita by state from Excelsior, November 29, 1993. M:\mx\asm\green\am3app3.tab  ANNEX 3 Appendix 3 Table 6 Distribution of Revenue Sharing to Municipalities in Sonora Dependent Variable: Per Capita Revenue Sharing by Municipality (in NS) Explanatory Variables 1990 1m Populationll 0.9 x 1()3 1.8xl()3 (0.11) (0.1) CONAPO Marginality Index 1.3xl()3 2.5 (1.36) (1.26) Population Density -19.3 -39.6 (0.47) (0.45) Constant 2,570 5.1 x 1()3 (2.17)" (2.03)" Adjusted R-Square 0.04 0.03 No. of Observations 70 70 Notes: Statistically significant at 1he 10 percent confidence level. Statistically significant at 1he 5 percent confidence level. Statistically significant at 1he 1 percent confidence level. II All years use 1990 INEGI Censu8 Data. 1I All year8 use 1992 GNP/Capita by state from Excelsior, November 29, 1993. M:\mx\asm\greeo\amt3app3.tab  • ANNEX 3 Appendix 3 Table 7 Distribution of Revenue Sharing Resources to Municipalities in Tabaseo Dependent Variable: Per Capita Revenue Sharing by Municipality (in N$I) Explanatory Variables 1990 1992 Populationll -4.8 x 107 -6.3 X 107 (2.52)** (2.02)* CONAPO Marginality Index -4.1 x 1()2 -10.6 x 1()3 (1.24) (1.29) Population Density 0.26 x 1()3 -0.19 (1.43) (0.61) Constant 0.25 0.25 (12.9)"* (7.41)'" Adjusted R-Square 0.55 0.36 No. of Observations 17 17 Notes: Statistically significant at the 10 percent confidence level. Statistically significant at the 5 percent confidence level. Statistically significant at the 1 percent confidence level. II All years use 1990 INEGI Census Data. 11 All years use 1992 GNP/Capita by state from Excelsior, November 29, 1993. M:\mx\asm\greeu\a0x3app3.tab  ANNEX 3 Appendix 3 Table 8 Federal Solidarity Expenditures by State Dependent Variable: Federal Solidarity Expenditures (millions N$) Explanatory Variables 1990 1991 1992 Population1! 14.2 20.0 x 1()3 26.7 x 1()3 (3.49)*" (3.51)*** (3.24)*** CONAPO Marginality Index 21.9 43.6 53.1 (2.57)** (3.66)*** (3.09)*** Per Capita2! State GNP 439.5 508.2 460.3 (1.50) (1.24) (0.77) Constant 35.1 68.7 x 1()3 104.6 x 1()3 (1.61) (2.23)** (2.37)** Adjusted R-Square 0.38 0.45 0.38 Notes: Statistically significant at the 10 percent confidence level. Statistically significant at the 5 percent confidence level. Statistically significant at the 1 percent confidence level. u All years use 1990 INEGI Census Data. " All years use 1992 GNP/Capita by state from Excelsior, November 29, 1993. M:\mx\asml.greeo\anx3appJ.tab •  ANNEX 3 Appendix 3 Table 9 Distribution by State of Federal Solidarity Municipal Funds Dependent Variable: State Allocation of Federal Municipal Funds (in thousands of N$) Explanatory Variables 1990 1991 1992 Populationll -0.6 x 1(f 1.2 x 1(f 3.4 x 1(f (0.61) (2.49)*- (3.14)'" CONAPO Marginality Index 6.6 x 1(f 4.3 x 1(f 8.2 x 1(f (2.32)-- (1.64) (3.64)*-- Per Capita21 State GNP 24.5 -6.8 9.9 (0.25) (0.06) (0.10) Constant 9.3 20.9 x 10 3 10.1 x 1(f (1.15) (1.89) (1.50) Adjusted R-Square 0.15 0.05 0.46 No. of Observations 19 20 27 Notes: Statistically significant at the 10 percent confidence level. Statistically significant at the S percent confidence level. Statistically significant at the 1 percent confidence level. II All years use 1990 INEGl Census Data. '11 All years use 1992 GNP/Capita by state from Excelsior, November 29, 1993. M:\nu\asm\greea\BDX3appl.tab  ANNEX 3 Appendix 3 Table 10 State Distribution of Federal Solidarity Production Funds Dependent Variable: State Allocation of Solidarity Production Funds (in thousands of N$) Explanatory Variables 1990 1991 1992 Population 1 0.6 1.3xl()l 2.1 x l()l (7.12)"- (1.45) (1.85)- Percent in Population 360.2 336.5 529.4 in towns with < 5,000 (1.98)- (2.12)*- (2.65)*- people Percent of Population with -239.8 -131.4 -588.2 < than two minimum wages (0.70) (0.45) (1.61) Per Capita21 State GNP 126.8 116.9 116.4 (1.70) (1. 79) (1.42) Constant 6.9 x 1()l -1.2 x 10 3 20.3 x 1()l (0.38) Adjusted R-Square 0.08 0.18 0.17 No. of Observations 27 28 29 Notes: Statistically significant at the 10 percent confidence level. StatiBiically significant at the 5 percent confidence level. Statistically significant at the 1 percent confidence level. U All years use 1990 INEGI Census Data. 'JI All years use 1992 GNP/Capita by state from Excelsior, November 29, 1993. M:\mx\asm\green\amr3app3.tab  ~ ANNEX 3 Appendix 3 Table 11 Distribution by State of SARH Delegation Expenditures Dependent Variable: State Delegation Expenditure (in thousands of N$) Explanatory Variables 1990 1991 1992 Population1 1.5 x 1()3 1.9 x 1()3 3.4 x 1()3 (2.30) (1.86)* (3.14)*** CONAPO Marginality Index 2.1 x 1()3 3.9 x 1()3 8.2 x l(f (1.32) (1.70) (3.64)*** Per Capita21 State GNP 113.5 172.6 9.9 (1.99)*- (2.10)** (0.10) Constant 9.9 x 1()3 12.7 x 10 3 16.4 x 1()3 (2.30) (2.05) (3.09) au Adjusted R-Square 0.14 0.15 0.26 Notes: Statistically significant at the 10 percent confidence level. Statistically significant at the S percent confidence level. Statistically significant at the 1 percent confidence level. 11 All years use 1990 INEGI Census Data. '11 All years use 1992 GNP/Capita by state from Excelsior, November 29, 1993. M:\m:I;\asm \green\aDx3appl.tab  ANNEX 4: INTEGRATION OF MEXICAN MARKETS 1. The Model 1. This Annex describes the model developed to examine the level of integration of Mexican markets. Market integration models essentially make a "Law of One Price" (LOP) argument. Suppose that a market for a specific commodity consists of two locations, 1 and 2. If trade takes place, imbalances in demand and supply will lead to large regional price disparities and hence arbitrage opportunities. Eventually, price fluctuations originating from location 1 will be transmitted to the price in location 2 and vice-versa. 2. Let PIt and PZI denote the price of the commodity under consideration in location 1 and 2. Correspondingly PIt and PZI may denote the domestic and the world price of the commodity. Early studies on market integration (or LOP) were based on the following type of regression [e.g., Isard (1977), Richardson (1978)]:1 (1) Then, the hypothesis that the slope coefficient equals unity and (possibly) the intercept term equals zero is tested. Formally: HO" p. + 1 = {:J = 1 against HI: Ho is not true. 3. Under Ho (1) becomes PIt = PZI' However, some statistical properties of the series involved in the regression, namely nonstationarity, may invalidate standard econometric tests and therefore give misleading results regarding market integration. Furthermore, the test is very restrictive, especially when applied to developing economies where the links between markets are not very strong, hence it is rather unlikely that the two prices will differ by a white noise only as Ho of (1) dictates. 4. The first problem can be overcome by examining the stationarity properties of the error term of (1) (Ardeni, 1989).1 Given the overwhelming evidence that prices are nonstationary, the existence of a stationary error term in (1), which implies co movement between the two prices, leads to the conclusion that markets are characterized by long-run integration. 3 However, if the slope coefficient is different from unity, the corresponding price differential is growing and such growth is not being accounted for; consequently stationarity of the error term is not a testable form comparable to the Ho of (1). 5. To account for the non-unity slope coefficient problem, one can restrict the parameters of (l) according to Ho and then test the following: (2) (PIt - P21) - 1(0). If the price differential as defined in (2) is stationary, then one may conclude that the market consisting of locations 1 and 2 is characterized by long-run integration. 6. It is clear from (2) that cointegration tests are not very powerful as they only make inferences about the existence of the moments of the distribution of (PIt - p,J and not about certain restrictions  -2- that may be required by economic theory [e.g., Ho of (1)]," Therefore, as was the case with the co integration regression (I), (2) cannot serve as a substitute for the Ho of (1); as demonstrated below, it can onJy serve as an intermediate step in establishing the validity of Ho of (1) or any other hypothesis of interest. 7. An advantage of (2) over (1), however, is that it implies the existence of the following dynamic specification: (3) As all series in (3) are stationary, several hypotheses can be tested with conventional F-tests. 5 For example, HO" p. + 1 = a = {3 = 1 against HI: Ho is not true. Under Ho (3) becomes PIt = P2I' which gives the same outcome as the null associated with (1). In a sense, (3) quantifies the long-run nature of (2), i.e., it tells us "how long is the long-run".' 8. The interpretation of the parameters in (3) is as follows: {3 indicates how much of a given change in the price of region 2 will be transmitted to the price of region 1 in the first period (initial adjustment); a indicates how much of the past price disequilibria between the prices of the two locations is eliminated in each period (error-correction or speed of adjustment); Both coefficients are expected to fall within the [0,1] interval. The closer to unity are a and {3, the stronger the market integration. 9. Because immediate adjustment is not likely to be the case, the next question is how can one transform the information contained in (3) in such a way so a succinct interpretation of both initial adjustment and error correction term can be given. Stated otherwise, we seek an answer to the following question: How many periods does it take for the price of the commodity in location 1 to adjust to a given price change in location 2? 10. The number of periods can be calculated by solving (3). In particular, the solution is related to the parameters a and P as k = l-(l-p)(l-a),,·I, where k is the percentage of adjustment which will take place within n periods.' It is clear that for values of a and P close to unity, a small n (number of periods) is required to give a k close to unity (i.e., full adjustment). Note that if {3 = 1 then k = 1. Finally, if lags are included in (3), {3 will be adjusted up to the period for which the last lag is included. 2. International :Market Integration 11. This section examines whether and to what extent the decisions taken by Mexican government through CONASUPO regarding commodity prices have been influenced by fluctuations in world prices. Accordingly, PIt and P2I denote the domestic and the world price of the commodity. We focused on the following commodities: corn, rice, sorghum, soybeans, and wheat. The sample covers the 1960-1991 period -- a total of 32 observations. Domestic prices are the "Precios Medios Rurales" reported in Econotecnica Agricola, No 9, SARH, 1991 and refer to guarantee prices (precios de concertacion). As world prices for corn, sorghum, and wheat the U.S. Gulf Ports prices were used, while for rice the New Orleans and for Soybeans the U.S. Rotterdam price were used -- all taken from International Financial Statistics, IMF. The official spot US$lPeso exchange rate was used to convert domestic prices into US$. Finally, prices were adjusted by the U.S. CPI.  -3- 12. The first step was to examine the stationarity properties of the prices being considered.8 All three tests indicated that both domestic and world prices are nonstationary. Differencing them once, however, induced stationarity thus concluding that prices are integrated of order one (we do not report results from these tests). The next step was to examine whether the price differentials were stationary. Results are reported in Table A4.1. For com, sorghum, and soybeans all three tests indicated that there exists co movement between world and domestic prices (at the 5 % level of significance). Tests regarding the prices of rice gave no indication of comovement while for the prices of wheat comovement was supported by one test only (ADF, 5% level). • 13. Since wheat and rice price differential were not stationary, we run regression (1) and examined stationary of the error term. In what follows, regression results concerning wheat prices are reported (subscripts 1 and 2 denote domestic and world price, t-values in parentheses): Pl~ = 92.63 + 0.60 P:, (4.01) (4.53) R2 0.41, F(2,29) = 20.55, DW = 0.644, DF = -2.63, ADF == -2.76, PP == -2.57. None of the statistics is in favor of a stationary residual and hence the hypothesis that prices move together in a proportional fashion is rejected. Although the cointegration regression can be run either way, because we implicitly assign an endogeneity pattern by assuming that the domestic price is "being set" by the world price and not vice-versa, we only consider regressions with the world price being the explanatory variable. To further examine the relationship between domestic and world prices for wheat and rice, a linear time trend was included in the regressions. Results regarding wheat are as follows: Pl~ == 202.06 - 3.60 Trend + 0.28 P:, (8.58) (-6.16) (2.96) R2 = 0.74, F(2,29) = 41.89, DW == 1.224, DF -3.44, ADF == -4.25, PP == -3.37. As evidenced by all three cointegration statistics, the results present strong evidence of proportional movement around a linear time trend. 14. The corresponding regression for rice is as follows: Pl~ == 73.67 + 0.34 P:, (3.46) (9.52) R2 0.75, F(2,29) == 90.54, DW == 1.463, DF = -4.20, ADF == -6.42, PP = -4.33. The rice-regression gave very strong evidence of co integration as all three statistics far exceeded their critical values.  -4- 15. The slope coefficients reported above may be interpreted as price transmission elasticities, along the lines discussed by Mundlak and Larson (1992). Such elasticity gives the proportion of world price variation transmitted into the domestic price. In the case of the wheat regression with time trend, the transmission elasticity is equal to 0.26 (calculated as 0.28*p2IWIplt W, evaluated at sample means). The corresponding elasticity for rice is 0.72 (calculated as 0.34*p2IR/Plt~. 16. The last step with respect to international market integration is to construct the error correction specification for the commodities for which the price differential was stationary. Panel B of Table A4.1 reports results consistent with relationship (3) for corn, sorghum, and soybeans.' All three models performed satisfactorily in the sense that both error correction and immediate adjustment • coefficients are highly significant. The hypothesis of immediate adjustment (i.e., Ho of (3» was uniformly rejected as the F-statistics far exceeded the 1 % critical value of 4.54 (64.94 for maize, 13.18 for sorghum, and 11.74 for soybeans). Similar F-statistics were derived when the intercept term was allowed to be different from zero, but the slope coefficients were restricted to unity (13.19, 15.53, and 11.57 for maize, sorghum, and wheat respectively, the F-value is 5.42, 1 % level). 17. When error-correction and immediate adjustment terms are combined into the single measure outlined earlier (column denoted by n, Panel B, Table A4.1), it is found that for any change observed in the world price of corn, it will take an average of 3.2 years for 90% of that change to be transmitted into the domestic price. The associated adjustment period for sorghum and soybeans is 2.9 and 2.8 years. Therefore, despite the heavy interventionist environment of the price setting mechanism, for three of the five commodities examined above, world price signals were transmitted into domestic markets rather quickly. 18. A comment with respect to the limitations of the model is in order. Examining the relationship between domestic and world prices, which, adjusted by the official exchange rate only, admittedly, ignores many issues such as deviations from equilibrium exchange rates, transportation costs, and marketing costs. Therefore, one may wonder whether rejection of long-run integration is a result of these factors or because the government does not adjust the domestic prices according to world price signals. The fact that some of the commodities indicated long-run integration is enough to rule out the latter possibility. However, some of the above mentioned factors may be responsible for deviations from immediate adjustment. 3. Domestic Market Integration 19. The section focusses on domestic markets. In particular, the markets of milled rice (arroz pulido), traditional beans (frijol de mayo), and white beans (haba) are examined. In the rice market and traditional bean market, four destinations (consumption regions) were considered: Guadalajara, Federal District (it includes Mexico City), Ecatepec, and Torreon. The source of rice and traditional beans for all four regions is Sinaloa. For the market of white beans the following four consumption regions were selected: Guadalajara, Federal District, Ecatepec, and Monterrey. Finally, white beans originate from two sources: the states of Puebla and Mexico. The prices in the consumption regions are determined in local wholesale markets reflecting supply and demand conditions. Note that at this level of trade government plays no role as the reported prices are the ones that retailers pay to the wholesalers under competitive market conditions. 20. The data were obtained from Anuario Estadistico de Granos y Semillas, 1991 & 1992, published by SNIM (Servicio Nacional de Informaci6n de Mercados), a price monitoring agency. The series are weekly and cover the 1991-92 time period -- a total of 104 observations. Prices refer  -5- to products of high quality, quoted in the morning of each Wednesday at the wholesale markets (Centros de Abastos). 21. Despite the fact that the sample consists of 104 observations, the time elapsed is admittedly short. However, two advantages must be mentioned. First, during these two years, inflation has been moderate so that using nominal prices does not present a problem (thus avoiding the assumption that inflation equally affects all prices in all regions). Second, transportation and storage costs are unlikely to have changed by much within this time period. Therefore, it is expected that the price differential must reflect mainly regional disparities in demand and supply conditions . • 22. The order of integration of all prices was first determined. With the exception of the price of white beans at Monterrey which was stationary, all other prices were integrated of order one (we do not report results regarding these tests). Therefore, at this stage one can conclude that the market of Monterrey is not integrated in the long run with the markets of Federal District, Guadalajara, and Ecatepec. 23. The order of integration of the price differentials was next examined. In principle, it is not necessary to examine all possible price differentials as stationarity of the price difference between locations 1 and 2 as well stationarity of the price difference between locations 2 and 3 would, by transitivity, imply stationarity of the price difference between locations 1 and 3. However, transitivity is an argument which may not be applicable to econometric models. Therefore, we examined the stationarity properties of all price differentials. 24. Results reported in Table A4.2 indicate that the three markets which are in physical proximity (i.e., Federal District, Ecatepec, Guadalajara) showed some degree of comovement, while distant markets (i.e., Torreon and Monterrey) were found to be segmented. Therefore, one may argue that long-run market integration, at least for the commodities examined, is a geography specific rather than a commodity specific issue, as all three commodities presented fairly similar outcome. However, results from the dynamic specification (fable A4.3) indicate that the transmission of price changes among markets is rather slow. Finally, figures 1 and 2 depict the adjustment paths for Guadalajara-Ecatepec and Ecatepec-F. District respectively for up to 31 weeks. The horizontal axis represents periods (in weeks) while the vertical axis represents adjustment (expressed in percentage units). 4. Conclusions 25. This Annex developed a model which examined the degree to which Mexican commodity markets are integrated on international as well as on domestic level. Since the Mexican government sets producer's prices, international integration, in a sense, quantifies the extend to which the government incorporated the signals sent by the world market of the respective commodities. It was found that, while such signals were conveyed in the prices of maize, sorghum, and soybeans with an average adjustment period of three years, this was not the case for rice and wheat. 26. On the domestic side, the degree of market integration for the commodities of milled rice, traditional and white beans was examined. While long-run integration was found in markets with physical proximity, this was not the case with distant markets. On the other hand, for the markets that were integrated in the long run, the dynamic specification indicated that price changes are transmitted at a very low speed from market to market, a result due to inadequate infrastructure that makes the movement of crops very slow.  -6- ENDNOTES 1. In some instances prices are expressed in logarithms. In other instances (1) explicitly includes transaction costs. 2. A series whose mean and variance exist is said to be stationary and is usually denoted as I(O}. If the mean and variance do not exist, the series is nonstationary and is generally denoted as I(d} and needs to be differenced d times to induce stationarity; if d = 1, the series is integrated of order 1. Here we will consider /(O) and 1(1} series. Two or more l(l} series, a non-trivial linear combination of which is I(O}, are said to be cointegrated (Engle and Granger, 1987). • 3. Nonstationarity of primary commodity prices has been well documented in Ardeni (1989) and Baffes (1991). Goodwin and Schroeder (1991) have shown that U.S. cattle prices are nonstationary. Clark and Youngblood (1992) have reached a similar conclusion regarding input prices in Canadian agriculture. 4. The term power is used in its economic sense. For power of stationarity tests in the statistical sense see DeJ ong et al (1992). s. Hendry, Pagan, and Sargan (1984) discuss this type of model in full detail. The formulation followed here is the one introduced by Engle and Granger (1987). To achieve white noise in (3), lags of the differenced series may be required; however, this is a data specific issue. Finally, it should be noted that the level of significance of the error-correction coefficient may serve as a cointegration test. 6. One can construct the multivariate counterpart of (3). For example, in the three-region market, the dynamic specification will include in the set of explanatory variables the price differential between regions 1 and 3 as well as the price difference of region 3. 7. Consider again the series k = 1-(I-fJ} (l-a)"'J. The second term is a geometric sequence with ratio (1-a). For n = 1, (i. e., the first period) k takes the value of fJ, which captures the current effect of (p" - P2t.J) on (P/l - pJt.I). For n = 2 (i.e., the second period) k takes the value of fJ+ (l-fJ}a, which is the effect of the previous period, fJ, plus the "corrected error" part, (l-fJ}a, and so on. 8. To determine the order of integration we utilized the following three procedures: (aJ Dickey-Fuller (OF), (b) augmented Dickey-Fuller (ADF) , and (c) Phillips-Perron (PP). Let z, denote the variable under consideration. The DF test is based on the regression .dzt = p. + fJZ,.I + ftJ where 4 denotes the difference operator (i. e., ..6.z, = zr - Zt.J) while p. and fJ are parameters to be estimated. A negative and significantly different from zero value of fJ indicates that Zr is I(O} [Fuller (1976); Dickey and Fuller (1981)], The ADF test accounts for the possibility that f t is not white noise; it is based on the following regression: .dz, = p. + fJz,.I + lags(Az,) + f,. Again, a negative and significantly different from zero value of fJ indicates that z, is 1(0}. The PP test is similar to the ADF one; their difference lies on the treatment of any "nuisance" serial correlation aside from that generated by the hypothesized unit root [Phillips (1987); Phillips and Perron (1988)]. To identify the presence of one unit root we test Hc': Z, is not I(O} against the alternative HI" Z, is /(O). 9. Even though the residual for the wheat trend-regression was stationary, since the price differential was not, we did not proceed with the error correction specification. As Meese (1986) and West (1987) observe, noncointegration (with unity slope coefficient in our case) can be attributed to omitted nonstationary variables. Therefore, an additional component has to be included to fully account for price variability. However, since this component is unknown, specification (3) will give biased estimates. Baffes and Shah (1993) discuss the subject in more detail.  -7- Table A4.1: Long- and Short-Run Integration - International Markets PANEL A: Long-Run Integration -- Stationarity oj Price Differentials Commodity DF ADF PP CORN -3.02@ -3.4f~ -3.03@ RICE -2.61 -2.88 -2.61 SORGHUM -3.65@ -3.12@ -3.84@ • SOYBEANS -3.68@ -3.44@ -3.68@ WHEAT -2.71 -3.04@ -2.68 PANELB: Short-Run Integration -- Error-Correction Model Commodity JL ex {J J(l DW n CORN 34.39@ 0.57@ 0.37@ 0.49 1.69 3.2 (3.62) (4.11) (2.02) SORGHUM 4.45 0.65@ 0.39111 0.57 1.55 2.9 (1.35) (5.13) (2.86) SOYBEANS 18.35@ 0.6OI/iI 0.50lIl 0.53 2.06 2.8 (1.83) (4.06) (4.37) NOTES: Panel A: The sample covers the 1960-1991 period (32 observations). For the definitions of DF, ADF, and PP see endnote 8. The 5% level of significance is -3.00. That is, a statistic less than -3.00 indicates that the market under consideration is integrated in the long run (as marked by @). The ADF test contained at least one lag; when a second lag was significant it was included in the regression, and so on. Panel B: The estimates are based on equation (3); the dependent variable is the first difference of the domestic price while the independent variables are the lagged price differential (world minus domestic) and the first difference of the world price. We did not use any lags of the differenced variables as in all cases the first lag was not significant at the 5 % level. The numbers in parentheses denote I-statistics. @ denotes parameter estimate significant at the 5% level. DW is the Durbin-Watson statistic. The last column (n) gives the number of periods (years) it takes for 90% of the adjustment to take place in the domestic price due to changes in the world prices. n is related to a and {j as 0.90 = l-(J-{j)(J-aTJ •  -8- Table A4.2: Long-Run Integration - Domestic Markets DF ADF PP Milled Rice F. DISTRICT - GUADALAJARA -3.82@ -4.34@ -3.800 F. DISTRICT - TORREON -1.S8 -1.98 -1.92 F. DISTRICT - ECATEPEC -3.490 -2.76 -3.37@ GUADALAJARA - TORREON -1.62 -2.06 -1.94 • GUADALAJARA - ECATEPEC -4.000 -4.62@ -4.070 TORREON - ECATEPEC -1.39 -1.88 -1.71 TraditiomU Beans F. DISTRICT - GUADALAJARA -2.40 -3.()9@ -2.71 F. DISTRICT - TORREON -2.34 -2.67 -2.29 F. DISTRICT - ECATEPEC -3.890 -3.67@ -3.790 GUADALAJARA - TORREON -2.44 -2.66 -2.60 GUADALAJARA - ECATEPEC -3.42@ -4.1S@ -3.690 TORREON - ECATEPEC -3.070 -2.93@ -3.070 White Beans F. DISTRICT - GUADALAJARA -4.3S@ -4.68@ -4.41@ F. DISTRICT - ECATEPEC -3.84@ -3.98@ -3.790 GUADALAJARA - ECATEPEC -S.l1@ -3.97@ -S.2SO NOTES: The sample covers the years 1991 and 1992 (104 weekly observations). For the definitions of DF, ADF, and PP see endnote 8. @ indicates stationary price differential at the S % level of significance -- the critical value is -2.89. Since the price of white beans at Monterrey was stationary, it is concluded that the market of Monterrey is not integrated with the rest of the markets (i.e.• Federal District, Ecatepec, and Guadalajara) even in the long run. ,  -9- Table A4.3: Short-Run Integration - Domestic Markets Milled Traditional White Region 1 Region 2 Rice Beans Beans F. DISTRICI' GUADALAJARA 22 13 GUADALAJARA F. DISTRICI' 22 31 7 ECATEPEC GUADALAJARA 12 16 GUADALAJARA ECATEPEC 15 22 5 F. DISTRICI' ECATEPEC 12 9 ECATEPEC F. DISTRICI' 13 29 21 TORREON ECATEPEC NS NA ECATEPEC TORREON NS 20 NA NOTES: Price in region 1 is the independent variable while price in region 2 is the dependent variable. The figures indicate the number of periods (weeks) it takes for 90% of the adjustment to occur for price changes of region 1 to be transmitted to region 2. For example, if the price of rice in Federal District goes up by 50%, it will take 22 weeks for the price of rice in Guadalajara to increase by 45 %. (--) indicates that the error correction term of the dynamic specification was not significant at the 10% level of significance. (NS) indicates that the corresponding price differential is not stationary; (NA) indicates it is not applicable. For the calculation of the number of weeks see notes in Table A4.1. For other notes see Table A4.2. •  ANNEX 5: SOIL CONSERVATION ANALYTICS 1. Table AS.l of this Annex gives the notation for the analysis; the key parameters are summarized in Table 10.7 of the main text. The simplest estimate of the costs of erosion is NPV of the economic cost of the annual crop yield loss over a relevant period of time. This is done by • where the variables and indices are as defmed in Table AS .1. 2. In theory, the value of crop output could fall to zero as the tenn e{-r+i.)t approaches zero. In practice, farmers would stop producing on severely eroded land if the value of output were less than the cost of production. It is therefore necessary to specify a maximum loss, beyond which farmers stop producing crops on the eroded. This maximum loss is expressed as a share of the value of output and set equal to 20 percent. Using equation (1) and the parameter values in Table 10.7, the NPV of the loss to erosion is USS2,190 million over a period of 22 years. The 22 years is the maximum period during which profitable crop production could occur; beginning in year 23, the farmer would fmd it more profitable to abandon the land than to keep farming it. Table AS.I Notation for Economic Analysis Variables y = crop yield a = area cultivated of a crop p = price of a crop or fertilizer c = cost of a crop production technique u = incremental supervision costs of a soil conservation technique v = incremental supervision costs of a land reserve program z = a land reserve payment to farmers r = a discount rate q = a quantity of erosion saved k = loss of crop yield from erosion « 0) m = gain in crop yield from soil conservation (> 0) GB = gross benefits to produce a crop NB = net benefits to produce a crop C = total cost to produce a crop E = total cost of erosion Indices t = subscript for year i = subscript for a crop j = subscript for a soil conservation technique n = subscript for land receiving no soil conservation technique A = subscript for tolerance level of soil loss & = subscript for lowest technically feasible level of soil loss  -2- 3. The incremental costs of achieving the tolerance level of erosion are given by I, (2) 4. The supervision costs of extending new soil conservation methods are assumed to be ... higher than nonnal extension costs, so u > O. Because cost per ton of erosion saved incurs increasing costs as farmers must work progressively harder to achieve those savings, it is assumed that c.j < c&j' The incremental costs of achieving the lowest feasible level of soil erosion are cij - cin = c&j + U. 5. The gross benefits of one hectare of crop production are yield times price, that is: (3) 6. The benefits of soil conservation initially derive from higher crop yields. The incremental benefits of a soil conservation technique are crop price times the incremental yield attributable to conservation, or: (4) GR· IJ - GRm = p.*(y .. - y.inJ· 1 IJ '\ The crop price is not affected by the incremental yield attributable to erosion control. The incremental yield (the tenn Yij - Yin) has two components. The yield without soil conservation, Yint> would slowly fall as erosion cut crop yields over time; the expression used to approximate this is yint = YinO*ekt • Without erosion, the variable 'k' is equal to 0 and the crop yield is the same in every year, The yield with soil conservation, Yijt' rises as the impact of erosion is attenuated. The expression to approximate this is Yijt = Yino*e(mt) for all years. The variable 'm' is a constant expressing the percentage increase in crop yields from what they would have been had erosion continued with the same force it had before the conservation investments were made. Figure 10.1 illustrates these basic relations. The marginal impact of soil conservation is (5) Ay·It = y.. IJt Ym·t = Ym·0*(1 +m) - Ym·o*ekt 7. The net benefits of achieving a tolerance level of erosion with technique j for crop i in any year are: (6) 8. The discounted present value of net benefits for technique j and crop i is: 4 The year subscript is not shown unless otherwise noted.  -3- 9. The cost per ton of erosion saved in the reserve is the NPV of the supervision costs plus that of the reserve payment to producers, divided by the quantity of erosion saved; those costs are given by the variables 'v' and 'z', respectively, in equation (8): (8) The variable 'q' in equation (8) is the difference in erosion on the uncultivated reserve and the average erosion on the same land under cultivation. The reserve payment to producers is • arbitrarily set equal to 20 percent of the gross benefits (yield times price) of output. Because a reserve program produces no gross economic benefits, its net benefits will always be negative, as is evident from equation (6). 10. The cost per ton of erosion saved by technical changes is the discounted present value of net benefits, divided by the quantity of erosion saved. For a combination of one crop and technique, and letting qj represent the erosion saved by technique 'j', that cost is: 1 ,